diff options
| author | Diego 'Flameeyes' Pettenò <flameeyes@gmail.com> | 2007-04-03 00:38:22 +0200 |
|---|---|---|
| committer | Diego 'Flameeyes' Pettenò <flameeyes@gmail.com> | 2007-04-03 00:38:22 +0200 |
| commit | 0ed2cd4f34189ec303dfac5a30de0abae0decba8 (patch) | |
| tree | 7a8d08d25ca7c81daa9d6cd65fd4f633fd676b41 /src/libffmpeg | |
| parent | 6081bc9a06ee97333769f77a9e5c18a15afb29da (diff) | |
| parent | 3dd7d925c2feb7868a49e7a1a0b953a5aab233f0 (diff) | |
| download | xine-lib-0ed2cd4f34189ec303dfac5a30de0abae0decba8.tar.gz xine-lib-0ed2cd4f34189ec303dfac5a30de0abae0decba8.tar.bz2 | |
Merge changes happened in 1.1 development.
Diffstat (limited to 'src/libffmpeg')
52 files changed, 13117 insertions, 170 deletions
diff --git a/src/libffmpeg/.cvsignore b/src/libffmpeg/.hgignore index 7d926a554..7d926a554 100644 --- a/src/libffmpeg/.cvsignore +++ b/src/libffmpeg/.hgignore diff --git a/src/libffmpeg/Makefile.am b/src/libffmpeg/Makefile.am index cf925bd3b..9ea37cd17 100644 --- a/src/libffmpeg/Makefile.am +++ b/src/libffmpeg/Makefile.am @@ -22,6 +22,9 @@ $(top_builddir)/contrib/ffmpeg/libpostproc/libpostproc.a: $(MAKE) -C $(top_builddir)/contrib/ffmpeg/ -f makefile.xine libpostproc/libpostproc.a endif +# ffmpeg_config.h is generated by configure +DISTCLEANFILES = ffmpeg_config.h + # this must always be included, even if the current machine has no DXR3... EXTRA_DIST = xine_encoder.c diff_to_ffmpeg_cvs.txt @@ -45,7 +48,7 @@ endif xineplug_decode_ff_la_CFLAGS = $(VISIBILITY_FLAG) $(AM_CFLAGS) xineplug_decode_ff_la_LDFLAGS = -avoid-version -module @IMPURE_TEXT_LDFLAGS@ xineplug_decode_ff_la_LIBADD = $(MLIB_LIBS) $(XINE_LIB) -lm $(ZLIB_LIBS) \ - $(link_ffmpeg) $(THREAD_LIBS) + $(link_ffmpeg) $(PTHREAD_LIBS) xineplug_decode_dvaudio_la_CFLAGS = $(VISIBILITY_FLAG) $(AM_CFLAGS) xineplug_decode_dvaudio_la_LDFLAGS = -avoid-version -module diff --git a/src/libffmpeg/audio_decoder.c b/src/libffmpeg/audio_decoder.c index c0a4503e2..7ab1f56b4 100644 --- a/src/libffmpeg/audio_decoder.c +++ b/src/libffmpeg/audio_decoder.c @@ -17,7 +17,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA * - * $Id: audio_decoder.c,v 1.29 2006/10/13 21:41:30 dgp85 Exp $ + * $Id: audio_decoder.c,v 1.33 2007/01/28 18:38:33 miguelfreitas Exp $ * * xine audio decoder plugin using ffmpeg * @@ -25,6 +25,7 @@ #ifdef HAVE_CONFIG_H #include "config.h" +#include "ffmpeg_config.h" #endif #include <stdlib.h> @@ -107,6 +108,8 @@ static const ff_codec_t ff_audio_lookup[] = { {BUF_AUDIO_TRUESPEECH, CODEC_ID_TRUESPEECH, "TrueSpeech (ffmpeg)"}, {BUF_AUDIO_TTA, CODEC_ID_TTA, "True Audio Lossless (ffmpeg)"}, {BUF_AUDIO_SMACKER, CODEC_ID_SMACKAUDIO, "Smacker (ffmpeg)"}, + {BUF_AUDIO_FLVADPCM, CODEC_ID_ADPCM_SWF, "Flash ADPCM (ffmpeg)"}, + {BUF_AUDIO_WAVPACK, CODEC_ID_WAVPACK, "WavPack (ffmpeg)"}, }; @@ -288,7 +291,8 @@ static void ff_audio_decode_data (audio_decoder_t *this_gen, buf_element_t *buf) offset = 0; while (this->size>0) { - bytes_consumed = avcodec_decode_audio (this->context, + decode_buffer_size = AVCODEC_MAX_AUDIO_FRAME_SIZE; + bytes_consumed = avcodec_decode_audio2 (this->context, (int16_t *)this->decode_buffer, &decode_buffer_size, &this->buf[offset], @@ -441,37 +445,106 @@ void *init_audio_plugin (xine_t *xine, void *data) { } static uint32_t supported_audio_types[] = { + #ifdef CONFIG_WMAV1_DECODER BUF_AUDIO_WMAV1, + #endif + #ifdef CONFIG_WMAV2_DECODER BUF_AUDIO_WMAV2, + #endif + #ifdef CONFIG_RA_144_DECODER BUF_AUDIO_14_4, + #endif + #ifdef CONFIG_RA_288_DECODER BUF_AUDIO_28_8, - BUF_AUDIO_MULAW, - BUF_AUDIO_ALAW, + #endif + #ifdef CONFIG_MP3_DECODER + BUF_AUDIO_MPEG, + #endif + #ifdef CONFIG_ADPCM_MS_DECODER BUF_AUDIO_MSADPCM, + #endif + #ifdef CONFIG_ADPCM_IMA_QT_DECODER BUF_AUDIO_QTIMAADPCM, + #endif + #ifdef CONFIG_ADPCM_IMA_WAV_DECODER BUF_AUDIO_MSIMAADPCM, + #endif + #ifdef CONFIG_ADPCM_IMA_DK3_DECODER BUF_AUDIO_DK3ADPCM, + #endif + #ifdef CONFIG_ADPCM_IMA_DK4_DECODER BUF_AUDIO_DK4ADPCM, + #endif + #ifdef CONFIG_ADPCM_IMA_WS_DECODER + BUF_AUDIO_VQA_IMA, + #endif + #ifdef CONFIG_ADPCM_IMA_SMJPEG_DECODER + BUF_AUDIO_SMJPEG_IMA, + #endif + #ifdef CONFIG_ADPCM_XA_DECODER BUF_AUDIO_XA_ADPCM, + #endif + #ifdef CONFIG_ADPCM_4XM_DECODER + BUF_AUDIO_4X_ADPCM, + #endif + #ifdef CONFIG_ADPCM_EA_DECODER + BUF_AUDIO_EA_ADPCM, + #endif + #ifdef CONFIG_PCM_MULAW_DECODER + BUF_AUDIO_MULAW, + #endif + #ifdef CONFIG_PCM_ALAW_DECODER + BUF_AUDIO_ALAW, + #endif + #ifdef CONFIG_ROQ_DPCM_DECODER BUF_AUDIO_ROQ, + #endif + #ifdef CONFIG_INTERPLAY_DPCM_DECODER BUF_AUDIO_INTERPLAY, - BUF_AUDIO_VQA_IMA, - BUF_AUDIO_4X_ADPCM, + #endif + #ifdef CONFIG_MACE3_DECODER BUF_AUDIO_MAC3, + #endif + #ifdef CONFIG_MACE6_DECODER BUF_AUDIO_MAC6, + #endif + #ifdef CONFIG_XAN_DPCM_DECODER BUF_AUDIO_XAN_DPCM, + #endif + #ifdef CONFIG_VMDAUDIO_DECODER BUF_AUDIO_VMD, - BUF_AUDIO_EA_ADPCM, - BUF_AUDIO_SMJPEG_IMA, + #endif + #ifdef CONFIG_FLAC_DECODER BUF_AUDIO_FLAC, - BUF_AUDIO_ALAC, + #endif + #ifdef CONFIG_SHORTEN_DECODER BUF_AUDIO_SHORTEN, - BUF_AUDIO_MPEG, + #endif + #ifdef CONFIG_ALAC_DECODER + BUF_AUDIO_ALAC, + #endif + #ifdef CONFIG_QDM2_DECODER BUF_AUDIO_QDESIGN2, + #endif + #ifdef CONFIG_COOK_DECODER BUF_AUDIO_COOK, + #endif + #ifdef CONFIG_TRUESPEECH_DECODER BUF_AUDIO_TRUESPEECH, + #endif + #ifdef CONFIG_TTA_DECODER BUF_AUDIO_TTA, + #endif + #ifdef CONFIG_SMACKAUDIO_DECODER BUF_AUDIO_SMACKER, + #endif + #ifdef CONFIG_ADPCM_SWF_DECODER + BUF_AUDIO_FLVADPCM, + #endif + #ifdef CONFIG_WAVPACK_DECODER + BUF_AUDIO_WAVPACK, + #endif + 0 }; diff --git a/src/libffmpeg/dvaudio_decoder.c b/src/libffmpeg/dvaudio_decoder.c index 71e58417b..ff9a8435c 100644 --- a/src/libffmpeg/dvaudio_decoder.c +++ b/src/libffmpeg/dvaudio_decoder.c @@ -17,7 +17,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA * - * $Id: dvaudio_decoder.c,v 1.11.2.1 2006/12/02 01:20:07 dgp85 Exp $ + * $Id: dvaudio_decoder.c,v 1.11 2006/07/10 22:08:29 dgp85 Exp $ * * dv audio decoder based on patch by Dan Dennedy <dan@dennedy.org> * @@ -91,6 +91,22 @@ typedef struct dvaudio_decoder_s { } dvaudio_decoder_t; + +enum dv_pack_type { + dv_header525 = 0x3f, /* see dv_write_pack for important details on */ + dv_header625 = 0xbf, /* these two packs */ + dv_timecode = 0x13, + dv_audio_source = 0x50, + dv_audio_control = 0x51, + dv_audio_recdate = 0x52, + dv_audio_rectime = 0x53, + dv_video_source = 0x60, + dv_video_control = 0x61, + dv_viedo_recdate = 0x62, + dv_video_rectime = 0x63, + dv_unknown_pack = 0xff, +}; + /* * This is the dumbest implementation of all -- it simply looks at * a fixed offset and if pack isn't there -- fails. We might want diff --git a/src/libffmpeg/libavcodec/.hgignore b/src/libffmpeg/libavcodec/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/alpha/.hgignore b/src/libffmpeg/libavcodec/alpha/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/alpha/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/armv4l/.hgignore b/src/libffmpeg/libavcodec/armv4l/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/armv4l/dsputil_arm_s.S b/src/libffmpeg/libavcodec/armv4l/dsputil_arm_s.S new file mode 100644 index 000000000..2a3ee9c50 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/dsputil_arm_s.S @@ -0,0 +1,696 @@ +@ +@ ARMv4L optimized DSP utils +@ Copyright (c) 2004 AGAWA Koji <i (AT) atty (DOT) jp> +@ +@ This file is part of FFmpeg. +@ +@ FFmpeg is free software; you can redistribute it and/or +@ modify it under the terms of the GNU Lesser General Public +@ License as published by the Free Software Foundation; either +@ version 2.1 of the License, or (at your option) any later version. +@ +@ FFmpeg is distributed in the hope that it will be useful, +@ but WITHOUT ANY WARRANTY; without even the implied warranty of +@ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU +@ Lesser General Public License for more details. +@ +@ You should have received a copy of the GNU Lesser General Public +@ License along with FFmpeg; if not, write to the Free Software +@ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA +@ + +.macro ADJ_ALIGN_QUADWORD_D shift, Rd0, Rd1, Rd2, Rd3, Rn0, Rn1, Rn2, Rn3, Rn4 + mov \Rd0, \Rn0, lsr #(\shift * 8) + mov \Rd1, \Rn1, lsr #(\shift * 8) + mov \Rd2, \Rn2, lsr #(\shift * 8) + mov \Rd3, \Rn3, lsr #(\shift * 8) + orr \Rd0, \Rd0, \Rn1, lsl #(32 - \shift * 8) + orr \Rd1, \Rd1, \Rn2, lsl #(32 - \shift * 8) + orr \Rd2, \Rd2, \Rn3, lsl #(32 - \shift * 8) + orr \Rd3, \Rd3, \Rn4, lsl #(32 - \shift * 8) +.endm +.macro ADJ_ALIGN_DOUBLEWORD shift, R0, R1, R2 + mov \R0, \R0, lsr #(\shift * 8) + orr \R0, \R0, \R1, lsl #(32 - \shift * 8) + mov \R1, \R1, lsr #(\shift * 8) + orr \R1, \R1, \R2, lsl #(32 - \shift * 8) +.endm +.macro ADJ_ALIGN_DOUBLEWORD_D shift, Rdst0, Rdst1, Rsrc0, Rsrc1, Rsrc2 + mov \Rdst0, \Rsrc0, lsr #(\shift * 8) + mov \Rdst1, \Rsrc1, lsr #(\shift * 8) + orr \Rdst0, \Rdst0, \Rsrc1, lsl #(32 - (\shift * 8)) + orr \Rdst1, \Rdst1, \Rsrc2, lsl #(32 - (\shift * 8)) +.endm + +.macro RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask + @ Rd = (Rn | Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1) + @ Rmask = 0xFEFEFEFE + @ Rn = destroy + eor \Rd0, \Rn0, \Rm0 + eor \Rd1, \Rn1, \Rm1 + orr \Rn0, \Rn0, \Rm0 + orr \Rn1, \Rn1, \Rm1 + and \Rd0, \Rd0, \Rmask + and \Rd1, \Rd1, \Rmask + sub \Rd0, \Rn0, \Rd0, lsr #1 + sub \Rd1, \Rn1, \Rd1, lsr #1 +.endm + +.macro NO_RND_AVG32 Rd0, Rd1, Rn0, Rn1, Rm0, Rm1, Rmask + @ Rd = (Rn & Rm) - (((Rn ^ Rm) & ~0x01010101) >> 1) + @ Rmask = 0xFEFEFEFE + @ Rn = destroy + eor \Rd0, \Rn0, \Rm0 + eor \Rd1, \Rn1, \Rm1 + and \Rn0, \Rn0, \Rm0 + and \Rn1, \Rn1, \Rm1 + and \Rd0, \Rd0, \Rmask + and \Rd1, \Rd1, \Rmask + add \Rd0, \Rn0, \Rd0, lsr #1 + add \Rd1, \Rn1, \Rd1, lsr #1 +.endm + +@ ---------------------------------------------------------------- + .align 8 + .global put_pixels16_arm +put_pixels16_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r11, lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + bic r1, r1, #3 + add r5, r5, r4, lsl #2 + ldrne pc, [r5] +1: + ldmia r1, {r4-r7} + add r1, r1, r2 + stmia r0, {r4-r7} + pld [r1] + subs r3, r3, #1 + add r0, r0, r2 + bne 1b + ldmfd sp!, {r4-r11, pc} + .align 8 +2: + ldmia r1, {r4-r8} + add r1, r1, r2 + ADJ_ALIGN_QUADWORD_D 1, r9, r10, r11, r12, r4, r5, r6, r7, r8 + pld [r1] + subs r3, r3, #1 + stmia r0, {r9-r12} + add r0, r0, r2 + bne 2b + ldmfd sp!, {r4-r11, pc} + .align 8 +3: + ldmia r1, {r4-r8} + add r1, r1, r2 + ADJ_ALIGN_QUADWORD_D 2, r9, r10, r11, r12, r4, r5, r6, r7, r8 + pld [r1] + subs r3, r3, #1 + stmia r0, {r9-r12} + add r0, r0, r2 + bne 3b + ldmfd sp!, {r4-r11, pc} + .align 8 +4: + ldmia r1, {r4-r8} + add r1, r1, r2 + ADJ_ALIGN_QUADWORD_D 3, r9, r10, r11, r12, r4, r5, r6, r7, r8 + pld [r1] + subs r3, r3, #1 + stmia r0, {r9-r12} + add r0, r0, r2 + bne 4b + ldmfd sp!, {r4-r11,pc} + .align 8 +5: + .word 1b + .word 2b + .word 3b + .word 4b + +@ ---------------------------------------------------------------- + .align 8 + .global put_pixels8_arm +put_pixels8_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r5,lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + bic r1, r1, #3 + add r5, r5, r4, lsl #2 + ldrne pc, [r5] +1: + ldmia r1, {r4-r5} + add r1, r1, r2 + subs r3, r3, #1 + pld [r1] + stmia r0, {r4-r5} + add r0, r0, r2 + bne 1b + ldmfd sp!, {r4-r5,pc} + .align 8 +2: + ldmia r1, {r4-r5, r12} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r12 + pld [r1] + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 2b + ldmfd sp!, {r4-r5,pc} + .align 8 +3: + ldmia r1, {r4-r5, r12} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r12 + pld [r1] + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 3b + ldmfd sp!, {r4-r5,pc} + .align 8 +4: + ldmia r1, {r4-r5, r12} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r12 + pld [r1] + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 4b + ldmfd sp!, {r4-r5,pc} + .align 8 +5: + .word 1b + .word 2b + .word 3b + .word 4b + +@ ---------------------------------------------------------------- + .align 8 + .global put_pixels8_x2_arm +put_pixels8_x2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r10,lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + ldr r12, [r5] + add r5, r5, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10 + pld [r1] + RND_AVG32 r8, r9, r4, r5, r6, r7, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 1b + ldmfd sp!, {r4-r10,pc} + .align 8 +2: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10 + ADJ_ALIGN_DOUBLEWORD_D 2, r8, r9, r4, r5, r10 + pld [r1] + RND_AVG32 r4, r5, r6, r7, r8, r9, r12 + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 2b + ldmfd sp!, {r4-r10,pc} + .align 8 +3: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r4, r5, r10 + ADJ_ALIGN_DOUBLEWORD_D 3, r8, r9, r4, r5, r10 + pld [r1] + RND_AVG32 r4, r5, r6, r7, r8, r9, r12 + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 3b + ldmfd sp!, {r4-r10,pc} + .align 8 +4: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r4, r5, r10 + pld [r1] + RND_AVG32 r8, r9, r6, r7, r5, r10, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 4b + ldmfd sp!, {r4-r10,pc} @@ update PC with LR content. + .align 8 +5: + .word 0xFEFEFEFE + .word 2b + .word 3b + .word 4b + + .align 8 + .global put_no_rnd_pixels8_x2_arm +put_no_rnd_pixels8_x2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r10,lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + ldr r12, [r5] + add r5, r5, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10 + pld [r1] + NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 1b + ldmfd sp!, {r4-r10,pc} + .align 8 +2: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 1, r6, r7, r4, r5, r10 + ADJ_ALIGN_DOUBLEWORD_D 2, r8, r9, r4, r5, r10 + pld [r1] + NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12 + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 2b + ldmfd sp!, {r4-r10,pc} + .align 8 +3: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r4, r5, r10 + ADJ_ALIGN_DOUBLEWORD_D 3, r8, r9, r4, r5, r10 + pld [r1] + NO_RND_AVG32 r4, r5, r6, r7, r8, r9, r12 + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 3b + ldmfd sp!, {r4-r10,pc} + .align 8 +4: + ldmia r1, {r4-r5, r10} + add r1, r1, r2 + ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r4, r5, r10 + pld [r1] + NO_RND_AVG32 r8, r9, r6, r7, r5, r10, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 4b + ldmfd sp!, {r4-r10,pc} @@ update PC with LR content. + .align 8 +5: + .word 0xFEFEFEFE + .word 2b + .word 3b + .word 4b + + +@ ---------------------------------------------------------------- + .align 8 + .global put_pixels8_y2_arm +put_pixels8_y2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r11,lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + mov r3, r3, lsr #1 + ldr r12, [r5] + add r5, r5, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + ldmia r1, {r4-r5} + add r1, r1, r2 +6: ldmia r1, {r6-r7} + add r1, r1, r2 + pld [r1] + RND_AVG32 r8, r9, r4, r5, r6, r7, r12 + ldmia r1, {r4-r5} + add r1, r1, r2 + stmia r0, {r8-r9} + add r0, r0, r2 + pld [r1] + RND_AVG32 r8, r9, r6, r7, r4, r5, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +2: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9 + RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6 + subs r3, r3, #1 + RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +3: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9 + RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6 + subs r3, r3, #1 + RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +4: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9 + RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6 + subs r3, r3, #1 + RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + + .align 8 +5: + .word 0xFEFEFEFE + .word 2b + .word 3b + .word 4b + + .align 8 + .global put_no_rnd_pixels8_y2_arm +put_no_rnd_pixels8_y2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r11,lr} @ R14 is also called LR + adr r5, 5f + ands r4, r1, #3 + mov r3, r3, lsr #1 + ldr r12, [r5] + add r5, r5, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + ldmia r1, {r4-r5} + add r1, r1, r2 +6: ldmia r1, {r6-r7} + add r1, r1, r2 + pld [r1] + NO_RND_AVG32 r8, r9, r4, r5, r6, r7, r12 + ldmia r1, {r4-r5} + add r1, r1, r2 + stmia r0, {r8-r9} + add r0, r0, r2 + pld [r1] + NO_RND_AVG32 r8, r9, r6, r7, r4, r5, r12 + subs r3, r3, #1 + stmia r0, {r8-r9} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +2: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r7, r8, r9 + NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 1, r4, r5, r6 + subs r3, r3, #1 + NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +3: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r7, r8, r9 + NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 2, r4, r5, r6 + subs r3, r3, #1 + NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +4: + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6 +6: ldmia r1, {r7-r9} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r7, r8, r9 + NO_RND_AVG32 r10, r11, r4, r5, r7, r8, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + ldmia r1, {r4-r6} + add r1, r1, r2 + pld [r1] + ADJ_ALIGN_DOUBLEWORD 3, r4, r5, r6 + subs r3, r3, #1 + NO_RND_AVG32 r10, r11, r7, r8, r4, r5, r12 + stmia r0, {r10-r11} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} + .align 8 +5: + .word 0xFEFEFEFE + .word 2b + .word 3b + .word 4b + +@ ---------------------------------------------------------------- +.macro RND_XY2_IT align, rnd + @ l1= (a & 0x03030303) + (b & 0x03030303) ?(+ 0x02020202) + @ h1= ((a & 0xFCFCFCFCUL) >> 2) + ((b & 0xFCFCFCFCUL) >> 2) +.if \align == 0 + ldmia r1, {r6-r8} +.elseif \align == 3 + ldmia r1, {r5-r7} +.else + ldmia r1, {r8-r10} +.endif + add r1, r1, r2 + pld [r1] +.if \align == 0 + ADJ_ALIGN_DOUBLEWORD_D 1, r4, r5, r6, r7, r8 +.elseif \align == 1 + ADJ_ALIGN_DOUBLEWORD_D 1, r4, r5, r8, r9, r10 + ADJ_ALIGN_DOUBLEWORD_D 2, r6, r7, r8, r9, r10 +.elseif \align == 2 + ADJ_ALIGN_DOUBLEWORD_D 2, r4, r5, r8, r9, r10 + ADJ_ALIGN_DOUBLEWORD_D 3, r6, r7, r8, r9, r10 +.elseif \align == 3 + ADJ_ALIGN_DOUBLEWORD_D 3, r4, r5, r5, r6, r7 +.endif + ldr r14, [r12, #0] @ 0x03030303 + tst r3, #1 + and r8, r4, r14 + and r9, r5, r14 + and r10, r6, r14 + and r11, r7, r14 +.if \rnd == 1 + ldreq r14, [r12, #16] @ 0x02020202 +.else + ldreq r14, [r12, #28] @ 0x01010101 +.endif + add r8, r8, r10 + add r9, r9, r11 + addeq r8, r8, r14 + addeq r9, r9, r14 + ldr r14, [r12, #20] @ 0xFCFCFCFC >> 2 + and r4, r14, r4, lsr #2 + and r5, r14, r5, lsr #2 + and r6, r14, r6, lsr #2 + and r7, r14, r7, lsr #2 + add r10, r4, r6 + add r11, r5, r7 +.endm + +.macro RND_XY2_EXPAND align, rnd + RND_XY2_IT \align, \rnd +6: stmfd sp!, {r8-r11} + RND_XY2_IT \align, \rnd + ldmfd sp!, {r4-r7} + add r4, r4, r8 + add r5, r5, r9 + add r6, r6, r10 + add r7, r7, r11 + ldr r14, [r12, #24] @ 0x0F0F0F0F + and r4, r14, r4, lsr #2 + and r5, r14, r5, lsr #2 + add r4, r4, r6 + add r5, r5, r7 + subs r3, r3, #1 + stmia r0, {r4-r5} + add r0, r0, r2 + bne 6b + ldmfd sp!, {r4-r11,pc} +.endm + + .align 8 + .global put_pixels8_xy2_arm +put_pixels8_xy2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r11,lr} @ R14 is also called LR + adrl r12, 5f + ands r4, r1, #3 + add r5, r12, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + RND_XY2_EXPAND 0, 1 + + .align 8 +2: + RND_XY2_EXPAND 1, 1 + + .align 8 +3: + RND_XY2_EXPAND 2, 1 + + .align 8 +4: + RND_XY2_EXPAND 3, 1 + +5: + .word 0x03030303 + .word 2b + .word 3b + .word 4b + .word 0x02020202 + .word 0xFCFCFCFC >> 2 + .word 0x0F0F0F0F + .word 0x01010101 + + .align 8 + .global put_no_rnd_pixels8_xy2_arm +put_no_rnd_pixels8_xy2_arm: + @ void func(uint8_t *block, const uint8_t *pixels, int line_size, int h) + @ block = word aligned, pixles = unaligned + pld [r1] + stmfd sp!, {r4-r11,lr} @ R14 is also called LR + adrl r12, 5f + ands r4, r1, #3 + add r5, r12, r4, lsl #2 + bic r1, r1, #3 + ldrne pc, [r5] +1: + RND_XY2_EXPAND 0, 0 + + .align 8 +2: + RND_XY2_EXPAND 1, 0 + + .align 8 +3: + RND_XY2_EXPAND 2, 0 + + .align 8 +4: + RND_XY2_EXPAND 3, 0 + +5: + .word 0x03030303 + .word 2b + .word 3b + .word 4b + .word 0x02020202 + .word 0xFCFCFCFC >> 2 + .word 0x0F0F0F0F + .word 0x01010101 diff --git a/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt.c b/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt.c new file mode 100644 index 000000000..d7401e760 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt.c @@ -0,0 +1,188 @@ +/* + * iWMMXt optimized DSP utils + * Copyright (c) 2004 AGAWA Koji + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "../dsputil.h" + +#define DEF(x, y) x ## _no_rnd_ ## y ##_iwmmxt +#define SET_RND(regd) __asm__ __volatile__ ("mov r12, #1 \n\t tbcsth " #regd ", r12":::"r12"); +#define WAVG2B "wavg2b" +#include "dsputil_iwmmxt_rnd.h" +#undef DEF +#undef SET_RND +#undef WAVG2B + +#define DEF(x, y) x ## _ ## y ##_iwmmxt +#define SET_RND(regd) __asm__ __volatile__ ("mov r12, #2 \n\t tbcsth " #regd ", r12":::"r12"); +#define WAVG2B "wavg2br" +#include "dsputil_iwmmxt_rnd.h" +#undef DEF +#undef SET_RND +#undef WAVG2BR + +// need scheduling +#define OP(AVG) \ + asm volatile ( \ + /* alignment */ \ + "and r12, %[pixels], #7 \n\t" \ + "bic %[pixels], %[pixels], #7 \n\t" \ + "tmcr wcgr1, r12 \n\t" \ + \ + "wldrd wr0, [%[pixels]] \n\t" \ + "wldrd wr1, [%[pixels], #8] \n\t" \ + "add %[pixels], %[pixels], %[line_size] \n\t" \ + "walignr1 wr4, wr0, wr1 \n\t" \ + \ + "1: \n\t" \ + \ + "wldrd wr2, [%[pixels]] \n\t" \ + "wldrd wr3, [%[pixels], #8] \n\t" \ + "add %[pixels], %[pixels], %[line_size] \n\t" \ + "pld [%[pixels]] \n\t" \ + "walignr1 wr5, wr2, wr3 \n\t" \ + AVG " wr6, wr4, wr5 \n\t" \ + "wstrd wr6, [%[block]] \n\t" \ + "add %[block], %[block], %[line_size] \n\t" \ + \ + "wldrd wr0, [%[pixels]] \n\t" \ + "wldrd wr1, [%[pixels], #8] \n\t" \ + "add %[pixels], %[pixels], %[line_size] \n\t" \ + "walignr1 wr4, wr0, wr1 \n\t" \ + "pld [%[pixels]] \n\t" \ + AVG " wr6, wr4, wr5 \n\t" \ + "wstrd wr6, [%[block]] \n\t" \ + "add %[block], %[block], %[line_size] \n\t" \ + \ + "subs %[h], %[h], #2 \n\t" \ + "bne 1b \n\t" \ + : [block]"+r"(block), [pixels]"+r"(pixels), [h]"+r"(h) \ + : [line_size]"r"(line_size) \ + : "memory", "r12"); +void put_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + OP("wavg2br"); +} +void put_no_rnd_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + OP("wavg2b"); +} +#undef OP + +void add_pixels_clamped_iwmmxt(const DCTELEM *block, uint8_t *pixels, int line_size) +{ + uint8_t *pixels2 = pixels + line_size; + + __asm__ __volatile__ ( + "mov r12, #4 \n\t" + "1: \n\t" + "pld [%[pixels], %[line_size2]] \n\t" + "pld [%[pixels2], %[line_size2]] \n\t" + "wldrd wr4, [%[pixels]] \n\t" + "wldrd wr5, [%[pixels2]] \n\t" + "pld [%[block], #32] \n\t" + "wunpckelub wr6, wr4 \n\t" + "wldrd wr0, [%[block]] \n\t" + "wunpckehub wr7, wr4 \n\t" + "wldrd wr1, [%[block], #8] \n\t" + "wunpckelub wr8, wr5 \n\t" + "wldrd wr2, [%[block], #16] \n\t" + "wunpckehub wr9, wr5 \n\t" + "wldrd wr3, [%[block], #24] \n\t" + "add %[block], %[block], #32 \n\t" + "waddhss wr10, wr0, wr6 \n\t" + "waddhss wr11, wr1, wr7 \n\t" + "waddhss wr12, wr2, wr8 \n\t" + "waddhss wr13, wr3, wr9 \n\t" + "wpackhus wr14, wr10, wr11 \n\t" + "wpackhus wr15, wr12, wr13 \n\t" + "wstrd wr14, [%[pixels]] \n\t" + "add %[pixels], %[pixels], %[line_size2] \n\t" + "subs r12, r12, #1 \n\t" + "wstrd wr15, [%[pixels2]] \n\t" + "add %[pixels2], %[pixels2], %[line_size2] \n\t" + "bne 1b \n\t" + : [block]"+r"(block), [pixels]"+r"(pixels), [pixels2]"+r"(pixels2) + : [line_size2]"r"(line_size << 1) + : "cc", "memory", "r12"); +} + +static void nop(uint8_t *block, const uint8_t *pixels, int line_size, int h) +{ + return; +} + +int mm_flags; /* multimedia extension flags */ + +int mm_support(void) +{ + return 0; /* TODO, implement proper detection */ +} + +void dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx) +{ + mm_flags = mm_support(); + + if (avctx->dsp_mask) { + if (avctx->dsp_mask & FF_MM_FORCE) + mm_flags |= (avctx->dsp_mask & 0xffff); + else + mm_flags &= ~(avctx->dsp_mask & 0xffff); + } + + if (!(mm_flags & MM_IWMMXT)) return; + + c->add_pixels_clamped = add_pixels_clamped_iwmmxt; + + c->put_pixels_tab[0][0] = put_pixels16_iwmmxt; + c->put_pixels_tab[0][1] = put_pixels16_x2_iwmmxt; + c->put_pixels_tab[0][2] = put_pixels16_y2_iwmmxt; + c->put_pixels_tab[0][3] = put_pixels16_xy2_iwmmxt; + c->put_no_rnd_pixels_tab[0][0] = put_pixels16_iwmmxt; + c->put_no_rnd_pixels_tab[0][1] = put_no_rnd_pixels16_x2_iwmmxt; + c->put_no_rnd_pixels_tab[0][2] = put_no_rnd_pixels16_y2_iwmmxt; + c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy2_iwmmxt; + + c->put_pixels_tab[1][0] = put_pixels8_iwmmxt; + c->put_pixels_tab[1][1] = put_pixels8_x2_iwmmxt; + c->put_pixels_tab[1][2] = put_pixels8_y2_iwmmxt; + c->put_pixels_tab[1][3] = put_pixels8_xy2_iwmmxt; + c->put_no_rnd_pixels_tab[1][0] = put_pixels8_iwmmxt; + c->put_no_rnd_pixels_tab[1][1] = put_no_rnd_pixels8_x2_iwmmxt; + c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_iwmmxt; + c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_iwmmxt; + + c->avg_pixels_tab[0][0] = avg_pixels16_iwmmxt; + c->avg_pixels_tab[0][1] = avg_pixels16_x2_iwmmxt; + c->avg_pixels_tab[0][2] = avg_pixels16_y2_iwmmxt; + c->avg_pixels_tab[0][3] = avg_pixels16_xy2_iwmmxt; + c->avg_no_rnd_pixels_tab[0][0] = avg_pixels16_iwmmxt; + c->avg_no_rnd_pixels_tab[0][1] = avg_no_rnd_pixels16_x2_iwmmxt; + c->avg_no_rnd_pixels_tab[0][2] = avg_no_rnd_pixels16_y2_iwmmxt; + c->avg_no_rnd_pixels_tab[0][3] = avg_no_rnd_pixels16_xy2_iwmmxt; + + c->avg_pixels_tab[1][0] = avg_pixels8_iwmmxt; + c->avg_pixels_tab[1][1] = avg_pixels8_x2_iwmmxt; + c->avg_pixels_tab[1][2] = avg_pixels8_y2_iwmmxt; + c->avg_pixels_tab[1][3] = avg_pixels8_xy2_iwmmxt; + c->avg_no_rnd_pixels_tab[1][0] = avg_no_rnd_pixels8_iwmmxt; + c->avg_no_rnd_pixels_tab[1][1] = avg_no_rnd_pixels8_x2_iwmmxt; + c->avg_no_rnd_pixels_tab[1][2] = avg_no_rnd_pixels8_y2_iwmmxt; + c->avg_no_rnd_pixels_tab[1][3] = avg_no_rnd_pixels8_xy2_iwmmxt; +} diff --git a/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt_rnd.h b/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt_rnd.h new file mode 100644 index 000000000..51ba61c47 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/dsputil_iwmmxt_rnd.h @@ -0,0 +1,1114 @@ +/* + * iWMMXt optimized DSP utils + * copyright (c) 2004 AGAWA Koji + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +void DEF(put, pixels8)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + __asm__ __volatile__ ( + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r4, %[pixels], %[line_size] \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "1: \n\t" + "wldrd wr0, [%[pixels]] \n\t" + "subs %[h], %[h], #2 \n\t" + "wldrd wr1, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr3, [r4] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wldrd wr4, [r4, #8] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr8, wr0, wr1 \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr10, wr3, wr4 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr10, [r5] \n\t" + "add r5, r5, %[line_size] \n\t" + "bne 1b \n\t" + : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h) + : + : "memory", "r4", "r5", "r12"); +} + +void DEF(avg, pixels8)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + __asm__ __volatile__ ( + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r4, %[pixels], %[line_size] \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "1: \n\t" + "wldrd wr0, [%[pixels]] \n\t" + "subs %[h], %[h], #2 \n\t" + "wldrd wr1, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr3, [r4] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wldrd wr4, [r4, #8] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr8, wr0, wr1 \n\t" + "wldrd wr0, [%[block]] \n\t" + "wldrd wr2, [r5] \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr10, wr3, wr4 \n\t" + WAVG2B" wr8, wr8, wr0 \n\t" + WAVG2B" wr10, wr10, wr2 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr10, [r5] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "add r5, r5, %[line_size] \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + "bne 1b \n\t" + : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h) + : + : "memory", "r4", "r5", "r12"); +} + +void DEF(put, pixels16)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + __asm__ __volatile__ ( + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r4, %[pixels], %[line_size] \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "1: \n\t" + "wldrd wr0, [%[pixels]] \n\t" + "wldrd wr1, [%[pixels], #8] \n\t" + "subs %[h], %[h], #2 \n\t" + "wldrd wr2, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr3, [r4] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr8, wr0, wr1 \n\t" + "wldrd wr4, [r4, #8] \n\t" + "walignr1 wr9, wr1, wr2 \n\t" + "wldrd wr5, [r4, #16] \n\t" + "add r4, r4, %[line_size] \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr10, wr3, wr4 \n\t" + "wstrd wr8, [%[block]] \n\t" + "walignr1 wr11, wr4, wr5 \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr10, [r5] \n\t" + "wstrd wr11, [r5, #8] \n\t" + "add r5, r5, %[line_size] \n\t" + "bne 1b \n\t" + : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h) + : + : "memory", "r4", "r5", "r12"); +} + +void DEF(avg, pixels16)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + __asm__ __volatile__ ( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r4, %[pixels], %[line_size]\n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "1: \n\t" + "wldrd wr0, [%[pixels]] \n\t" + "wldrd wr1, [%[pixels], #8] \n\t" + "subs %[h], %[h], #2 \n\t" + "wldrd wr2, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr3, [r4] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr8, wr0, wr1 \n\t" + "wldrd wr4, [r4, #8] \n\t" + "walignr1 wr9, wr1, wr2 \n\t" + "wldrd wr5, [r4, #16] \n\t" + "add r4, r4, %[line_size] \n\t" + "wldrd wr0, [%[block]] \n\t" + "pld [r4] \n\t" + "wldrd wr1, [%[block], #8] \n\t" + "pld [r4, #32] \n\t" + "wldrd wr2, [r5] \n\t" + "walignr1 wr10, wr3, wr4 \n\t" + "wldrd wr3, [r5, #8] \n\t" + WAVG2B" wr8, wr8, wr0 \n\t" + WAVG2B" wr9, wr9, wr1 \n\t" + WAVG2B" wr10, wr10, wr2 \n\t" + "wstrd wr8, [%[block]] \n\t" + "walignr1 wr11, wr4, wr5 \n\t" + WAVG2B" wr11, wr11, wr3 \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr10, [r5] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "wstrd wr11, [r5, #8] \n\t" + "add r5, r5, %[line_size] \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + "bne 1b \n\t" + : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h) + : + : "memory", "r4", "r5", "r12"); +} + +void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "add r4, %[pixels], %[line_size]\n\t" + "tmcr wcgr2, r12 \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr13, [r4] \n\t" + "pld [%[pixels]] \n\t" + "wldrd wr14, [r4, #8] \n\t" + "pld [%[pixels], #32] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr2, wr13, wr14 \n\t" + "wmoveq wr4, wr11 \n\t" + "wmoveq wr6, wr14 \n\t" + "walignr2ne wr4, wr10, wr11 \n\t" + "walignr2ne wr6, wr13, wr14 \n\t" + WAVG2B" wr0, wr0, wr4 \n\t" + WAVG2B" wr2, wr2, wr6 \n\t" + "wstrd wr0, [%[block]] \n\t" + "subs %[h], %[h], #2 \n\t" + "wstrd wr2, [r5] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "add r5, r5, %[line_size] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "r4", "r5", "r12", "memory"); +} + +void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "add r4, %[pixels], %[line_size]\n\t" + "tmcr wcgr2, r12 \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr13, [r4] \n\t" + "pld [%[pixels]] \n\t" + "wldrd wr14, [r4, #8] \n\t" + "pld [%[pixels], #32] \n\t" + "wldrd wr15, [r4, #16] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + "walignr1 wr2, wr13, wr14 \n\t" + "walignr1 wr3, wr14, wr15 \n\t" + "wmoveq wr4, wr11 \n\t" + "wmoveq wr5, wr12 \n\t" + "wmoveq wr6, wr14 \n\t" + "wmoveq wr7, wr15 \n\t" + "walignr2ne wr4, wr10, wr11 \n\t" + "walignr2ne wr5, wr11, wr12 \n\t" + "walignr2ne wr6, wr13, wr14 \n\t" + "walignr2ne wr7, wr14, wr15 \n\t" + WAVG2B" wr0, wr0, wr4 \n\t" + WAVG2B" wr1, wr1, wr5 \n\t" + "wstrd wr0, [%[block]] \n\t" + WAVG2B" wr2, wr2, wr6 \n\t" + "wstrd wr1, [%[block], #8] \n\t" + WAVG2B" wr3, wr3, wr7 \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr2, [r5] \n\t" + "subs %[h], %[h], #2 \n\t" + "wstrd wr3, [r5, #8] \n\t" + "add r5, r5, %[line_size] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "r4", "r5", "r12", "memory"); +} + +void DEF(avg, pixels8_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "add r4, %[pixels], %[line_size]\n\t" + "tmcr wcgr2, r12 \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr13, [r4] \n\t" + "pld [%[pixels]] \n\t" + "wldrd wr14, [r4, #8] \n\t" + "pld [%[pixels], #32] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr2, wr13, wr14 \n\t" + "wmoveq wr4, wr11 \n\t" + "wmoveq wr6, wr14 \n\t" + "walignr2ne wr4, wr10, wr11 \n\t" + "wldrd wr10, [%[block]] \n\t" + "walignr2ne wr6, wr13, wr14 \n\t" + "wldrd wr12, [r5] \n\t" + WAVG2B" wr0, wr0, wr4 \n\t" + WAVG2B" wr2, wr2, wr6 \n\t" + WAVG2B" wr0, wr0, wr10 \n\t" + WAVG2B" wr2, wr2, wr12 \n\t" + "wstrd wr0, [%[block]] \n\t" + "subs %[h], %[h], #2 \n\t" + "wstrd wr2, [r5] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "add r5, r5, %[line_size] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "r4", "r5", "r12", "memory"); +} + +void DEF(avg, pixels16_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "add r4, %[pixels], %[line_size]\n\t" + "tmcr wcgr2, r12 \n\t" + "add r5, %[block], %[line_size] \n\t" + "mov %[line_size], %[line_size], lsl #1 \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "wldrd wr13, [r4] \n\t" + "pld [%[pixels]] \n\t" + "wldrd wr14, [r4, #8] \n\t" + "pld [%[pixels], #32] \n\t" + "wldrd wr15, [r4, #16] \n\t" + "add r4, r4, %[line_size] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "pld [r4] \n\t" + "pld [r4, #32] \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + "walignr1 wr2, wr13, wr14 \n\t" + "walignr1 wr3, wr14, wr15 \n\t" + "wmoveq wr4, wr11 \n\t" + "wmoveq wr5, wr12 \n\t" + "wmoveq wr6, wr14 \n\t" + "wmoveq wr7, wr15 \n\t" + "walignr2ne wr4, wr10, wr11 \n\t" + "walignr2ne wr5, wr11, wr12 \n\t" + "walignr2ne wr6, wr13, wr14 \n\t" + "walignr2ne wr7, wr14, wr15 \n\t" + "wldrd wr10, [%[block]] \n\t" + WAVG2B" wr0, wr0, wr4 \n\t" + "wldrd wr11, [%[block], #8] \n\t" + WAVG2B" wr1, wr1, wr5 \n\t" + "wldrd wr12, [r5] \n\t" + WAVG2B" wr2, wr2, wr6 \n\t" + "wldrd wr13, [r5, #8] \n\t" + WAVG2B" wr3, wr3, wr7 \n\t" + WAVG2B" wr0, wr0, wr10 \n\t" + WAVG2B" wr1, wr1, wr11 \n\t" + WAVG2B" wr2, wr2, wr12 \n\t" + WAVG2B" wr3, wr3, wr13 \n\t" + "wstrd wr0, [%[block]] \n\t" + "subs %[h], %[h], #2 \n\t" + "wstrd wr1, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wstrd wr2, [r5] \n\t" + "pld [%[block]] \n\t" + "wstrd wr3, [r5, #8] \n\t" + "add r5, r5, %[line_size] \n\t" + "pld [%[block], #32] \n\t" + "pld [r5] \n\t" + "pld [r5, #32] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + :"r4", "r5", "r12", "memory"); +} + +void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "pld [%[block]] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr4, wr10, wr11 \n\t" + "wldrd wr10, [%[block]] \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr8, wr8, wr10 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "pld [%[block]] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "wldrd wr10, [%[block]] \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr8, wr8, wr10 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "subs %[h], %[h], #2 \n\t" + "pld [%[block]] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "cc", "memory", "r12"); +} + +void DEF(put, pixels16_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr4, wr10, wr11 \n\t" + "walignr1 wr5, wr11, wr12 \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr9, wr1, wr5 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr9, wr1, wr5 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "subs %[h], %[h], #2 \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "r4", "r5", "r12", "memory"); +} + +void DEF(avg, pixels16_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + int stride = line_size; + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "and r12, %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "pld [%[block]] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + + "1: \n\t" + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr4, wr10, wr11 \n\t" + "walignr1 wr5, wr11, wr12 \n\t" + "wldrd wr10, [%[block]] \n\t" + "wldrd wr11, [%[block], #8] \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr9, wr1, wr5 \n\t" + WAVG2B" wr8, wr8, wr10 \n\t" + WAVG2B" wr9, wr9, wr11 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "wldrd wr10, [%[pixels]] \n\t" + "wldrd wr11, [%[pixels], #8] \n\t" + "pld [%[block]] \n\t" + "wldrd wr12, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr0, wr10, wr11 \n\t" + "walignr1 wr1, wr11, wr12 \n\t" + "wldrd wr10, [%[block]] \n\t" + "wldrd wr11, [%[block], #8] \n\t" + WAVG2B" wr8, wr0, wr4 \n\t" + WAVG2B" wr9, wr1, wr5 \n\t" + WAVG2B" wr8, wr8, wr10 \n\t" + WAVG2B" wr9, wr9, wr11 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "subs %[h], %[h], #2 \n\t" + "pld [%[block]] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride) + : + : "r4", "r5", "r12", "memory"); +} + +void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "mov r12, #2 \n\t" + "pld [%[pixels], #32] \n\t" + "tmcr wcgr0, r12 \n\t" /* for shift value */ + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "add r12, r12, #1 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "tmcr wcgr2, r12 \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "cmp r12, #8 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + + "1: \n\t" + // [wr0 wr1 wr2 wr3] + // [wr4 wr5 wr6 wr7] <= * + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr6, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr4, wr6 \n\t" + "wunpckehub wr5, wr6 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr4, wr4, wr8 \n\t" + "waddhus wr5, wr5, wr9 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "subs %[h], %[h], #2 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block) + : [line_size]"r"(line_size) + : "r12", "memory"); +} + +void DEF(put, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[pixels]] \n\t" + "mov r12, #2 \n\t" + "pld [%[pixels], #32] \n\t" + "tmcr wcgr0, r12 \n\t" /* for shift value */ + /* alignment */ + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "tmcr wcgr2, r12 \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr3, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr2, wr3 \n\t" + "wunpckehub wr3, wr3 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr2, wr2, wr10 \n\t" + "waddhus wr3, wr3, wr11 \n\t" + + "1: \n\t" + // [wr0 wr1 wr2 wr3] + // [wr4 wr5 wr6 wr7] <= * + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr6, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr7, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr4, wr6 \n\t" + "wunpckehub wr5, wr6 \n\t" + "wunpckelub wr6, wr7 \n\t" + "wunpckehub wr7, wr7 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr4, wr4, wr8 \n\t" + "waddhus wr5, wr5, wr9 \n\t" + "waddhus wr6, wr6, wr10 \n\t" + "waddhus wr7, wr7, wr11 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr10, wr2, wr6 \n\t" + "waddhus wr11, wr3, wr7 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "waddhus wr10, wr10, wr15 \n\t" + "waddhus wr11, wr11, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wsrlhg wr10, wr10, wcgr0 \n\t" + "wsrlhg wr11, wr11, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "wpackhus wr9, wr10, wr11 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr3, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr2, wr3 \n\t" + "wunpckehub wr3, wr3 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr2, wr2, wr10 \n\t" + "waddhus wr3, wr3, wr11 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr10, wr2, wr6 \n\t" + "waddhus wr11, wr3, wr7 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "waddhus wr10, wr10, wr15 \n\t" + "waddhus wr11, wr11, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wsrlhg wr10, wr10, wcgr0 \n\t" + "wsrlhg wr11, wr11, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "wpackhus wr9, wr10, wr11 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + "subs %[h], %[h], #2 \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block) + : [line_size]"r"(line_size) + : "r12", "memory"); +} + +void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "pld [%[pixels]] \n\t" + "mov r12, #2 \n\t" + "pld [%[pixels], #32] \n\t" + "tmcr wcgr0, r12 \n\t" /* for shift value */ + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "add r12, r12, #1 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "tmcr wcgr2, r12 \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "cmp r12, #8 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + + "1: \n\t" + // [wr0 wr1 wr2 wr3] + // [wr4 wr5 wr6 wr7] <= * + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr6, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr4, wr6 \n\t" + "wunpckehub wr5, wr6 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr4, wr4, wr8 \n\t" + "waddhus wr5, wr5, wr9 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "wldrd wr12, [%[block]] \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + WAVG2B" wr8, wr8, wr12 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "wldrd wr12, [%[pixels]] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr13, [%[pixels], #8] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "wmoveq wr10, wr13 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "wldrd wr12, [%[block]] \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "subs %[h], %[h], #2 \n\t" + WAVG2B" wr8, wr8, wr12 \n\t" + "wstrd wr8, [%[block]] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block) + : [line_size]"r"(line_size) + : "r12", "memory"); +} + +void DEF(avg, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h) +{ + // [wr0 wr1 wr2 wr3] for previous line + // [wr4 wr5 wr6 wr7] for current line + SET_RND(wr15); // =2 for rnd and =1 for no_rnd version + __asm__ __volatile__( + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "pld [%[pixels]] \n\t" + "mov r12, #2 \n\t" + "pld [%[pixels], #32] \n\t" + "tmcr wcgr0, r12 \n\t" /* for shift value */ + /* alignment */ + "and r12, %[pixels], #7 \n\t" + "bic %[pixels], %[pixels], #7 \n\t" + "tmcr wcgr1, r12 \n\t" + "add r12, r12, #1 \n\t" + "tmcr wcgr2, r12 \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "pld [%[pixels]] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr3, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr2, wr3 \n\t" + "wunpckehub wr3, wr3 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr2, wr2, wr10 \n\t" + "waddhus wr3, wr3, wr11 \n\t" + + "1: \n\t" + // [wr0 wr1 wr2 wr3] + // [wr4 wr5 wr6 wr7] <= * + "wldrd wr12, [%[pixels]] \n\t" + "cmp r12, #8 \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr6, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr7, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr4, wr6 \n\t" + "wunpckehub wr5, wr6 \n\t" + "wunpckelub wr6, wr7 \n\t" + "wunpckehub wr7, wr7 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr4, wr4, wr8 \n\t" + "waddhus wr5, wr5, wr9 \n\t" + "waddhus wr6, wr6, wr10 \n\t" + "waddhus wr7, wr7, wr11 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr10, wr2, wr6 \n\t" + "waddhus wr11, wr3, wr7 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "waddhus wr10, wr10, wr15 \n\t" + "waddhus wr11, wr11, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wldrd wr12, [%[block]] \n\t" + "wldrd wr13, [%[block], #8] \n\t" + "wsrlhg wr10, wr10, wcgr0 \n\t" + "wsrlhg wr11, wr11, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "wpackhus wr9, wr10, wr11 \n\t" + WAVG2B" wr8, wr8, wr12 \n\t" + WAVG2B" wr9, wr9, wr13 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + + // [wr0 wr1 wr2 wr3] <= * + // [wr4 wr5 wr6 wr7] + "wldrd wr12, [%[pixels]] \n\t" + "pld [%[block]] \n\t" + "wldrd wr13, [%[pixels], #8] \n\t" + "pld [%[block], #32] \n\t" + "wldrd wr14, [%[pixels], #16] \n\t" + "add %[pixels], %[pixels], %[line_size] \n\t" + "walignr1 wr2, wr12, wr13 \n\t" + "pld [%[pixels]] \n\t" + "pld [%[pixels], #32] \n\t" + "walignr1 wr3, wr13, wr14 \n\t" + "wmoveq wr10, wr13 \n\t" + "wmoveq wr11, wr14 \n\t" + "walignr2ne wr10, wr12, wr13 \n\t" + "walignr2ne wr11, wr13, wr14 \n\t" + "wunpckelub wr0, wr2 \n\t" + "wunpckehub wr1, wr2 \n\t" + "wunpckelub wr2, wr3 \n\t" + "wunpckehub wr3, wr3 \n\t" + "wunpckelub wr8, wr10 \n\t" + "wunpckehub wr9, wr10 \n\t" + "wunpckelub wr10, wr11 \n\t" + "wunpckehub wr11, wr11 \n\t" + "waddhus wr0, wr0, wr8 \n\t" + "waddhus wr1, wr1, wr9 \n\t" + "waddhus wr2, wr2, wr10 \n\t" + "waddhus wr3, wr3, wr11 \n\t" + "waddhus wr8, wr0, wr4 \n\t" + "waddhus wr9, wr1, wr5 \n\t" + "waddhus wr10, wr2, wr6 \n\t" + "waddhus wr11, wr3, wr7 \n\t" + "waddhus wr8, wr8, wr15 \n\t" + "waddhus wr9, wr9, wr15 \n\t" + "waddhus wr10, wr10, wr15 \n\t" + "waddhus wr11, wr11, wr15 \n\t" + "wsrlhg wr8, wr8, wcgr0 \n\t" + "wsrlhg wr9, wr9, wcgr0 \n\t" + "wldrd wr12, [%[block]] \n\t" + "wldrd wr13, [%[block], #8] \n\t" + "wsrlhg wr10, wr10, wcgr0 \n\t" + "wsrlhg wr11, wr11, wcgr0 \n\t" + "wpackhus wr8, wr8, wr9 \n\t" + "wpackhus wr9, wr10, wr11 \n\t" + WAVG2B" wr8, wr8, wr12 \n\t" + WAVG2B" wr9, wr9, wr13 \n\t" + "wstrd wr8, [%[block]] \n\t" + "wstrd wr9, [%[block], #8] \n\t" + "add %[block], %[block], %[line_size] \n\t" + "subs %[h], %[h], #2 \n\t" + "pld [%[block]] \n\t" + "pld [%[block], #32] \n\t" + "bne 1b \n\t" + : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block) + : [line_size]"r"(line_size) + : "r12", "memory"); +} diff --git a/src/libffmpeg/libavcodec/armv4l/mathops.h b/src/libffmpeg/libavcodec/armv4l/mathops.h new file mode 100644 index 000000000..7ddd0ec6e --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/mathops.h @@ -0,0 +1,49 @@ +/* + * simple math operations + * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifdef FRAC_BITS +# define MULL(a, b) \ + ({ int lo, hi;\ + asm("smull %0, %1, %2, %3 \n\t"\ + "mov %0, %0, lsr %4\n\t"\ + "add %1, %0, %1, lsl %5\n\t"\ + : "=&r"(lo), "=&r"(hi)\ + : "r"(b), "r"(a), "i"(FRAC_BITS), "i"(32-FRAC_BITS));\ + hi; }) +#endif + +#define MULH(a, b) \ + ({ int lo, hi;\ + asm ("smull %0, %1, %2, %3" : "=&r"(lo), "=&r"(hi) : "r"(b), "r"(a));\ + hi; }) + +#if defined(HAVE_ARMV5TE) + +/* signed 16x16 -> 32 multiply add accumulate */ +# define MAC16(rt, ra, rb) \ + asm ("smlabb %0, %2, %3, %0" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb)); +/* signed 16x16 -> 32 multiply */ +# define MUL16(ra, rb) \ + ({ int __rt; \ + asm ("smulbb %0, %1, %2" : "=r" (__rt) : "r" (ra), "r" (rb)); \ + __rt; }) + +#endif diff --git a/src/libffmpeg/libavcodec/armv4l/mpegvideo_armv5te.c b/src/libffmpeg/libavcodec/armv4l/mpegvideo_armv5te.c new file mode 100644 index 000000000..a8d09b8ce --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/mpegvideo_armv5te.c @@ -0,0 +1,213 @@ +/* + * Optimization of some functions from mpegvideo.c for armv5te + * Copyright (c) 2007 Siarhei Siamashka <ssvb@users.sourceforge.net> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/* + * Some useful links for those who may be interested in optimizing code for ARM. + * ARM Architecture Reference Manual: http://www.arm.com/community/academy/resources.html + * Instructions timings and optimization guide for ARM9E: http://www.arm.com/pdfs/DDI0222B_9EJS_r1p2.pdf + */ + +#include "../dsputil.h" +#include "../mpegvideo.h" +#include "../avcodec.h" + + +#ifdef ENABLE_ARM_TESTS +/** + * h263 dequantizer supplementary function, it is performance critical and needs to + * have optimized implementations for each architecture. Is also used as a reference + * implementation in regression tests + */ +static inline void dct_unquantize_h263_helper_c(DCTELEM *block, int qmul, int qadd, int count) +{ + int i, level; + for (i = 0; i < count; i++) { + level = block[i]; + if (level) { + if (level < 0) { + level = level * qmul - qadd; + } else { + level = level * qmul + qadd; + } + block[i] = level; + } + } +} +#endif + +/* GCC 3.1 or higher is required to support symbolic names in assembly code */ +#if (__GNUC__ > 3) || ((__GNUC__ == 3) && (__GNUC_MINOR__ >= 1)) + +/** + * Special optimized version of dct_unquantize_h263_helper_c, it requires the block + * to be at least 8 bytes aligned, and may process more elements than requested. + * But it is guaranteed to never process more than 64 elements provided that + * xxcount argument is <= 64, so it is safe. This macro is optimized for a common + * distribution of values for nCoeffs (they are mostly multiple of 8 plus one or + * two extra elements). So this macro processes data as 8 elements per loop iteration + * and contains optional 2 elements processing in the end. + * + * Inner loop should take 6 cycles per element on arm926ej-s (Nokia 770) + */ +#define dct_unquantize_h263_special_helper_armv5te(xxblock, xxqmul, xxqadd, xxcount) \ +({ DCTELEM *xblock = xxblock; \ + int xqmul = xxqmul, xqadd = xxqadd, xcount = xxcount, xtmp; \ + int xdata1, xdata2; \ +__asm__ __volatile__( \ + "subs %[count], #2 \n\t" \ + "ble 2f \n\t" \ + "ldrd r4, [%[block], #0] \n\t" \ + "1: \n\t" \ + "ldrd r6, [%[block], #8] \n\t" \ +\ + "rsbs %[data1], %[zero], r4, asr #16 \n\t" \ + "addgt %[data1], %[qadd], #0 \n\t" \ + "rsblt %[data1], %[qadd], #0 \n\t" \ + "smlatbne %[data1], r4, %[qmul], %[data1] \n\t" \ +\ + "rsbs %[data2], %[zero], r5, asr #16 \n\t" \ + "addgt %[data2], %[qadd], #0 \n\t" \ + "rsblt %[data2], %[qadd], #0 \n\t" \ + "smlatbne %[data2], r5, %[qmul], %[data2] \n\t" \ +\ + "rsbs %[tmp], %[zero], r4, asl #16 \n\t" \ + "addgt %[tmp], %[qadd], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne r4, r4, %[qmul], %[tmp] \n\t" \ +\ + "rsbs %[tmp], %[zero], r5, asl #16 \n\t" \ + "addgt %[tmp], %[qadd], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne r5, r5, %[qmul], %[tmp] \n\t" \ +\ + "strh r4, [%[block]], #2 \n\t" \ + "strh %[data1], [%[block]], #2 \n\t" \ + "strh r5, [%[block]], #2 \n\t" \ + "strh %[data2], [%[block]], #2 \n\t" \ +\ + "rsbs %[data1], %[zero], r6, asr #16 \n\t" \ + "addgt %[data1], %[qadd], #0 \n\t" \ + "rsblt %[data1], %[qadd], #0 \n\t" \ + "smlatbne %[data1], r6, %[qmul], %[data1] \n\t" \ +\ + "rsbs %[data2], %[zero], r7, asr #16 \n\t" \ + "addgt %[data2], %[qadd], #0 \n\t" \ + "rsblt %[data2], %[qadd], #0 \n\t" \ + "smlatbne %[data2], r7, %[qmul], %[data2] \n\t" \ +\ + "rsbs %[tmp], %[zero], r6, asl #16 \n\t" \ + "addgt %[tmp], %[qadd], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne r6, r6, %[qmul], %[tmp] \n\t" \ +\ + "rsbs %[tmp], %[zero], r7, asl #16 \n\t" \ + "addgt %[tmp], %[qadd], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne r7, r7, %[qmul], %[tmp] \n\t" \ +\ + "strh r6, [%[block]], #2 \n\t" \ + "strh %[data1], [%[block]], #2 \n\t" \ + "strh r7, [%[block]], #2 \n\t" \ + "strh %[data2], [%[block]], #2 \n\t" \ +\ + "subs %[count], #8 \n\t" \ + "ldrgtd r4, [%[block], #0] \n\t" /* load data early to avoid load/use pipeline stall */ \ + "bgt 1b \n\t" \ +\ + "adds %[count], #2 \n\t" \ + "ble 3f \n\t" \ + "2: \n\t" \ + "ldrsh %[data1], [%[block], #0] \n\t" \ + "ldrsh %[data2], [%[block], #2] \n\t" \ + "mov %[tmp], %[qadd] \n\t" \ + "cmp %[data1], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne %[data1], %[data1], %[qmul], %[tmp] \n\t" \ + "mov %[tmp], %[qadd] \n\t" \ + "cmp %[data2], #0 \n\t" \ + "rsblt %[tmp], %[qadd], #0 \n\t" \ + "smlabbne %[data2], %[data2], %[qmul], %[tmp] \n\t" \ + "strh %[data1], [%[block]], #2 \n\t" \ + "strh %[data2], [%[block]], #2 \n\t" \ + "3: \n\t" \ + : [block] "+&r" (xblock), [count] "+&r" (xcount), [tmp] "=&r" (xtmp), \ + [data1] "=&r" (xdata1), [data2] "=&r" (xdata2) \ + : [qmul] "r" (xqmul), [qadd] "r" (xqadd), [zero] "r" (0) \ + : "r4", "r5", "r6", "r7", "cc", "memory" \ +); \ +}) + +static void dct_unquantize_h263_intra_armv5te(MpegEncContext *s, + DCTELEM *block, int n, int qscale) +{ + int i, level, qmul, qadd; + int nCoeffs; + + assert(s->block_last_index[n]>=0); + + qmul = qscale << 1; + + if (!s->h263_aic) { + if (n < 4) + level = block[0] * s->y_dc_scale; + else + level = block[0] * s->c_dc_scale; + qadd = (qscale - 1) | 1; + }else{ + qadd = 0; + level = block[0]; + } + if(s->ac_pred) + nCoeffs=63; + else + nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ]; + + dct_unquantize_h263_special_helper_armv5te(block, qmul, qadd, nCoeffs + 1); + block[0] = level; +} + +static void dct_unquantize_h263_inter_armv5te(MpegEncContext *s, + DCTELEM *block, int n, int qscale) +{ + int i, level, qmul, qadd; + int nCoeffs; + + assert(s->block_last_index[n]>=0); + + qadd = (qscale - 1) | 1; + qmul = qscale << 1; + + nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ]; + + dct_unquantize_h263_special_helper_armv5te(block, qmul, qadd, nCoeffs + 1); +} + +#define HAVE_DCT_UNQUANTIZE_H263_ARMV5TE_OPTIMIZED + +#endif + +void MPV_common_init_armv5te(MpegEncContext *s) +{ +#ifdef HAVE_DCT_UNQUANTIZE_H263_ARMV5TE_OPTIMIZED + s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_armv5te; + s->dct_unquantize_h263_inter = dct_unquantize_h263_inter_armv5te; +#endif +} diff --git a/src/libffmpeg/libavcodec/armv4l/mpegvideo_iwmmxt.c b/src/libffmpeg/libavcodec/armv4l/mpegvideo_iwmmxt.c new file mode 100644 index 000000000..1336ac5f8 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/mpegvideo_iwmmxt.c @@ -0,0 +1,119 @@ +/* + * copyright (c) 2004 AGAWA Koji + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "../dsputil.h" +#include "../mpegvideo.h" +#include "../avcodec.h" + +static void dct_unquantize_h263_intra_iwmmxt(MpegEncContext *s, + DCTELEM *block, int n, int qscale) +{ + int level, qmul, qadd; + int nCoeffs; + DCTELEM *block_orig = block; + + assert(s->block_last_index[n]>=0); + + qmul = qscale << 1; + + if (!s->h263_aic) { + if (n < 4) + level = block[0] * s->y_dc_scale; + else + level = block[0] * s->c_dc_scale; + qadd = (qscale - 1) | 1; + }else{ + qadd = 0; + level = block[0]; + } + if(s->ac_pred) + nCoeffs=63; + else + nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ]; + + __asm__ __volatile__ ( +/* "movd %1, %%mm6 \n\t" //qmul */ +/* "packssdw %%mm6, %%mm6 \n\t" */ +/* "packssdw %%mm6, %%mm6 \n\t" */ + "tbcsth wr6, %[qmul] \n\t" +/* "movd %2, %%mm5 \n\t" //qadd */ +/* "packssdw %%mm5, %%mm5 \n\t" */ +/* "packssdw %%mm5, %%mm5 \n\t" */ + "tbcsth wr5, %[qadd] \n\t" + "wzero wr7 \n\t" /* "pxor %%mm7, %%mm7 \n\t" */ + "wzero wr4 \n\t" /* "pxor %%mm4, %%mm4 \n\t" */ + "wsubh wr7, wr5, wr7 \n\t" /* "psubw %%mm5, %%mm7 \n\t" */ + "1: \n\t" + "wldrd wr2, [%[block]] \n\t" /* "movq (%0, %3), %%mm0 \n\t" */ + "wldrd wr3, [%[block], #8] \n\t" /* "movq 8(%0, %3), %%mm1 \n\t" */ + "wmulsl wr0, wr6, wr2 \n\t" /* "pmullw %%mm6, %%mm0 \n\t" */ + "wmulsl wr1, wr6, wr3 \n\t" /* "pmullw %%mm6, %%mm1 \n\t" */ +/* "movq (%0, %3), %%mm2 \n\t" */ +/* "movq 8(%0, %3), %%mm3 \n\t" */ + "wcmpgtsh wr2, wr4, wr2 \n\t" /* "pcmpgtw %%mm4, %%mm2 \n\t" // block[i] < 0 ? -1 : 0 */ + "wcmpgtsh wr3, wr4, wr2 \n\t" /* "pcmpgtw %%mm4, %%mm3 \n\t" // block[i] < 0 ? -1 : 0 */ + "wxor wr0, wr2, wr0 \n\t" /* "pxor %%mm2, %%mm0 \n\t" */ + "wxor wr1, wr3, wr1 \n\t" /* "pxor %%mm3, %%mm1 \n\t" */ + "waddh wr0, wr7, wr0 \n\t" /* "paddw %%mm7, %%mm0 \n\t" */ + "waddh wr1, wr7, wr1 \n\t" /* "paddw %%mm7, %%mm1 \n\t" */ + "wxor wr2, wr0, wr2 \n\t" /* "pxor %%mm0, %%mm2 \n\t" */ + "wxor wr3, wr1, wr3 \n\t" /* "pxor %%mm1, %%mm3 \n\t" */ + "wcmpeqh wr0, wr7, wr0 \n\t" /* "pcmpeqw %%mm7, %%mm0 \n\t" // block[i] == 0 ? -1 : 0 */ + "wcmpeqh wr1, wr7, wr1 \n\t" /* "pcmpeqw %%mm7, %%mm1 \n\t" // block[i] == 0 ? -1 : 0 */ + "wandn wr0, wr2, wr0 \n\t" /* "pandn %%mm2, %%mm0 \n\t" */ + "wandn wr1, wr3, wr1 \n\t" /* "pandn %%mm3, %%mm1 \n\t" */ + "wstrd wr0, [%[block]] \n\t" /* "movq %%mm0, (%0, %3) \n\t" */ + "wstrd wr1, [%[block], #8] \n\t" /* "movq %%mm1, 8(%0, %3) \n\t" */ + "add %[block], %[block], #16 \n\t" /* "addl $16, %3 \n\t" */ + "subs %[i], %[i], #1 \n\t" + "bne 1b \n\t" /* "jng 1b \n\t" */ + :[block]"+r"(block) + :[i]"r"((nCoeffs + 8) / 8), [qmul]"r"(qmul), [qadd]"r"(qadd) + :"memory"); + + block_orig[0] = level; +} + +#if 0 +static void dct_unquantize_h263_inter_iwmmxt(MpegEncContext *s, + DCTELEM *block, int n, int qscale) +{ + int nCoeffs; + + assert(s->block_last_index[n]>=0); + + if(s->ac_pred) + nCoeffs=63; + else + nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ]; + + ippiQuantInvInter_Compact_H263_16s_I(block, nCoeffs+1, qscale); +} +#endif + +void MPV_common_init_iwmmxt(MpegEncContext *s) +{ + if (!(mm_flags & MM_IWMMXT)) return; + + s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_iwmmxt; +#if 0 + s->dct_unquantize_h263_inter = dct_unquantize_h263_inter_iwmmxt; +#endif +} diff --git a/src/libffmpeg/libavcodec/armv4l/simple_idct_armv5te.S b/src/libffmpeg/libavcodec/armv4l/simple_idct_armv5te.S new file mode 100644 index 000000000..28bee0643 --- /dev/null +++ b/src/libffmpeg/libavcodec/armv4l/simple_idct_armv5te.S @@ -0,0 +1,718 @@ +/* + * Simple IDCT + * + * Copyright (c) 2001 Michael Niedermayer <michaelni@gmx.at> + * Copyright (c) 2006 Mans Rullgard <mru@inprovide.com> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#define W1 22725 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W2 21407 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W3 19266 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W4 16383 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W5 12873 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W6 8867 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define W7 4520 /* cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5 */ +#define ROW_SHIFT 11 +#define COL_SHIFT 20 + +#define W13 (W1 | (W3 << 16)) +#define W26 (W2 | (W6 << 16)) +#define W57 (W5 | (W7 << 16)) + + .text + .align +w13: .long W13 +w26: .long W26 +w57: .long W57 + + .align + .func idct_row_armv5te +idct_row_armv5te: + str lr, [sp, #-4]! + + ldrd v1, [a1, #8] + ldrd a3, [a1] /* a3 = row[1:0], a4 = row[3:2] */ + orrs v1, v1, v2 + cmpeq v1, a4 + cmpeq v1, a3, lsr #16 + beq row_dc_only + + mov v1, #(1<<(ROW_SHIFT-1)) + mov ip, #16384 + sub ip, ip, #1 /* ip = W4 */ + smlabb v1, ip, a3, v1 /* v1 = W4*row[0]+(1<<(RS-1)) */ + ldr ip, [pc, #(w26-.-8)] /* ip = W2 | (W6 << 16) */ + smultb a2, ip, a4 + smulbb lr, ip, a4 + add v2, v1, a2 + sub v3, v1, a2 + sub v4, v1, lr + add v1, v1, lr + + ldr ip, [pc, #(w13-.-8)] /* ip = W1 | (W3 << 16) */ + ldr lr, [pc, #(w57-.-8)] /* lr = W5 | (W7 << 16) */ + smulbt v5, ip, a3 + smultt v6, lr, a4 + smlatt v5, ip, a4, v5 + smultt a2, ip, a3 + smulbt v7, lr, a3 + sub v6, v6, a2 + smulbt a2, ip, a4 + smultt fp, lr, a3 + sub v7, v7, a2 + smulbt a2, lr, a4 + ldrd a3, [a1, #8] /* a3=row[5:4] a4=row[7:6] */ + sub fp, fp, a2 + + orrs a2, a3, a4 + beq 1f + + smlabt v5, lr, a3, v5 + smlabt v6, ip, a3, v6 + smlatt v5, lr, a4, v5 + smlabt v6, lr, a4, v6 + smlatt v7, lr, a3, v7 + smlatt fp, ip, a3, fp + smulbt a2, ip, a4 + smlatt v7, ip, a4, v7 + sub fp, fp, a2 + + ldr ip, [pc, #(w26-.-8)] /* ip = W2 | (W6 << 16) */ + mov a2, #16384 + sub a2, a2, #1 /* a2 = W4 */ + smulbb a2, a2, a3 /* a2 = W4*row[4] */ + smultb lr, ip, a4 /* lr = W6*row[6] */ + add v1, v1, a2 /* v1 += W4*row[4] */ + add v1, v1, lr /* v1 += W6*row[6] */ + add v4, v4, a2 /* v4 += W4*row[4] */ + sub v4, v4, lr /* v4 -= W6*row[6] */ + smulbb lr, ip, a4 /* lr = W2*row[6] */ + sub v2, v2, a2 /* v2 -= W4*row[4] */ + sub v2, v2, lr /* v2 -= W2*row[6] */ + sub v3, v3, a2 /* v3 -= W4*row[4] */ + add v3, v3, lr /* v3 += W2*row[6] */ + +1: add a2, v1, v5 + mov a3, a2, lsr #11 + bic a3, a3, #0x1f0000 + sub a2, v2, v6 + mov a2, a2, lsr #11 + add a3, a3, a2, lsl #16 + add a2, v3, v7 + mov a4, a2, lsr #11 + bic a4, a4, #0x1f0000 + add a2, v4, fp + mov a2, a2, lsr #11 + add a4, a4, a2, lsl #16 + strd a3, [a1] + + sub a2, v4, fp + mov a3, a2, lsr #11 + bic a3, a3, #0x1f0000 + sub a2, v3, v7 + mov a2, a2, lsr #11 + add a3, a3, a2, lsl #16 + add a2, v2, v6 + mov a4, a2, lsr #11 + bic a4, a4, #0x1f0000 + sub a2, v1, v5 + mov a2, a2, lsr #11 + add a4, a4, a2, lsl #16 + strd a3, [a1, #8] + + ldr pc, [sp], #4 + +row_dc_only: + orr a3, a3, a3, lsl #16 + bic a3, a3, #0xe000 + mov a3, a3, lsl #3 + mov a4, a3 + strd a3, [a1] + strd a3, [a1, #8] + + ldr pc, [sp], #4 + .endfunc + + .macro idct_col + ldr a4, [a1] /* a4 = col[1:0] */ + mov ip, #16384 + sub ip, ip, #1 /* ip = W4 */ +#if 0 + mov v1, #(1<<(COL_SHIFT-1)) + smlabt v2, ip, a4, v1 /* v2 = W4*col[1] + (1<<(COL_SHIFT-1)) */ + smlabb v1, ip, a4, v1 /* v1 = W4*col[0] + (1<<(COL_SHIFT-1)) */ + ldr a4, [a1, #(16*4)] +#else + mov v1, #((1<<(COL_SHIFT-1))/W4) /* this matches the C version */ + add v2, v1, a4, asr #16 + rsb v2, v2, v2, lsl #14 + mov a4, a4, lsl #16 + add v1, v1, a4, asr #16 + ldr a4, [a1, #(16*4)] + rsb v1, v1, v1, lsl #14 +#endif + + smulbb lr, ip, a4 + smulbt a3, ip, a4 + sub v3, v1, lr + sub v5, v1, lr + add v7, v1, lr + add v1, v1, lr + sub v4, v2, a3 + sub v6, v2, a3 + add fp, v2, a3 + ldr ip, [pc, #(w26-.-8)] + ldr a4, [a1, #(16*2)] + add v2, v2, a3 + + smulbb lr, ip, a4 + smultb a3, ip, a4 + add v1, v1, lr + sub v7, v7, lr + add v3, v3, a3 + sub v5, v5, a3 + smulbt lr, ip, a4 + smultt a3, ip, a4 + add v2, v2, lr + sub fp, fp, lr + add v4, v4, a3 + ldr a4, [a1, #(16*6)] + sub v6, v6, a3 + + smultb lr, ip, a4 + smulbb a3, ip, a4 + add v1, v1, lr + sub v7, v7, lr + sub v3, v3, a3 + add v5, v5, a3 + smultt lr, ip, a4 + smulbt a3, ip, a4 + add v2, v2, lr + sub fp, fp, lr + sub v4, v4, a3 + add v6, v6, a3 + + stmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp} + + ldr ip, [pc, #(w13-.-8)] + ldr a4, [a1, #(16*1)] + ldr lr, [pc, #(w57-.-8)] + smulbb v1, ip, a4 + smultb v3, ip, a4 + smulbb v5, lr, a4 + smultb v7, lr, a4 + smulbt v2, ip, a4 + smultt v4, ip, a4 + smulbt v6, lr, a4 + smultt fp, lr, a4 + rsb v4, v4, #0 + ldr a4, [a1, #(16*3)] + rsb v3, v3, #0 + + smlatb v1, ip, a4, v1 + smlatb v3, lr, a4, v3 + smulbb a3, ip, a4 + smulbb a2, lr, a4 + sub v5, v5, a3 + sub v7, v7, a2 + smlatt v2, ip, a4, v2 + smlatt v4, lr, a4, v4 + smulbt a3, ip, a4 + smulbt a2, lr, a4 + sub v6, v6, a3 + ldr a4, [a1, #(16*5)] + sub fp, fp, a2 + + smlabb v1, lr, a4, v1 + smlabb v3, ip, a4, v3 + smlatb v5, lr, a4, v5 + smlatb v7, ip, a4, v7 + smlabt v2, lr, a4, v2 + smlabt v4, ip, a4, v4 + smlatt v6, lr, a4, v6 + ldr a3, [a1, #(16*7)] + smlatt fp, ip, a4, fp + + smlatb v1, lr, a3, v1 + smlabb v3, lr, a3, v3 + smlatb v5, ip, a3, v5 + smulbb a4, ip, a3 + smlatt v2, lr, a3, v2 + sub v7, v7, a4 + smlabt v4, lr, a3, v4 + smulbt a4, ip, a3 + smlatt v6, ip, a3, v6 + sub fp, fp, a4 + .endm + + .align + .func idct_col_armv5te +idct_col_armv5te: + str lr, [sp, #-4]! + + idct_col + + ldmfd sp!, {a3, a4} + adds a2, a3, v1 + mov a2, a2, lsr #20 + orrmi a2, a2, #0xf000 + add ip, a4, v2 + mov ip, ip, asr #20 + orr a2, a2, ip, lsl #16 + str a2, [a1] + subs a3, a3, v1 + mov a2, a3, lsr #20 + orrmi a2, a2, #0xf000 + sub a4, a4, v2 + mov a4, a4, asr #20 + orr a2, a2, a4, lsl #16 + ldmfd sp!, {a3, a4} + str a2, [a1, #(16*7)] + + subs a2, a3, v3 + mov a2, a2, lsr #20 + orrmi a2, a2, #0xf000 + sub ip, a4, v4 + mov ip, ip, asr #20 + orr a2, a2, ip, lsl #16 + str a2, [a1, #(16*1)] + adds a3, a3, v3 + mov a2, a3, lsr #20 + orrmi a2, a2, #0xf000 + add a4, a4, v4 + mov a4, a4, asr #20 + orr a2, a2, a4, lsl #16 + ldmfd sp!, {a3, a4} + str a2, [a1, #(16*6)] + + adds a2, a3, v5 + mov a2, a2, lsr #20 + orrmi a2, a2, #0xf000 + add ip, a4, v6 + mov ip, ip, asr #20 + orr a2, a2, ip, lsl #16 + str a2, [a1, #(16*2)] + subs a3, a3, v5 + mov a2, a3, lsr #20 + orrmi a2, a2, #0xf000 + sub a4, a4, v6 + mov a4, a4, asr #20 + orr a2, a2, a4, lsl #16 + ldmfd sp!, {a3, a4} + str a2, [a1, #(16*5)] + + adds a2, a3, v7 + mov a2, a2, lsr #20 + orrmi a2, a2, #0xf000 + add ip, a4, fp + mov ip, ip, asr #20 + orr a2, a2, ip, lsl #16 + str a2, [a1, #(16*3)] + subs a3, a3, v7 + mov a2, a3, lsr #20 + orrmi a2, a2, #0xf000 + sub a4, a4, fp + mov a4, a4, asr #20 + orr a2, a2, a4, lsl #16 + str a2, [a1, #(16*4)] + + ldr pc, [sp], #4 + .endfunc + + .align + .func idct_col_put_armv5te +idct_col_put_armv5te: + str lr, [sp, #-4]! + + idct_col + + ldmfd sp!, {a3, a4} + ldr lr, [sp, #32] + add a2, a3, v1 + movs a2, a2, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add ip, a4, v2 + movs ip, ip, asr #20 + movmi ip, #0 + cmp ip, #255 + movgt ip, #255 + orr a2, a2, ip, lsl #8 + sub a3, a3, v1 + movs a3, a3, asr #20 + movmi a3, #0 + cmp a3, #255 + movgt a3, #255 + sub a4, a4, v2 + movs a4, a4, asr #20 + movmi a4, #0 + cmp a4, #255 + ldr v1, [sp, #28] + movgt a4, #255 + strh a2, [v1] + add a2, v1, #2 + str a2, [sp, #28] + orr a2, a3, a4, lsl #8 + rsb v2, lr, lr, lsl #3 + ldmfd sp!, {a3, a4} + strh a2, [v2, v1]! + + sub a2, a3, v3 + movs a2, a2, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + sub ip, a4, v4 + movs ip, ip, asr #20 + movmi ip, #0 + cmp ip, #255 + movgt ip, #255 + orr a2, a2, ip, lsl #8 + strh a2, [v1, lr]! + add a3, a3, v3 + movs a2, a3, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add a4, a4, v4 + movs a4, a4, asr #20 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a2, a4, lsl #8 + ldmfd sp!, {a3, a4} + strh a2, [v2, -lr]! + + add a2, a3, v5 + movs a2, a2, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add ip, a4, v6 + movs ip, ip, asr #20 + movmi ip, #0 + cmp ip, #255 + movgt ip, #255 + orr a2, a2, ip, lsl #8 + strh a2, [v1, lr]! + sub a3, a3, v5 + movs a2, a3, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + sub a4, a4, v6 + movs a4, a4, asr #20 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a2, a4, lsl #8 + ldmfd sp!, {a3, a4} + strh a2, [v2, -lr]! + + add a2, a3, v7 + movs a2, a2, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add ip, a4, fp + movs ip, ip, asr #20 + movmi ip, #0 + cmp ip, #255 + movgt ip, #255 + orr a2, a2, ip, lsl #8 + strh a2, [v1, lr] + sub a3, a3, v7 + movs a2, a3, asr #20 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + sub a4, a4, fp + movs a4, a4, asr #20 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a2, a4, lsl #8 + strh a2, [v2, -lr] + + ldr pc, [sp], #4 + .endfunc + + .align + .func idct_col_add_armv5te +idct_col_add_armv5te: + str lr, [sp, #-4]! + + idct_col + + ldr lr, [sp, #36] + + ldmfd sp!, {a3, a4} + ldrh ip, [lr] + add a2, a3, v1 + mov a2, a2, asr #20 + sub a3, a3, v1 + and v1, ip, #255 + adds a2, a2, v1 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add v1, a4, v2 + mov v1, v1, asr #20 + adds v1, v1, ip, lsr #8 + movmi v1, #0 + cmp v1, #255 + movgt v1, #255 + orr a2, a2, v1, lsl #8 + ldr v1, [sp, #32] + sub a4, a4, v2 + rsb v2, v1, v1, lsl #3 + ldrh ip, [v2, lr]! + strh a2, [lr] + mov a3, a3, asr #20 + and a2, ip, #255 + adds a3, a3, a2 + movmi a3, #0 + cmp a3, #255 + movgt a3, #255 + mov a4, a4, asr #20 + adds a4, a4, ip, lsr #8 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + add a2, lr, #2 + str a2, [sp, #28] + orr a2, a3, a4, lsl #8 + strh a2, [v2] + + ldmfd sp!, {a3, a4} + ldrh ip, [lr, v1]! + sub a2, a3, v3 + mov a2, a2, asr #20 + add a3, a3, v3 + and v3, ip, #255 + adds a2, a2, v3 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + sub v3, a4, v4 + mov v3, v3, asr #20 + adds v3, v3, ip, lsr #8 + movmi v3, #0 + cmp v3, #255 + movgt v3, #255 + orr a2, a2, v3, lsl #8 + add a4, a4, v4 + ldrh ip, [v2, -v1]! + strh a2, [lr] + mov a3, a3, asr #20 + and a2, ip, #255 + adds a3, a3, a2 + movmi a3, #0 + cmp a3, #255 + movgt a3, #255 + mov a4, a4, asr #20 + adds a4, a4, ip, lsr #8 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a3, a4, lsl #8 + strh a2, [v2] + + ldmfd sp!, {a3, a4} + ldrh ip, [lr, v1]! + add a2, a3, v5 + mov a2, a2, asr #20 + sub a3, a3, v5 + and v3, ip, #255 + adds a2, a2, v3 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add v3, a4, v6 + mov v3, v3, asr #20 + adds v3, v3, ip, lsr #8 + movmi v3, #0 + cmp v3, #255 + movgt v3, #255 + orr a2, a2, v3, lsl #8 + sub a4, a4, v6 + ldrh ip, [v2, -v1]! + strh a2, [lr] + mov a3, a3, asr #20 + and a2, ip, #255 + adds a3, a3, a2 + movmi a3, #0 + cmp a3, #255 + movgt a3, #255 + mov a4, a4, asr #20 + adds a4, a4, ip, lsr #8 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a3, a4, lsl #8 + strh a2, [v2] + + ldmfd sp!, {a3, a4} + ldrh ip, [lr, v1]! + add a2, a3, v7 + mov a2, a2, asr #20 + sub a3, a3, v7 + and v3, ip, #255 + adds a2, a2, v3 + movmi a2, #0 + cmp a2, #255 + movgt a2, #255 + add v3, a4, fp + mov v3, v3, asr #20 + adds v3, v3, ip, lsr #8 + movmi v3, #0 + cmp v3, #255 + movgt v3, #255 + orr a2, a2, v3, lsl #8 + sub a4, a4, fp + ldrh ip, [v2, -v1]! + strh a2, [lr] + mov a3, a3, asr #20 + and a2, ip, #255 + adds a3, a3, a2 + movmi a3, #0 + cmp a3, #255 + movgt a3, #255 + mov a4, a4, asr #20 + adds a4, a4, ip, lsr #8 + movmi a4, #0 + cmp a4, #255 + movgt a4, #255 + orr a2, a3, a4, lsl #8 + strh a2, [v2] + + ldr pc, [sp], #4 + .endfunc + + .align + .global simple_idct_armv5te + .func simple_idct_armv5te +simple_idct_armv5te: + stmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp, lr} + + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + + sub a1, a1, #(16*7) + + bl idct_col_armv5te + add a1, a1, #4 + bl idct_col_armv5te + add a1, a1, #4 + bl idct_col_armv5te + add a1, a1, #4 + bl idct_col_armv5te + + ldmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp, pc} + .endfunc + + .align + .global simple_idct_add_armv5te + .func simple_idct_add_armv5te +simple_idct_add_armv5te: + stmfd sp!, {a1, a2, v1, v2, v3, v4, v5, v6, v7, fp, lr} + + mov a1, a3 + + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + + sub a1, a1, #(16*7) + + bl idct_col_add_armv5te + add a1, a1, #4 + bl idct_col_add_armv5te + add a1, a1, #4 + bl idct_col_add_armv5te + add a1, a1, #4 + bl idct_col_add_armv5te + + add sp, sp, #8 + ldmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp, pc} + .endfunc + + .align + .global simple_idct_put_armv5te + .func simple_idct_put_armv5te +simple_idct_put_armv5te: + stmfd sp!, {a1, a2, v1, v2, v3, v4, v5, v6, v7, fp, lr} + + mov a1, a3 + + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + add a1, a1, #16 + bl idct_row_armv5te + + sub a1, a1, #(16*7) + + bl idct_col_put_armv5te + add a1, a1, #4 + bl idct_col_put_armv5te + add a1, a1, #4 + bl idct_col_put_armv5te + add a1, a1, #4 + bl idct_col_put_armv5te + + add sp, sp, #8 + ldmfd sp!, {v1, v2, v3, v4, v5, v6, v7, fp, pc} + .endfunc diff --git a/src/libffmpeg/libavcodec/bytestream.h b/src/libffmpeg/libavcodec/bytestream.h new file mode 100644 index 000000000..ae5438b49 --- /dev/null +++ b/src/libffmpeg/libavcodec/bytestream.h @@ -0,0 +1,89 @@ +/* + * Bytestream functions + * copyright (c) 2006 Baptiste Coudurier <baptiste.coudurier@free.fr> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef FFMPEG_BYTESTREAM_H +#define FFMPEG_BYTESTREAM_H + +static av_always_inline unsigned int bytestream_get_le32(uint8_t **b) +{ + (*b) += 4; + return AV_RL32(*b - 4); +} + +static av_always_inline unsigned int bytestream_get_le16(uint8_t **b) +{ + (*b) += 2; + return AV_RL16(*b - 2); +} + +static av_always_inline unsigned int bytestream_get_byte(uint8_t **b) +{ + (*b)++; + return (*b)[-1]; +} + +static av_always_inline unsigned int bytestream_get_buffer(uint8_t **b, uint8_t *dst, unsigned int size) +{ + memcpy(dst, *b, size); + (*b) += size; + return size; +} + +static av_always_inline void bytestream_put_be32(uint8_t **b, const unsigned int value) +{ + *(*b)++ = value >> 24; + *(*b)++ = value >> 16; + *(*b)++ = value >> 8; + *(*b)++ = value; +}; + +static av_always_inline void bytestream_put_be16(uint8_t **b, const unsigned int value) +{ + *(*b)++ = value >> 8; + *(*b)++ = value; +} + +static av_always_inline void bytestream_put_le32(uint8_t **b, const unsigned int value) +{ + *(*b)++ = value; + *(*b)++ = value >> 8; + *(*b)++ = value >> 16; + *(*b)++ = value >> 24; +} + +static av_always_inline void bytestream_put_le16(uint8_t **b, const unsigned int value) +{ + *(*b)++ = value; + *(*b)++ = value >> 8; +} + +static av_always_inline void bytestream_put_byte(uint8_t **b, const unsigned int value) +{ + *(*b)++ = value; +} + +static av_always_inline void bytestream_put_buffer(uint8_t **b, const uint8_t *src, unsigned int size) +{ + memcpy(*b, src, size); + (*b) += size; +} + +#endif /* FFMPEG_BYTESTREAM_H */ diff --git a/src/libffmpeg/libavcodec/eval.h b/src/libffmpeg/libavcodec/eval.h new file mode 100644 index 000000000..b52199cf4 --- /dev/null +++ b/src/libffmpeg/libavcodec/eval.h @@ -0,0 +1,84 @@ +/* + * simple arithmetic expression evaluator + * + * Copyright (c) 2002 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file eval.h + * eval header. + */ + +#ifndef AVCODEC_EVAL_H +#define AVCODEC_EVAL_H + +#if LIBAVCODEC_VERSION_INT < ((52<<16)+(0<<8)+0) +double ff_eval(char *s, double *const_value, const char **const_name, + double (**func1)(void *, double), const char **func1_name, + double (**func2)(void *, double, double), char **func2_name, + void *opaque); +#endif + +/** + * Parses and evaluates an expression. + * Note, this is significantly slower than ff_parse_eval() + * @param s expression as a zero terminated string for example "1+2^3+5*5+sin(2/3)" + * @param func1 NULL terminated array of function pointers for functions which take 1 argument + * @param func2 NULL terminated array of function pointers for functions which take 2 arguments + * @param const_name NULL terminated array of zero terminated strings of constant identifers for example {"PI", "E", 0} + * @param func1_name NULL terminated array of zero terminated strings of func1 identifers + * @param func2_name NULL terminated array of zero terminated strings of func2 identifers + * @param error pointer to a char* which is set to an error message if something goes wrong + * @param const_value a zero terminated array of values for the identifers from const_name + * @param opaque a pointer which will be passed to all functions from func1 and func2 + * @return the value of the expression + */ +double ff_eval2(char *s, double *const_value, const char **const_name, + double (**func1)(void *, double), const char **func1_name, + double (**func2)(void *, double, double), char **func2_name, + void *opaque, char **error); + +typedef struct ff_expr_s AVEvalExpr; + +/** + * Parses a expression. + * @param s expression as a zero terminated string for example "1+2^3+5*5+sin(2/3)" + * @param func1 NULL terminated array of function pointers for functions which take 1 argument + * @param func2 NULL terminated array of function pointers for functions which take 2 arguments + * @param const_name NULL terminated array of zero terminated strings of constant identifers for example {"PI", "E", 0} + * @param func1_name NULL terminated array of zero terminated strings of func1 identifers + * @param func2_name NULL terminated array of zero terminated strings of func2 identifers + * @param error pointer to a char* which is set to an error message if something goes wrong + * @return AVEvalExpr which must be freed with ff_eval_free by the user when its not needed anymore + * NULL if anything went wrong + */ +AVEvalExpr * ff_parse(char *s, const char **const_name, + double (**func1)(void *, double), const char **func1_name, + double (**func2)(void *, double, double), char **func2_name, + char **error); +/** + * Evaluates a previously parsed expression. + * @param const_value a zero terminated array of values for the identifers from ff_parse const_name + * @param opaque a pointer which will be passed to all functions from func1 and func2 + * @return the value of the expression + */ +double ff_parse_eval(AVEvalExpr * e, double *const_value, void *opaque); +void ff_eval_free(AVEvalExpr * e); + +#endif /* AVCODEC_EVAL_H */ diff --git a/src/libffmpeg/libavcodec/i386/.hgignore b/src/libffmpeg/libavcodec/i386/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/i386/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/i386/fft_3dn.c b/src/libffmpeg/libavcodec/i386/fft_3dn.c new file mode 100644 index 000000000..8087f1932 --- /dev/null +++ b/src/libffmpeg/libavcodec/i386/fft_3dn.c @@ -0,0 +1,125 @@ +/* + * FFT/MDCT transform with 3DNow! optimizations + * Copyright (c) 2006 Zuxy MENG Jie, Loren Merritt + * Based on fft_sse.c copyright (c) 2002 Fabrice Bellard. + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "../dsputil.h" + +static const int p1m1[2] __attribute__((aligned(8))) = + { 0, 1 << 31 }; + +static const int m1p1[2] __attribute__((aligned(8))) = + { 1 << 31, 0 }; + +void ff_fft_calc_3dn(FFTContext *s, FFTComplex *z) +{ + int ln = s->nbits; + long i, j; + long nblocks, nloops; + FFTComplex *p, *cptr; + + asm volatile( + /* FEMMS is not a must here but recommended by AMD */ + "femms \n\t" + "movq %0, %%mm7 \n\t" + ::"m"(*(s->inverse ? m1p1 : p1m1)) + ); + + i = 8 << ln; + asm volatile( + "1: \n\t" + "sub $32, %0 \n\t" + "movq (%0,%1), %%mm0 \n\t" + "movq 16(%0,%1), %%mm1 \n\t" + "movq 8(%0,%1), %%mm2 \n\t" + "movq 24(%0,%1), %%mm3 \n\t" + "movq %%mm0, %%mm4 \n\t" + "movq %%mm1, %%mm5 \n\t" + "pfadd %%mm2, %%mm0 \n\t" + "pfadd %%mm3, %%mm1 \n\t" + "pfsub %%mm2, %%mm4 \n\t" + "pfsub %%mm3, %%mm5 \n\t" + "movq %%mm0, %%mm2 \n\t" + "punpckldq %%mm5, %%mm6 \n\t" + "punpckhdq %%mm6, %%mm5 \n\t" + "movq %%mm4, %%mm3 \n\t" + "pxor %%mm7, %%mm5 \n\t" + "pfadd %%mm1, %%mm0 \n\t" + "pfadd %%mm5, %%mm4 \n\t" + "pfsub %%mm1, %%mm2 \n\t" + "pfsub %%mm5, %%mm3 \n\t" + "movq %%mm0, (%0,%1) \n\t" + "movq %%mm4, 8(%0,%1) \n\t" + "movq %%mm2, 16(%0,%1) \n\t" + "movq %%mm3, 24(%0,%1) \n\t" + "jg 1b \n\t" + :"+r"(i) + :"r"(z) + ); + /* pass 2 .. ln-1 */ + + nblocks = 1 << (ln-3); + nloops = 1 << 2; + cptr = s->exptab1; + do { + p = z; + j = nblocks; + do { + i = nloops*8; + asm volatile( + "1: \n\t" + "sub $16, %0 \n\t" + "movq (%1,%0), %%mm0 \n\t" + "movq 8(%1,%0), %%mm1 \n\t" + "movq (%2,%0), %%mm2 \n\t" + "movq 8(%2,%0), %%mm3 \n\t" + "movq %%mm2, %%mm4 \n\t" + "movq %%mm3, %%mm5 \n\t" + "punpckldq %%mm2, %%mm2 \n\t" + "punpckldq %%mm3, %%mm3 \n\t" + "punpckhdq %%mm4, %%mm4 \n\t" + "punpckhdq %%mm5, %%mm5 \n\t" + "pfmul (%3,%0,2), %%mm2 \n\t" // cre*re cim*re + "pfmul 8(%3,%0,2), %%mm3 \n\t" + "pfmul 16(%3,%0,2), %%mm4 \n\t" // -cim*im cre*im + "pfmul 24(%3,%0,2), %%mm5 \n\t" + "pfadd %%mm2, %%mm4 \n\t" // cre*re-cim*im cim*re+cre*im + "pfadd %%mm3, %%mm5 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm1, %%mm3 \n\t" + "pfadd %%mm4, %%mm0 \n\t" + "pfadd %%mm5, %%mm1 \n\t" + "pfsub %%mm4, %%mm2 \n\t" + "pfsub %%mm5, %%mm3 \n\t" + "movq %%mm0, (%1,%0) \n\t" + "movq %%mm1, 8(%1,%0) \n\t" + "movq %%mm2, (%2,%0) \n\t" + "movq %%mm3, 8(%2,%0) \n\t" + "jg 1b \n\t" + :"+r"(i) + :"r"(p), "r"(p + nloops), "r"(cptr) + ); + p += nloops*2; + } while (--j); + cptr += nloops*2; + nblocks >>= 1; + nloops <<= 1; + } while (nblocks != 0); + asm volatile("femms"); +} diff --git a/src/libffmpeg/libavcodec/i386/fft_3dn2.c b/src/libffmpeg/libavcodec/i386/fft_3dn2.c new file mode 100644 index 000000000..a4fe5f0b6 --- /dev/null +++ b/src/libffmpeg/libavcodec/i386/fft_3dn2.c @@ -0,0 +1,210 @@ +/* + * FFT/MDCT transform with Extended 3DNow! optimizations + * Copyright (c) 2006 Zuxy MENG Jie, Loren Merritt + * Based on fft_sse.c copyright (c) 2002 Fabrice Bellard. + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#include "../dsputil.h" + +static const int p1m1[2] __attribute__((aligned(8))) = + { 0, 1 << 31 }; + +static const int m1p1[2] __attribute__((aligned(8))) = + { 1 << 31, 0 }; + +void ff_fft_calc_3dn2(FFTContext *s, FFTComplex *z) +{ + int ln = s->nbits; + long i, j; + long nblocks, nloops; + FFTComplex *p, *cptr; + + asm volatile( + /* FEMMS is not a must here but recommended by AMD */ + "femms \n\t" + "movq %0, %%mm7 \n\t" + ::"m"(*(s->inverse ? m1p1 : p1m1)) + ); + + i = 8 << ln; + asm volatile( + "1: \n\t" + "sub $32, %0 \n\t" + "movq (%0,%1), %%mm0 \n\t" + "movq 16(%0,%1), %%mm1 \n\t" + "movq 8(%0,%1), %%mm2 \n\t" + "movq 24(%0,%1), %%mm3 \n\t" + "movq %%mm0, %%mm4 \n\t" + "movq %%mm1, %%mm5 \n\t" + "pfadd %%mm2, %%mm0 \n\t" + "pfadd %%mm3, %%mm1 \n\t" + "pfsub %%mm2, %%mm4 \n\t" + "pfsub %%mm3, %%mm5 \n\t" + "movq %%mm0, %%mm2 \n\t" + "pswapd %%mm5, %%mm5 \n\t" + "movq %%mm4, %%mm3 \n\t" + "pxor %%mm7, %%mm5 \n\t" + "pfadd %%mm1, %%mm0 \n\t" + "pfadd %%mm5, %%mm4 \n\t" + "pfsub %%mm1, %%mm2 \n\t" + "pfsub %%mm5, %%mm3 \n\t" + "movq %%mm0, (%0,%1) \n\t" + "movq %%mm4, 8(%0,%1) \n\t" + "movq %%mm2, 16(%0,%1) \n\t" + "movq %%mm3, 24(%0,%1) \n\t" + "jg 1b \n\t" + :"+r"(i) + :"r"(z) + ); + /* pass 2 .. ln-1 */ + + nblocks = 1 << (ln-3); + nloops = 1 << 2; + cptr = s->exptab1; + do { + p = z; + j = nblocks; + do { + i = nloops*8; + asm volatile( + "1: \n\t" + "sub $16, %0 \n\t" + "movq (%1,%0), %%mm0 \n\t" + "movq 8(%1,%0), %%mm1 \n\t" + "movq (%2,%0), %%mm2 \n\t" + "movq 8(%2,%0), %%mm3 \n\t" + "movq (%3,%0,2), %%mm4 \n\t" + "movq 8(%3,%0,2), %%mm5 \n\t" + "pswapd %%mm4, %%mm6 \n\t" // no need for cptr[2] & cptr[3] + "pswapd %%mm5, %%mm7 \n\t" + "pfmul %%mm2, %%mm4 \n\t" // cre*re cim*im + "pfmul %%mm3, %%mm5 \n\t" + "pfmul %%mm2, %%mm6 \n\t" // cim*re cre*im + "pfmul %%mm3, %%mm7 \n\t" + "pfpnacc %%mm6, %%mm4 \n\t" // cre*re-cim*im cim*re+cre*im + "pfpnacc %%mm7, %%mm5 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm1, %%mm3 \n\t" + "pfadd %%mm4, %%mm0 \n\t" + "pfadd %%mm5, %%mm1 \n\t" + "pfsub %%mm4, %%mm2 \n\t" + "pfsub %%mm5, %%mm3 \n\t" + "movq %%mm0, (%1,%0) \n\t" + "movq %%mm1, 8(%1,%0) \n\t" + "movq %%mm2, (%2,%0) \n\t" + "movq %%mm3, 8(%2,%0) \n\t" + "jg 1b \n\t" + :"+r"(i) + :"r"(p), "r"(p + nloops), "r"(cptr) + ); + p += nloops*2; + } while (--j); + cptr += nloops*2; + nblocks >>= 1; + nloops <<= 1; + } while (nblocks != 0); + asm volatile("femms"); +} + +void ff_imdct_calc_3dn2(MDCTContext *s, FFTSample *output, + const FFTSample *input, FFTSample *tmp) +{ + long k, n8, n4, n2, n; + const uint16_t *revtab = s->fft.revtab; + const FFTSample *tcos = s->tcos; + const FFTSample *tsin = s->tsin; + const FFTSample *in1, *in2; + FFTComplex *z = (FFTComplex *)tmp; + + n = 1 << s->nbits; + n2 = n >> 1; + n4 = n >> 2; + n8 = n >> 3; + + /* pre rotation */ + in1 = input; + in2 = input + n2 - 1; + for(k = 0; k < n4; k++) { + // FIXME a single block is faster, but gcc 2.95 and 3.4.x on 32bit can't compile it + asm volatile( + "movd %0, %%mm0 \n\t" + "movd %2, %%mm1 \n\t" + "punpckldq %1, %%mm0 \n\t" + "punpckldq %3, %%mm1 \n\t" + "movq %%mm0, %%mm2 \n\t" + "pfmul %%mm1, %%mm0 \n\t" + "pswapd %%mm1, %%mm1 \n\t" + "pfmul %%mm1, %%mm2 \n\t" + "pfpnacc %%mm2, %%mm0 \n\t" + ::"m"(in2[-2*k]), "m"(in1[2*k]), + "m"(tcos[k]), "m"(tsin[k]) + ); + asm volatile( + "movq %%mm0, %0 \n\t" + :"=m"(z[revtab[k]]) + ); + } + + ff_fft_calc(&s->fft, z); + + /* post rotation + reordering */ + for(k = 0; k < n4; k++) { + asm volatile( + "movq %0, %%mm0 \n\t" + "movd %1, %%mm1 \n\t" + "punpckldq %2, %%mm1 \n\t" + "movq %%mm0, %%mm2 \n\t" + "pfmul %%mm1, %%mm0 \n\t" + "pswapd %%mm1, %%mm1 \n\t" + "pfmul %%mm1, %%mm2 \n\t" + "pfpnacc %%mm2, %%mm0 \n\t" + "movq %%mm0, %0 \n\t" + :"+m"(z[k]) + :"m"(tcos[k]), "m"(tsin[k]) + ); + } + + k = n-8; + asm volatile("movd %0, %%mm7" ::"r"(1<<31)); + asm volatile( + "1: \n\t" + "movq (%4,%0), %%mm0 \n\t" // z[n8+k] + "neg %0 \n\t" + "pswapd -8(%4,%0), %%mm1 \n\t" // z[n8-1-k] + "movq %%mm0, %%mm2 \n\t" + "pxor %%mm7, %%mm2 \n\t" + "punpckldq %%mm1, %%mm2 \n\t" + "pswapd %%mm2, %%mm3 \n\t" + "punpckhdq %%mm1, %%mm0 \n\t" + "pswapd %%mm0, %%mm4 \n\t" + "pxor %%mm7, %%mm0 \n\t" + "pxor %%mm7, %%mm4 \n\t" + "movq %%mm3, -8(%3,%0) \n\t" // output[n-2-2*k] = { z[n8-1-k].im, -z[n8+k].re } + "movq %%mm4, -8(%2,%0) \n\t" // output[n2-2-2*k]= { -z[n8-1-k].re, z[n8+k].im } + "neg %0 \n\t" + "movq %%mm0, (%1,%0) \n\t" // output[2*k] = { -z[n8+k].im, z[n8-1-k].re } + "movq %%mm2, (%2,%0) \n\t" // output[n2+2*k] = { -z[n8+k].re, z[n8-1-k].im } + "sub $8, %0 \n\t" + "jge 1b \n\t" + :"+r"(k) + :"r"(output), "r"(output+n2), "r"(output+n), "r"(z+n8) + :"memory" + ); + asm volatile("femms"); +} + diff --git a/src/libffmpeg/libavcodec/i386/mathops.h b/src/libffmpeg/libavcodec/i386/mathops.h new file mode 100644 index 000000000..3553a4025 --- /dev/null +++ b/src/libffmpeg/libavcodec/i386/mathops.h @@ -0,0 +1,41 @@ +/* + * simple math operations + * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifdef FRAC_BITS +# define MULL(ra, rb) \ + ({ int rt, dummy; asm (\ + "imull %3 \n\t"\ + "shrdl %4, %%edx, %%eax \n\t"\ + : "=a"(rt), "=d"(dummy)\ + : "a" (ra), "rm" (rb), "i"(FRAC_BITS));\ + rt; }) +#endif + +#define MULH(ra, rb) \ + ({ int rt, dummy;\ + asm ("imull %3\n\t" : "=d"(rt), "=a"(dummy): "a" (ra), "rm" (rb));\ + rt; }) + +#define MUL64(ra, rb) \ + ({ int64_t rt;\ + asm ("imull %2\n\t" : "=A"(rt) : "a" (ra), "g" (rb));\ + rt; }) + diff --git a/src/libffmpeg/libavcodec/libpostproc/.hgignore b/src/libffmpeg/libavcodec/libpostproc/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/libpostproc/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/mathops.h b/src/libffmpeg/libavcodec/mathops.h new file mode 100644 index 000000000..c6ec70597 --- /dev/null +++ b/src/libffmpeg/libavcodec/mathops.h @@ -0,0 +1,69 @@ +/* + * simple math operations + * Copyright (c) 2001, 2002 Fabrice Bellard. + * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#ifndef MATHOPS_H +#define MATHOPS_H + +#ifdef ARCH_X86_32 + +#include "i386/mathops.h" + +#elif defined(ARCH_ARMV4L) + +#include "armv4l/mathops.h" + +#elif defined(ARCH_PPC) + +#include "ppc/mathops.h" + +#endif + +/* generic implementation */ + +#ifndef MULL +# define MULL(a,b) (((int64_t)(a) * (int64_t)(b)) >> FRAC_BITS) +#endif + +#ifndef MULH +//gcc 3.4 creates an incredibly bloated mess out of this +//# define MULH(a,b) (((int64_t)(a) * (int64_t)(b))>>32) + +static av_always_inline int MULH(int a, int b){ + return ((int64_t)(a) * (int64_t)(b))>>32; +} +#endif + +#ifndef MUL64 +# define MUL64(a,b) ((int64_t)(a) * (int64_t)(b)) +#endif + +/* signed 16x16 -> 32 multiply add accumulate */ +#ifndef MAC16 +# define MAC16(rt, ra, rb) rt += (ra) * (rb) +#endif + +/* signed 16x16 -> 32 multiply */ +#ifndef MUL16 +# define MUL16(ra, rb) ((ra) * (rb)) +#endif + +#endif //MATHOPS_H + diff --git a/src/libffmpeg/libavcodec/mlib/.hgignore b/src/libffmpeg/libavcodec/mlib/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/mlib/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/opt.c b/src/libffmpeg/libavcodec/opt.c new file mode 100644 index 000000000..70babd587 --- /dev/null +++ b/src/libffmpeg/libavcodec/opt.c @@ -0,0 +1,381 @@ +/* + * AVOptions + * Copyright (c) 2005 Michael Niedermayer <michaelni@gmx.at> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + * + */ + +/** + * @file opt.c + * AVOptions + * @author Michael Niedermayer <michaelni@gmx.at> + */ + +#include "avcodec.h" +#include "opt.h" +#include "eval.h" + +//FIXME order them and do a bin search +static const AVOption *find_opt(void *v, const char *name, const char *unit){ + AVClass *c= *(AVClass**)v; //FIXME silly way of storing AVClass + const AVOption *o= c->option; + + for(;o && o->name; o++){ + if(!strcmp(o->name, name) && (!unit || !strcmp(o->unit, unit)) ) + return o; + } + return NULL; +} + +const AVOption *av_next_option(void *obj, const AVOption *last){ + if(last && last[1].name) return ++last; + else if(last) return NULL; + else return (*(AVClass**)obj)->option; +} + +static const AVOption *av_set_number(void *obj, const char *name, double num, int den, int64_t intnum){ + const AVOption *o= find_opt(obj, name, NULL); + void *dst; + if(!o || o->offset<=0) + return NULL; + + if(o->max*den < num*intnum || o->min*den > num*intnum) { + av_log(NULL, AV_LOG_ERROR, "Value %lf for parameter '%s' out of range.\n", num, name); + return NULL; + } + + dst= ((uint8_t*)obj) + o->offset; + + switch(o->type){ + case FF_OPT_TYPE_FLAGS: + case FF_OPT_TYPE_INT: *(int *)dst= lrintf(num/den)*intnum; break; + case FF_OPT_TYPE_INT64: *(int64_t *)dst= lrintf(num/den)*intnum; break; + case FF_OPT_TYPE_FLOAT: *(float *)dst= num*intnum/den; break; + case FF_OPT_TYPE_DOUBLE:*(double *)dst= num*intnum/den; break; + case FF_OPT_TYPE_RATIONAL: + if((int)num == num) *(AVRational*)dst= (AVRational){num*intnum, den}; + else *(AVRational*)dst= av_d2q(num*intnum/den, 1<<24); + default: + return NULL; + } + return o; +} + +static const AVOption *set_all_opt(void *v, const char *unit, double d){ + AVClass *c= *(AVClass**)v; //FIXME silly way of storing AVClass + const AVOption *o= c->option; + const AVOption *ret=NULL; + + for(;o && o->name; o++){ + if(o->type != FF_OPT_TYPE_CONST && o->unit && !strcmp(o->unit, unit)){ + double tmp= d; + if(o->type == FF_OPT_TYPE_FLAGS) + tmp= av_get_int(v, o->name, NULL) | (int64_t)d; + + av_set_number(v, o->name, tmp, 1, 1); + ret= o; + } + } + return ret; +} + +static double const_values[]={ + M_PI, + M_E, + FF_QP2LAMBDA, + 0 +}; + +static const char *const_names[]={ + "PI", + "E", + "QP2LAMBDA", + 0 +}; + +const AVOption *av_set_string(void *obj, const char *name, const char *val){ + const AVOption *o= find_opt(obj, name, NULL); + if(o && o->offset==0 && o->type == FF_OPT_TYPE_CONST && o->unit){ + return set_all_opt(obj, o->unit, o->default_val); + } + if(!o || !val || o->offset<=0) + return NULL; + if(o->type != FF_OPT_TYPE_STRING){ + for(;;){ + int i; + char buf[256]; + int cmd=0; + double d; + char *error = NULL; + + if(*val == '+' || *val == '-') + cmd= *(val++); + + for(i=0; i<sizeof(buf)-1 && val[i] && val[i]!='+' && val[i]!='-'; i++) + buf[i]= val[i]; + buf[i]=0; + val+= i; + + d = ff_eval2(buf, const_values, const_names, NULL, NULL, NULL, NULL, NULL, &error); + if(isnan(d)) { + const AVOption *o_named= find_opt(obj, buf, o->unit); + if(o_named && o_named->type == FF_OPT_TYPE_CONST) + d= o_named->default_val; + else if(!strcmp(buf, "default")) d= o->default_val; + else if(!strcmp(buf, "max" )) d= o->max; + else if(!strcmp(buf, "min" )) d= o->min; + else { + if (!error) + av_log(NULL, AV_LOG_ERROR, "Unable to parse option value \"%s\": %s\n", val, error); + return NULL; + } + } + if(o->type == FF_OPT_TYPE_FLAGS){ + if (cmd=='+') d= av_get_int(obj, name, NULL) | (int64_t)d; + else if(cmd=='-') d= av_get_int(obj, name, NULL) &~(int64_t)d; + }else if(cmd=='-') + d= -d; + + av_set_number(obj, name, d, 1, 1); + if(!*val) + return o; + } + return NULL; + } + + memcpy(((uint8_t*)obj) + o->offset, val, sizeof(val)); + return o; +} + +const AVOption *av_set_double(void *obj, const char *name, double n){ + return av_set_number(obj, name, n, 1, 1); +} + +const AVOption *av_set_q(void *obj, const char *name, AVRational n){ + return av_set_number(obj, name, n.num, n.den, 1); +} + +const AVOption *av_set_int(void *obj, const char *name, int64_t n){ + return av_set_number(obj, name, 1, 1, n); +} + +/** + * + * @param buf a buffer which is used for returning non string values as strings, can be NULL + * @param buf_len allocated length in bytes of buf + */ +const char *av_get_string(void *obj, const char *name, const AVOption **o_out, char *buf, int buf_len){ + const AVOption *o= find_opt(obj, name, NULL); + void *dst; + if(!o || o->offset<=0) + return NULL; + if(o->type != FF_OPT_TYPE_STRING && (!buf || !buf_len)) + return NULL; + + dst= ((uint8_t*)obj) + o->offset; + if(o_out) *o_out= o; + + if(o->type == FF_OPT_TYPE_STRING) + return dst; + + switch(o->type){ + case FF_OPT_TYPE_FLAGS: snprintf(buf, buf_len, "0x%08X",*(int *)dst);break; + case FF_OPT_TYPE_INT: snprintf(buf, buf_len, "%d" , *(int *)dst);break; + case FF_OPT_TYPE_INT64: snprintf(buf, buf_len, "%"PRId64, *(int64_t*)dst);break; + case FF_OPT_TYPE_FLOAT: snprintf(buf, buf_len, "%f" , *(float *)dst);break; + case FF_OPT_TYPE_DOUBLE: snprintf(buf, buf_len, "%f" , *(double *)dst);break; + case FF_OPT_TYPE_RATIONAL: snprintf(buf, buf_len, "%d/%d", ((AVRational*)dst)->num, ((AVRational*)dst)->den);break; + default: return NULL; + } + return buf; +} + +static int av_get_number(void *obj, const char *name, const AVOption **o_out, double *num, int *den, int64_t *intnum){ + const AVOption *o= find_opt(obj, name, NULL); + void *dst; + if(!o || o->offset<=0) + goto error; + + dst= ((uint8_t*)obj) + o->offset; + + if(o_out) *o_out= o; + + switch(o->type){ + case FF_OPT_TYPE_FLAGS: + case FF_OPT_TYPE_INT: *intnum= *(int *)dst;return 0; + case FF_OPT_TYPE_INT64: *intnum= *(int64_t*)dst;return 0; + case FF_OPT_TYPE_FLOAT: *num= *(float *)dst;return 0; + case FF_OPT_TYPE_DOUBLE: *num= *(double *)dst;return 0; + case FF_OPT_TYPE_RATIONAL: *intnum= ((AVRational*)dst)->num; + *den = ((AVRational*)dst)->den; + return 0; + } +error: + *den=*intnum=0; + return -1; +} + +double av_get_double(void *obj, const char *name, const AVOption **o_out){ + int64_t intnum=1; + double num=1; + int den=1; + + av_get_number(obj, name, o_out, &num, &den, &intnum); + return num*intnum/den; +} + +AVRational av_get_q(void *obj, const char *name, const AVOption **o_out){ + int64_t intnum=1; + double num=1; + int den=1; + + av_get_number(obj, name, o_out, &num, &den, &intnum); + if(num == 1.0 && (int)intnum == intnum) + return (AVRational){intnum, den}; + else + return av_d2q(num*intnum/den, 1<<24); +} + +int64_t av_get_int(void *obj, const char *name, const AVOption **o_out){ + int64_t intnum=1; + double num=1; + int den=1; + + av_get_number(obj, name, o_out, &num, &den, &intnum); + return num*intnum/den; +} + +static void opt_list(void *obj, void *av_log_obj, const char *unit) +{ + const AVOption *opt=NULL; + + while((opt= av_next_option(obj, opt))){ + if(!(opt->flags & (AV_OPT_FLAG_ENCODING_PARAM|AV_OPT_FLAG_DECODING_PARAM))) + continue; + + /* Don't print CONST's on level one. + * Don't print anything but CONST's on level two. + * Only print items from the requested unit. + */ + if (!unit && opt->type==FF_OPT_TYPE_CONST) + continue; + else if (unit && opt->type!=FF_OPT_TYPE_CONST) + continue; + else if (unit && opt->type==FF_OPT_TYPE_CONST && strcmp(unit, opt->unit)) + continue; + else if (unit && opt->type == FF_OPT_TYPE_CONST) + av_log(av_log_obj, AV_LOG_INFO, " %-15s ", opt->name); + else + av_log(av_log_obj, AV_LOG_INFO, "-%-17s ", opt->name); + + switch( opt->type ) + { + case FF_OPT_TYPE_FLAGS: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<flags>" ); + break; + case FF_OPT_TYPE_INT: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<int>" ); + break; + case FF_OPT_TYPE_INT64: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<int64>" ); + break; + case FF_OPT_TYPE_DOUBLE: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<double>" ); + break; + case FF_OPT_TYPE_FLOAT: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<float>" ); + break; + case FF_OPT_TYPE_STRING: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<string>" ); + break; + case FF_OPT_TYPE_RATIONAL: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "<rational>" ); + break; + case FF_OPT_TYPE_CONST: + default: + av_log( av_log_obj, AV_LOG_INFO, "%-7s ", "" ); + break; + } + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_ENCODING_PARAM) ? 'E' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_DECODING_PARAM) ? 'D' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_VIDEO_PARAM ) ? 'V' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_AUDIO_PARAM ) ? 'A' : '.'); + av_log(av_log_obj, AV_LOG_INFO, "%c", (opt->flags & AV_OPT_FLAG_SUBTITLE_PARAM) ? 'S' : '.'); + + if(opt->help) + av_log(av_log_obj, AV_LOG_INFO, " %s", opt->help); + av_log(av_log_obj, AV_LOG_INFO, "\n"); + if (opt->unit && opt->type != FF_OPT_TYPE_CONST) { + opt_list(obj, av_log_obj, opt->unit); + } + } +} + +int av_opt_show(void *obj, void *av_log_obj){ + if(!obj) + return -1; + + av_log(av_log_obj, AV_LOG_INFO, "%s AVOptions:\n", (*(AVClass**)obj)->class_name); + + opt_list(obj, av_log_obj, NULL); + + return 0; +} + +/** Set the values of the AVCodecContext or AVFormatContext structure. + * They are set to the defaults specified in the according AVOption options + * array default_val field. + * + * @param s AVCodecContext or AVFormatContext for which the defaults will be set + */ +void av_opt_set_defaults(void *s) +{ + const AVOption *opt = NULL; + while ((opt = av_next_option(s, opt)) != NULL) { + switch(opt->type) { + case FF_OPT_TYPE_CONST: + /* Nothing to be done here */ + break; + case FF_OPT_TYPE_FLAGS: + case FF_OPT_TYPE_INT: { + int val; + val = opt->default_val; + av_set_int(s, opt->name, val); + } + break; + case FF_OPT_TYPE_FLOAT: { + double val; + val = opt->default_val; + av_set_double(s, opt->name, val); + } + break; + case FF_OPT_TYPE_RATIONAL: { + AVRational val; + val = av_d2q(opt->default_val, INT_MAX); + av_set_q(s, opt->name, val); + } + break; + case FF_OPT_TYPE_STRING: + /* Cannot set default for string as default_val is of type * double */ + break; + default: + av_log(s, AV_LOG_DEBUG, "AVOption type %d of option %s not implemented yet\n", opt->type, opt->name); + } + } +} + diff --git a/src/libffmpeg/libavcodec/parser.h b/src/libffmpeg/libavcodec/parser.h new file mode 100644 index 000000000..3496b341f --- /dev/null +++ b/src/libffmpeg/libavcodec/parser.h @@ -0,0 +1,63 @@ +/* + * AVCodecParser prototypes and definitions + * Copyright (c) 2003 Fabrice Bellard. + * Copyright (c) 2003 Michael Niedermayer. + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef FFMPEG_PARSER_H +#define FFMPEG_PARSER_H + +typedef struct ParseContext{ + uint8_t *buffer; + int index; + int last_index; + unsigned int buffer_size; + uint32_t state; ///< contains the last few bytes in MSB order + int frame_start_found; + int overread; ///< the number of bytes which where irreversibly read from the next frame + int overread_index; ///< the index into ParseContext.buffer of the overreaded bytes +} ParseContext; + +struct MpegEncContext; + +typedef struct ParseContext1{ + ParseContext pc; +/* XXX/FIXME PC1 vs. PC */ + /* MPEG2 specific */ + AVRational frame_rate; + int progressive_sequence; + int width, height; + + /* XXX: suppress that, needed by MPEG4 */ + struct MpegEncContext *enc; + int first_picture; +} ParseContext1; + +#define END_NOT_FOUND (-100) + +int ff_combine_frame(ParseContext *pc, int next, uint8_t **buf, int *buf_size); +int ff_mpeg4video_split(AVCodecContext *avctx, const uint8_t *buf, + int buf_size); +void ff_parse_close(AVCodecParserContext *s); +void ff_parse1_close(AVCodecParserContext *s); + +/* h263dec.c */ +int ff_mpeg4_find_frame_end(ParseContext *pc, const uint8_t *buf, int buf_size); + +#endif /* !FFMPEG_PARSER_H */ diff --git a/src/libffmpeg/libavcodec/ppc/.hgignore b/src/libffmpeg/libavcodec/ppc/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/ppc/float_altivec.c b/src/libffmpeg/libavcodec/ppc/float_altivec.c new file mode 100644 index 000000000..22c2de61a --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/float_altivec.c @@ -0,0 +1,193 @@ +/* + * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "../dsputil.h" + +#include "gcc_fixes.h" + +#include "dsputil_altivec.h" + +static void vector_fmul_altivec(float *dst, const float *src, int len) +{ + int i; + vector float d0, d1, s, zero = (vector float)vec_splat_u32(0); + for(i=0; i<len-7; i+=8) { + d0 = vec_ld(0, dst+i); + s = vec_ld(0, src+i); + d1 = vec_ld(16, dst+i); + d0 = vec_madd(d0, s, zero); + d1 = vec_madd(d1, vec_ld(16,src+i), zero); + vec_st(d0, 0, dst+i); + vec_st(d1, 16, dst+i); + } +} + +static void vector_fmul_reverse_altivec(float *dst, const float *src0, + const float *src1, int len) +{ + int i; + vector float d, s0, s1, h0, l0, + s2, s3, zero = (vector float)vec_splat_u32(0); + src1 += len-4; + for(i=0; i<len-7; i+=8) { + s1 = vec_ld(0, src1-i); // [a,b,c,d] + s0 = vec_ld(0, src0+i); + l0 = vec_mergel(s1, s1); // [c,c,d,d] + s3 = vec_ld(-16, src1-i); + h0 = vec_mergeh(s1, s1); // [a,a,b,b] + s2 = vec_ld(16, src0+i); + s1 = vec_mergeh(vec_mergel(l0,h0), // [d,b,d,b] + vec_mergeh(l0,h0)); // [c,a,c,a] + // [d,c,b,a] + l0 = vec_mergel(s3, s3); + d = vec_madd(s0, s1, zero); + h0 = vec_mergeh(s3, s3); + vec_st(d, 0, dst+i); + s3 = vec_mergeh(vec_mergel(l0,h0), + vec_mergeh(l0,h0)); + d = vec_madd(s2, s3, zero); + vec_st(d, 16, dst+i); + } +} + +static void vector_fmul_add_add_altivec(float *dst, const float *src0, + const float *src1, const float *src2, + int src3, int len, int step) +{ + int i; + vector float d, s0, s1, s2, t0, t1, edges; + vector unsigned char align = vec_lvsr(0,dst), + mask = vec_lvsl(0, dst); + +#if 0 //FIXME: there is still something wrong + if (step == 2) { + int y; + vector float d0, d1, s3, t2; + vector unsigned int sel = + vec_mergeh(vec_splat_u32(-1), vec_splat_u32(0)); + t1 = vec_ld(16, dst); + for (i=0,y=0; i<len-3; i+=4,y+=8) { + + s0 = vec_ld(0,src0+i); + s1 = vec_ld(0,src1+i); + s2 = vec_ld(0,src2+i); + +// t0 = vec_ld(0, dst+y); //[x x x|a] +// t1 = vec_ld(16, dst+y); //[b c d|e] + t2 = vec_ld(31, dst+y); //[f g h|x] + + d = vec_madd(s0,s1,s2); // [A B C D] + + // [A A B B] + + // [C C D D] + + d0 = vec_perm(t0, t1, mask); // [a b c d] + + d0 = vec_sel(vec_mergeh(d, d), d0, sel); // [A b B d] + + edges = vec_perm(t1, t0, mask); + + t0 = vec_perm(edges, d0, align); // [x x x|A] + + t1 = vec_perm(d0, edges, align); // [b B d|e] + + vec_stl(t0, 0, dst+y); + + d1 = vec_perm(t1, t2, mask); // [e f g h] + + d1 = vec_sel(vec_mergel(d, d), d1, sel); // [C f D h] + + edges = vec_perm(t2, t1, mask); + + t1 = vec_perm(edges, d1, align); // [b B d|C] + + t2 = vec_perm(d1, edges, align); // [f D h|x] + + vec_stl(t1, 16, dst+y); + + t0 = t1; + + vec_stl(t2, 31, dst+y); + + t1 = t2; + } + } else + #endif + if (step == 1 && src3 == 0) + for (i=0; i<len-3; i+=4) { + t0 = vec_ld(0, dst+i); + t1 = vec_ld(15, dst+i); + s0 = vec_ld(0, src0+i); + s1 = vec_ld(0, src1+i); + s2 = vec_ld(0, src2+i); + edges = vec_perm(t1 ,t0, mask); + d = vec_madd(s0,s1,s2); + t1 = vec_perm(d, edges, align); + t0 = vec_perm(edges, d, align); + vec_st(t1, 15, dst+i); + vec_st(t0, 0, dst+i); + } + else + ff_vector_fmul_add_add_c(dst, src0, src1, src2, src3, len, step); +} + +void float_to_int16_altivec(int16_t *dst, const float *src, int len) +{ + int i; + vector float s0, s1; + vector signed int t0, t1; + vector signed short d0, d1, d; + vector unsigned char align; + if(((long)dst)&15) //FIXME + for(i=0; i<len-7; i+=8) { + s0 = vec_ld(0, src+i); + s1 = vec_ld(16, src+i); + t0 = vec_cts(s0, 0); + d0 = vec_ld(0, dst+i); + t1 = vec_cts(s1, 0); + d1 = vec_ld(15, dst+i); + d = vec_packs(t0,t1); + d1 = vec_perm(d1, d0, vec_lvsl(0,dst+i)); + align = vec_lvsr(0, dst+i); + d0 = vec_perm(d1, d, align); + d1 = vec_perm(d, d1, align); + vec_st(d0, 0, dst+i); + vec_st(d1,15, dst+i); + } + else + for(i=0; i<len-7; i+=8) { + s0 = vec_ld(0, src+i); + s1 = vec_ld(16, src+i); + t0 = vec_cts(s0, 0); + t1 = vec_cts(s1, 0); + d = vec_packs(t0,t1); + vec_st(d, 0, dst+i); + } +} + +void float_init_altivec(DSPContext* c, AVCodecContext *avctx) +{ + c->vector_fmul = vector_fmul_altivec; + c->vector_fmul_reverse = vector_fmul_reverse_altivec; + c->vector_fmul_add_add = vector_fmul_add_add_altivec; + if(!(avctx->flags & CODEC_FLAG_BITEXACT)) + c->float_to_int16 = float_to_int16_altivec; +} diff --git a/src/libffmpeg/libavcodec/ppc/h264_altivec.c b/src/libffmpeg/libavcodec/ppc/h264_altivec.c new file mode 100644 index 000000000..bac620e82 --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/h264_altivec.c @@ -0,0 +1,565 @@ +/* + * Copyright (c) 2004 Romain Dolbeau <romain@dolbeau.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "../dsputil.h" + +#include "gcc_fixes.h" + +#include "dsputil_altivec.h" +#include "types_altivec.h" + +#define PUT_OP_U8_ALTIVEC(d, s, dst) d = s +#define AVG_OP_U8_ALTIVEC(d, s, dst) d = vec_avg(dst, s) + +#define OP_U8_ALTIVEC PUT_OP_U8_ALTIVEC +#define PREFIX_h264_chroma_mc8_altivec put_h264_chroma_mc8_altivec +#define PREFIX_h264_chroma_mc8_num altivec_put_h264_chroma_mc8_num +#define PREFIX_h264_qpel16_h_lowpass_altivec put_h264_qpel16_h_lowpass_altivec +#define PREFIX_h264_qpel16_h_lowpass_num altivec_put_h264_qpel16_h_lowpass_num +#define PREFIX_h264_qpel16_v_lowpass_altivec put_h264_qpel16_v_lowpass_altivec +#define PREFIX_h264_qpel16_v_lowpass_num altivec_put_h264_qpel16_v_lowpass_num +#define PREFIX_h264_qpel16_hv_lowpass_altivec put_h264_qpel16_hv_lowpass_altivec +#define PREFIX_h264_qpel16_hv_lowpass_num altivec_put_h264_qpel16_hv_lowpass_num +#include "h264_template_altivec.c" +#undef OP_U8_ALTIVEC +#undef PREFIX_h264_chroma_mc8_altivec +#undef PREFIX_h264_chroma_mc8_num +#undef PREFIX_h264_qpel16_h_lowpass_altivec +#undef PREFIX_h264_qpel16_h_lowpass_num +#undef PREFIX_h264_qpel16_v_lowpass_altivec +#undef PREFIX_h264_qpel16_v_lowpass_num +#undef PREFIX_h264_qpel16_hv_lowpass_altivec +#undef PREFIX_h264_qpel16_hv_lowpass_num + +#define OP_U8_ALTIVEC AVG_OP_U8_ALTIVEC +#define PREFIX_h264_chroma_mc8_altivec avg_h264_chroma_mc8_altivec +#define PREFIX_h264_chroma_mc8_num altivec_avg_h264_chroma_mc8_num +#define PREFIX_h264_qpel16_h_lowpass_altivec avg_h264_qpel16_h_lowpass_altivec +#define PREFIX_h264_qpel16_h_lowpass_num altivec_avg_h264_qpel16_h_lowpass_num +#define PREFIX_h264_qpel16_v_lowpass_altivec avg_h264_qpel16_v_lowpass_altivec +#define PREFIX_h264_qpel16_v_lowpass_num altivec_avg_h264_qpel16_v_lowpass_num +#define PREFIX_h264_qpel16_hv_lowpass_altivec avg_h264_qpel16_hv_lowpass_altivec +#define PREFIX_h264_qpel16_hv_lowpass_num altivec_avg_h264_qpel16_hv_lowpass_num +#include "h264_template_altivec.c" +#undef OP_U8_ALTIVEC +#undef PREFIX_h264_chroma_mc8_altivec +#undef PREFIX_h264_chroma_mc8_num +#undef PREFIX_h264_qpel16_h_lowpass_altivec +#undef PREFIX_h264_qpel16_h_lowpass_num +#undef PREFIX_h264_qpel16_v_lowpass_altivec +#undef PREFIX_h264_qpel16_v_lowpass_num +#undef PREFIX_h264_qpel16_hv_lowpass_altivec +#undef PREFIX_h264_qpel16_hv_lowpass_num + +#define H264_MC(OPNAME, SIZE, CODETYPE) \ +static void OPNAME ## h264_qpel ## SIZE ## _mc00_ ## CODETYPE (uint8_t *dst, uint8_t *src, int stride){\ + OPNAME ## pixels ## SIZE ## _ ## CODETYPE(dst, src, stride, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc10_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){ \ + DECLARE_ALIGNED_16(uint8_t, half[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc20_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + OPNAME ## h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(dst, src, stride, stride);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc30_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, half[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(half, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+1, half, stride, stride, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc01_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, half[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src, half, stride, stride, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc02_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + OPNAME ## h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(dst, src, stride, stride);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc03_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, half[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(half, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, src+stride, half, stride, stride, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc11_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc31_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc13_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc33_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc22_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(int16_t, tmp[SIZE*(SIZE+8)]);\ + OPNAME ## h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(dst, tmp, src, stride, SIZE, stride);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc21_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfHV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(int16_t, tmp[SIZE*(SIZE+8)]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src, SIZE, stride);\ + put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc23_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfH[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfHV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(int16_t, tmp[SIZE*(SIZE+8)]);\ + put_h264_qpel ## SIZE ## _h_lowpass_ ## CODETYPE(halfH, src + stride, SIZE, stride);\ + put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfH, halfHV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc12_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfHV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(int16_t, tmp[SIZE*(SIZE+8)]);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src, SIZE, stride);\ + put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\ +}\ +\ +static void OPNAME ## h264_qpel ## SIZE ## _mc32_ ## CODETYPE(uint8_t *dst, uint8_t *src, int stride){\ + DECLARE_ALIGNED_16(uint8_t, halfV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(uint8_t, halfHV[SIZE*SIZE]);\ + DECLARE_ALIGNED_16(int16_t, tmp[SIZE*(SIZE+8)]);\ + put_h264_qpel ## SIZE ## _v_lowpass_ ## CODETYPE(halfV, src+1, SIZE, stride);\ + put_h264_qpel ## SIZE ## _hv_lowpass_ ## CODETYPE(halfHV, tmp, src, SIZE, SIZE, stride);\ + OPNAME ## pixels ## SIZE ## _l2_ ## CODETYPE(dst, halfV, halfHV, stride, SIZE, SIZE);\ +}\ + +/* this code assume that stride % 16 == 0 */ +void put_no_rnd_h264_chroma_mc8_altivec(uint8_t * dst, uint8_t * src, int stride, int h, int x, int y) { + signed int ABCD[4] __attribute__((aligned(16))) = + {((8 - x) * (8 - y)), + ((x) * (8 - y)), + ((8 - x) * (y)), + ((x) * (y))}; + register int i; + vector unsigned char fperm; + const vector signed int vABCD = vec_ld(0, ABCD); + const vector signed short vA = vec_splat((vector signed short)vABCD, 1); + const vector signed short vB = vec_splat((vector signed short)vABCD, 3); + const vector signed short vC = vec_splat((vector signed short)vABCD, 5); + const vector signed short vD = vec_splat((vector signed short)vABCD, 7); + const vector signed int vzero = vec_splat_s32(0); + const vector signed short v28ss = vec_sub(vec_sl(vec_splat_s16(1),vec_splat_u16(5)),vec_splat_s16(4)); + const vector unsigned short v6us = vec_splat_u16(6); + register int loadSecond = (((unsigned long)src) % 16) <= 7 ? 0 : 1; + register int reallyBadAlign = (((unsigned long)src) % 16) == 15 ? 1 : 0; + + vector unsigned char vsrcAuc, vsrcBuc, vsrcperm0, vsrcperm1; + vector unsigned char vsrc0uc, vsrc1uc; + vector signed short vsrc0ssH, vsrc1ssH; + vector unsigned char vsrcCuc, vsrc2uc, vsrc3uc; + vector signed short vsrc2ssH, vsrc3ssH, psum; + vector unsigned char vdst, ppsum, fsum; + + if (((unsigned long)dst) % 16 == 0) { + fperm = (vector unsigned char)AVV(0x10, 0x11, 0x12, 0x13, + 0x14, 0x15, 0x16, 0x17, + 0x08, 0x09, 0x0A, 0x0B, + 0x0C, 0x0D, 0x0E, 0x0F); + } else { + fperm = (vector unsigned char)AVV(0x00, 0x01, 0x02, 0x03, + 0x04, 0x05, 0x06, 0x07, + 0x18, 0x19, 0x1A, 0x1B, + 0x1C, 0x1D, 0x1E, 0x1F); + } + + vsrcAuc = vec_ld(0, src); + + if (loadSecond) + vsrcBuc = vec_ld(16, src); + vsrcperm0 = vec_lvsl(0, src); + vsrcperm1 = vec_lvsl(1, src); + + vsrc0uc = vec_perm(vsrcAuc, vsrcBuc, vsrcperm0); + if (reallyBadAlign) + vsrc1uc = vsrcBuc; + else + vsrc1uc = vec_perm(vsrcAuc, vsrcBuc, vsrcperm1); + + vsrc0ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc0uc); + vsrc1ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc1uc); + + if (!loadSecond) {// -> !reallyBadAlign + for (i = 0 ; i < h ; i++) { + + + vsrcCuc = vec_ld(stride + 0, src); + + vsrc2uc = vec_perm(vsrcCuc, vsrcCuc, vsrcperm0); + vsrc3uc = vec_perm(vsrcCuc, vsrcCuc, vsrcperm1); + + vsrc2ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc2uc); + vsrc3ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc3uc); + + psum = vec_mladd(vA, vsrc0ssH, vec_splat_s16(0)); + psum = vec_mladd(vB, vsrc1ssH, psum); + psum = vec_mladd(vC, vsrc2ssH, psum); + psum = vec_mladd(vD, vsrc3ssH, psum); + psum = vec_add(v28ss, psum); + psum = vec_sra(psum, v6us); + + vdst = vec_ld(0, dst); + ppsum = (vector unsigned char)vec_packsu(psum, psum); + fsum = vec_perm(vdst, ppsum, fperm); + + vec_st(fsum, 0, dst); + + vsrc0ssH = vsrc2ssH; + vsrc1ssH = vsrc3ssH; + + dst += stride; + src += stride; + } + } else { + vector unsigned char vsrcDuc; + for (i = 0 ; i < h ; i++) { + vsrcCuc = vec_ld(stride + 0, src); + vsrcDuc = vec_ld(stride + 16, src); + + vsrc2uc = vec_perm(vsrcCuc, vsrcDuc, vsrcperm0); + if (reallyBadAlign) + vsrc3uc = vsrcDuc; + else + vsrc3uc = vec_perm(vsrcCuc, vsrcDuc, vsrcperm1); + + vsrc2ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc2uc); + vsrc3ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc3uc); + + psum = vec_mladd(vA, vsrc0ssH, vec_splat_s16(0)); + psum = vec_mladd(vB, vsrc1ssH, psum); + psum = vec_mladd(vC, vsrc2ssH, psum); + psum = vec_mladd(vD, vsrc3ssH, psum); + psum = vec_add(v28ss, psum); + psum = vec_sr(psum, v6us); + + vdst = vec_ld(0, dst); + ppsum = (vector unsigned char)vec_pack(psum, psum); + fsum = vec_perm(vdst, ppsum, fperm); + + vec_st(fsum, 0, dst); + + vsrc0ssH = vsrc2ssH; + vsrc1ssH = vsrc3ssH; + + dst += stride; + src += stride; + } + } +} + +static inline void put_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1, + const uint8_t * src2, int dst_stride, + int src_stride1, int h) +{ + int i; + vector unsigned char a, b, d, tmp1, tmp2, mask, mask_, edges, align; + + mask_ = vec_lvsl(0, src2); + + for (i = 0; i < h; i++) { + + tmp1 = vec_ld(i * src_stride1, src1); + mask = vec_lvsl(i * src_stride1, src1); + tmp2 = vec_ld(i * src_stride1 + 15, src1); + + a = vec_perm(tmp1, tmp2, mask); + + tmp1 = vec_ld(i * 16, src2); + tmp2 = vec_ld(i * 16 + 15, src2); + + b = vec_perm(tmp1, tmp2, mask_); + + tmp1 = vec_ld(0, dst); + mask = vec_lvsl(0, dst); + tmp2 = vec_ld(15, dst); + + d = vec_avg(a, b); + + edges = vec_perm(tmp2, tmp1, mask); + + align = vec_lvsr(0, dst); + + tmp2 = vec_perm(d, edges, align); + tmp1 = vec_perm(edges, d, align); + + vec_st(tmp2, 15, dst); + vec_st(tmp1, 0 , dst); + + dst += dst_stride; + } +} + +static inline void avg_pixels16_l2_altivec( uint8_t * dst, const uint8_t * src1, + const uint8_t * src2, int dst_stride, + int src_stride1, int h) +{ + int i; + vector unsigned char a, b, d, tmp1, tmp2, mask, mask_, edges, align; + + mask_ = vec_lvsl(0, src2); + + for (i = 0; i < h; i++) { + + tmp1 = vec_ld(i * src_stride1, src1); + mask = vec_lvsl(i * src_stride1, src1); + tmp2 = vec_ld(i * src_stride1 + 15, src1); + + a = vec_perm(tmp1, tmp2, mask); + + tmp1 = vec_ld(i * 16, src2); + tmp2 = vec_ld(i * 16 + 15, src2); + + b = vec_perm(tmp1, tmp2, mask_); + + tmp1 = vec_ld(0, dst); + mask = vec_lvsl(0, dst); + tmp2 = vec_ld(15, dst); + + d = vec_avg(vec_perm(tmp1, tmp2, mask), vec_avg(a, b)); + + edges = vec_perm(tmp2, tmp1, mask); + + align = vec_lvsr(0, dst); + + tmp2 = vec_perm(d, edges, align); + tmp1 = vec_perm(edges, d, align); + + vec_st(tmp2, 15, dst); + vec_st(tmp1, 0 , dst); + + dst += dst_stride; + } +} + +/* Implemented but could be faster +#define put_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) put_pixels16_l2(d,s1,s2,ds,s1s,16,h) +#define avg_pixels16_l2_altivec(d,s1,s2,ds,s1s,h) avg_pixels16_l2(d,s1,s2,ds,s1s,16,h) + */ + + H264_MC(put_, 16, altivec) + H264_MC(avg_, 16, altivec) + + +/**************************************************************************** + * IDCT transform: + ****************************************************************************/ + +#define IDCT8_1D_ALTIVEC(s0, s1, s2, s3, s4, s5, s6, s7, d0, d1, d2, d3, d4, d5, d6, d7) {\ + /* a0 = SRC(0) + SRC(4); */ \ + vec_s16_t a0v = vec_add(s0, s4); \ + /* a2 = SRC(0) - SRC(4); */ \ + vec_s16_t a2v = vec_sub(s0, s4); \ + /* a4 = (SRC(2)>>1) - SRC(6); */ \ + vec_s16_t a4v = vec_sub(vec_sra(s2, onev), s6); \ + /* a6 = (SRC(6)>>1) + SRC(2); */ \ + vec_s16_t a6v = vec_add(vec_sra(s6, onev), s2); \ + /* b0 = a0 + a6; */ \ + vec_s16_t b0v = vec_add(a0v, a6v); \ + /* b2 = a2 + a4; */ \ + vec_s16_t b2v = vec_add(a2v, a4v); \ + /* b4 = a2 - a4; */ \ + vec_s16_t b4v = vec_sub(a2v, a4v); \ + /* b6 = a0 - a6; */ \ + vec_s16_t b6v = vec_sub(a0v, a6v); \ + /* a1 = SRC(5) - SRC(3) - SRC(7) - (SRC(7)>>1); */ \ + /* a1 = (SRC(5)-SRC(3)) - (SRC(7) + (SRC(7)>>1)); */ \ + vec_s16_t a1v = vec_sub( vec_sub(s5, s3), vec_add(s7, vec_sra(s7, onev)) ); \ + /* a3 = SRC(7) + SRC(1) - SRC(3) - (SRC(3)>>1); */ \ + /* a3 = (SRC(7)+SRC(1)) - (SRC(3) + (SRC(3)>>1)); */ \ + vec_s16_t a3v = vec_sub( vec_add(s7, s1), vec_add(s3, vec_sra(s3, onev)) );\ + /* a5 = SRC(7) - SRC(1) + SRC(5) + (SRC(5)>>1); */ \ + /* a5 = (SRC(7)-SRC(1)) + SRC(5) + (SRC(5)>>1); */ \ + vec_s16_t a5v = vec_add( vec_sub(s7, s1), vec_add(s5, vec_sra(s5, onev)) );\ + /* a7 = SRC(5)+SRC(3) + SRC(1) + (SRC(1)>>1); */ \ + vec_s16_t a7v = vec_add( vec_add(s5, s3), vec_add(s1, vec_sra(s1, onev)) );\ + /* b1 = (a7>>2) + a1; */ \ + vec_s16_t b1v = vec_add( vec_sra(a7v, twov), a1v); \ + /* b3 = a3 + (a5>>2); */ \ + vec_s16_t b3v = vec_add(a3v, vec_sra(a5v, twov)); \ + /* b5 = (a3>>2) - a5; */ \ + vec_s16_t b5v = vec_sub( vec_sra(a3v, twov), a5v); \ + /* b7 = a7 - (a1>>2); */ \ + vec_s16_t b7v = vec_sub( a7v, vec_sra(a1v, twov)); \ + /* DST(0, b0 + b7); */ \ + d0 = vec_add(b0v, b7v); \ + /* DST(1, b2 + b5); */ \ + d1 = vec_add(b2v, b5v); \ + /* DST(2, b4 + b3); */ \ + d2 = vec_add(b4v, b3v); \ + /* DST(3, b6 + b1); */ \ + d3 = vec_add(b6v, b1v); \ + /* DST(4, b6 - b1); */ \ + d4 = vec_sub(b6v, b1v); \ + /* DST(5, b4 - b3); */ \ + d5 = vec_sub(b4v, b3v); \ + /* DST(6, b2 - b5); */ \ + d6 = vec_sub(b2v, b5v); \ + /* DST(7, b0 - b7); */ \ + d7 = vec_sub(b0v, b7v); \ +} + +#define ALTIVEC_STORE_SUM_CLIP(dest, idctv, perm_ldv, perm_stv, sel) { \ + /* unaligned load */ \ + vec_u8_t hv = vec_ld( 0, dest ); \ + vec_u8_t lv = vec_ld( 7, dest ); \ + vec_u8_t dstv = vec_perm( hv, lv, (vec_u8_t)perm_ldv ); \ + vec_s16_t idct_sh6 = vec_sra(idctv, sixv); \ + vec_u16_t dst16 = (vec_u16_t)vec_mergeh(zero_u8v, dstv); \ + vec_s16_t idstsum = vec_adds(idct_sh6, (vec_s16_t)dst16); \ + vec_u8_t idstsum8 = vec_packsu(zero_s16v, idstsum); \ + vec_u8_t edgehv; \ + /* unaligned store */ \ + vec_u8_t bodyv = vec_perm( idstsum8, idstsum8, perm_stv );\ + vec_u8_t edgelv = vec_perm( sel, zero_u8v, perm_stv ); \ + lv = vec_sel( lv, bodyv, edgelv ); \ + vec_st( lv, 7, dest ); \ + hv = vec_ld( 0, dest ); \ + edgehv = vec_perm( zero_u8v, sel, perm_stv ); \ + hv = vec_sel( hv, bodyv, edgehv ); \ + vec_st( hv, 0, dest ); \ + } + +void ff_h264_idct8_add_altivec( uint8_t *dst, DCTELEM *dct, int stride ) { + vec_s16_t s0, s1, s2, s3, s4, s5, s6, s7; + vec_s16_t d0, d1, d2, d3, d4, d5, d6, d7; + vec_s16_t idct0, idct1, idct2, idct3, idct4, idct5, idct6, idct7; + + vec_u8_t perm_ldv = vec_lvsl(0, dst); + vec_u8_t perm_stv = vec_lvsr(8, dst); + + const vec_u16_t onev = vec_splat_u16(1); + const vec_u16_t twov = vec_splat_u16(2); + const vec_u16_t sixv = vec_splat_u16(6); + + const vec_u8_t sel = (vec_u8_t) AVV(0,0,0,0,0,0,0,0, + -1,-1,-1,-1,-1,-1,-1,-1); + LOAD_ZERO; + + dct[0] += 32; // rounding for the >>6 at the end + + s0 = vec_ld(0x00, (int16_t*)dct); + s1 = vec_ld(0x10, (int16_t*)dct); + s2 = vec_ld(0x20, (int16_t*)dct); + s3 = vec_ld(0x30, (int16_t*)dct); + s4 = vec_ld(0x40, (int16_t*)dct); + s5 = vec_ld(0x50, (int16_t*)dct); + s6 = vec_ld(0x60, (int16_t*)dct); + s7 = vec_ld(0x70, (int16_t*)dct); + + IDCT8_1D_ALTIVEC(s0, s1, s2, s3, s4, s5, s6, s7, + d0, d1, d2, d3, d4, d5, d6, d7); + + TRANSPOSE8( d0, d1, d2, d3, d4, d5, d6, d7 ); + + IDCT8_1D_ALTIVEC(d0, d1, d2, d3, d4, d5, d6, d7, + idct0, idct1, idct2, idct3, idct4, idct5, idct6, idct7); + + ALTIVEC_STORE_SUM_CLIP(&dst[0*stride], idct0, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[1*stride], idct1, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[2*stride], idct2, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[3*stride], idct3, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[4*stride], idct4, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[5*stride], idct5, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[6*stride], idct6, perm_ldv, perm_stv, sel); + ALTIVEC_STORE_SUM_CLIP(&dst[7*stride], idct7, perm_ldv, perm_stv, sel); +} + +void dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) { + +#ifdef HAVE_ALTIVEC + if (has_altivec()) { + c->put_h264_chroma_pixels_tab[0] = put_h264_chroma_mc8_altivec; + c->put_no_rnd_h264_chroma_pixels_tab[0] = put_no_rnd_h264_chroma_mc8_altivec; + c->avg_h264_chroma_pixels_tab[0] = avg_h264_chroma_mc8_altivec; + c->h264_idct8_add = ff_h264_idct8_add_altivec; + +#define dspfunc(PFX, IDX, NUM) \ + c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_altivec; \ + c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_altivec; \ + c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_altivec; \ + c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_altivec; \ + c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_altivec; \ + c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_altivec; \ + c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_altivec; \ + c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_altivec; \ + c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_altivec; \ + c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_altivec; \ + c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_altivec; \ + c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_altivec; \ + c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_altivec; \ + c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_altivec; \ + c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_altivec; \ + c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_altivec + + dspfunc(put_h264_qpel, 0, 16); + dspfunc(avg_h264_qpel, 0, 16); +#undef dspfunc + + } else +#endif /* HAVE_ALTIVEC */ + { + // Non-AltiVec PPC optimisations + + // ... pending ... + } +} diff --git a/src/libffmpeg/libavcodec/ppc/h264_template_altivec.c b/src/libffmpeg/libavcodec/ppc/h264_template_altivec.c new file mode 100644 index 000000000..e8ad67f2f --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/h264_template_altivec.c @@ -0,0 +1,719 @@ +/* + * Copyright (c) 2004 Romain Dolbeau <romain@dolbeau.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/* this code assume that stride % 16 == 0 */ +void PREFIX_h264_chroma_mc8_altivec(uint8_t * dst, uint8_t * src, int stride, int h, int x, int y) { + POWERPC_PERF_DECLARE(PREFIX_h264_chroma_mc8_num, 1); + signed int ABCD[4] __attribute__((aligned(16))) = + {((8 - x) * (8 - y)), + ((x) * (8 - y)), + ((8 - x) * (y)), + ((x) * (y))}; + register int i; + vector unsigned char fperm; + const vector signed int vABCD = vec_ld(0, ABCD); + const vector signed short vA = vec_splat((vector signed short)vABCD, 1); + const vector signed short vB = vec_splat((vector signed short)vABCD, 3); + const vector signed short vC = vec_splat((vector signed short)vABCD, 5); + const vector signed short vD = vec_splat((vector signed short)vABCD, 7); + const vector signed int vzero = vec_splat_s32(0); + const vector signed short v32ss = vec_sl(vec_splat_s16(1),vec_splat_u16(5)); + const vector unsigned short v6us = vec_splat_u16(6); + register int loadSecond = (((unsigned long)src) % 16) <= 7 ? 0 : 1; + register int reallyBadAlign = (((unsigned long)src) % 16) == 15 ? 1 : 0; + + vector unsigned char vsrcAuc, vsrcBuc, vsrcperm0, vsrcperm1; + vector unsigned char vsrc0uc, vsrc1uc; + vector signed short vsrc0ssH, vsrc1ssH; + vector unsigned char vsrcCuc, vsrc2uc, vsrc3uc; + vector signed short vsrc2ssH, vsrc3ssH, psum; + vector unsigned char vdst, ppsum, vfdst, fsum; + + POWERPC_PERF_START_COUNT(PREFIX_h264_chroma_mc8_num, 1); + + if (((unsigned long)dst) % 16 == 0) { + fperm = (vector unsigned char)AVV(0x10, 0x11, 0x12, 0x13, + 0x14, 0x15, 0x16, 0x17, + 0x08, 0x09, 0x0A, 0x0B, + 0x0C, 0x0D, 0x0E, 0x0F); + } else { + fperm = (vector unsigned char)AVV(0x00, 0x01, 0x02, 0x03, + 0x04, 0x05, 0x06, 0x07, + 0x18, 0x19, 0x1A, 0x1B, + 0x1C, 0x1D, 0x1E, 0x1F); + } + + vsrcAuc = vec_ld(0, src); + + if (loadSecond) + vsrcBuc = vec_ld(16, src); + vsrcperm0 = vec_lvsl(0, src); + vsrcperm1 = vec_lvsl(1, src); + + vsrc0uc = vec_perm(vsrcAuc, vsrcBuc, vsrcperm0); + if (reallyBadAlign) + vsrc1uc = vsrcBuc; + else + vsrc1uc = vec_perm(vsrcAuc, vsrcBuc, vsrcperm1); + + vsrc0ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc0uc); + vsrc1ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc1uc); + + if (!loadSecond) {// -> !reallyBadAlign + for (i = 0 ; i < h ; i++) { + + + vsrcCuc = vec_ld(stride + 0, src); + + vsrc2uc = vec_perm(vsrcCuc, vsrcCuc, vsrcperm0); + vsrc3uc = vec_perm(vsrcCuc, vsrcCuc, vsrcperm1); + + vsrc2ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc2uc); + vsrc3ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc3uc); + + psum = vec_mladd(vA, vsrc0ssH, vec_splat_s16(0)); + psum = vec_mladd(vB, vsrc1ssH, psum); + psum = vec_mladd(vC, vsrc2ssH, psum); + psum = vec_mladd(vD, vsrc3ssH, psum); + psum = vec_add(v32ss, psum); + psum = vec_sra(psum, v6us); + + vdst = vec_ld(0, dst); + ppsum = (vector unsigned char)vec_packsu(psum, psum); + vfdst = vec_perm(vdst, ppsum, fperm); + + OP_U8_ALTIVEC(fsum, vfdst, vdst); + + vec_st(fsum, 0, dst); + + vsrc0ssH = vsrc2ssH; + vsrc1ssH = vsrc3ssH; + + dst += stride; + src += stride; + } + } else { + vector unsigned char vsrcDuc; + for (i = 0 ; i < h ; i++) { + vsrcCuc = vec_ld(stride + 0, src); + vsrcDuc = vec_ld(stride + 16, src); + + vsrc2uc = vec_perm(vsrcCuc, vsrcDuc, vsrcperm0); + if (reallyBadAlign) + vsrc3uc = vsrcDuc; + else + vsrc3uc = vec_perm(vsrcCuc, vsrcDuc, vsrcperm1); + + vsrc2ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc2uc); + vsrc3ssH = (vector signed short)vec_mergeh((vector unsigned char)vzero, + (vector unsigned char)vsrc3uc); + + psum = vec_mladd(vA, vsrc0ssH, vec_splat_s16(0)); + psum = vec_mladd(vB, vsrc1ssH, psum); + psum = vec_mladd(vC, vsrc2ssH, psum); + psum = vec_mladd(vD, vsrc3ssH, psum); + psum = vec_add(v32ss, psum); + psum = vec_sr(psum, v6us); + + vdst = vec_ld(0, dst); + ppsum = (vector unsigned char)vec_pack(psum, psum); + vfdst = vec_perm(vdst, ppsum, fperm); + + OP_U8_ALTIVEC(fsum, vfdst, vdst); + + vec_st(fsum, 0, dst); + + vsrc0ssH = vsrc2ssH; + vsrc1ssH = vsrc3ssH; + + dst += stride; + src += stride; + } + } + POWERPC_PERF_STOP_COUNT(PREFIX_h264_chroma_mc8_num, 1); +} + +/* this code assume stride % 16 == 0 */ +static void PREFIX_h264_qpel16_h_lowpass_altivec(uint8_t * dst, uint8_t * src, int dstStride, int srcStride) { + POWERPC_PERF_DECLARE(PREFIX_h264_qpel16_h_lowpass_num, 1); + register int i; + + const vector signed int vzero = vec_splat_s32(0); + const vector unsigned char permM2 = vec_lvsl(-2, src); + const vector unsigned char permM1 = vec_lvsl(-1, src); + const vector unsigned char permP0 = vec_lvsl(+0, src); + const vector unsigned char permP1 = vec_lvsl(+1, src); + const vector unsigned char permP2 = vec_lvsl(+2, src); + const vector unsigned char permP3 = vec_lvsl(+3, src); + const vector signed short v5ss = vec_splat_s16(5); + const vector unsigned short v5us = vec_splat_u16(5); + const vector signed short v20ss = vec_sl(vec_splat_s16(5),vec_splat_u16(2)); + const vector signed short v16ss = vec_sl(vec_splat_s16(1),vec_splat_u16(4)); + const vector unsigned char dstperm = vec_lvsr(0, dst); + const vector unsigned char neg1 = + (const vector unsigned char) vec_splat_s8(-1); + + const vector unsigned char dstmask = + vec_perm((const vector unsigned char)vzero, + neg1, dstperm); + + vector unsigned char srcM2, srcM1, srcP0, srcP1, srcP2, srcP3; + + register int align = ((((unsigned long)src) - 2) % 16); + + vector signed short srcP0A, srcP0B, srcP1A, srcP1B, + srcP2A, srcP2B, srcP3A, srcP3B, + srcM1A, srcM1B, srcM2A, srcM2B, + sum1A, sum1B, sum2A, sum2B, sum3A, sum3B, + pp1A, pp1B, pp2A, pp2B, pp3A, pp3B, + psumA, psumB, sumA, sumB; + + vector unsigned char sum, dst1, dst2, vdst, fsum, + rsum, fdst1, fdst2; + + POWERPC_PERF_START_COUNT(PREFIX_h264_qpel16_h_lowpass_num, 1); + + for (i = 0 ; i < 16 ; i ++) { + vector unsigned char srcR1 = vec_ld(-2, src); + vector unsigned char srcR2 = vec_ld(14, src); + + switch (align) { + default: { + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = vec_perm(srcR1, srcR2, permP2); + srcP3 = vec_perm(srcR1, srcR2, permP3); + } break; + case 11: { + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = vec_perm(srcR1, srcR2, permP2); + srcP3 = srcR2; + } break; + case 12: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = srcR2; + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 13: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = srcR2; + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 14: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = srcR2; + srcP1 = vec_perm(srcR2, srcR3, permP1); + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 15: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = srcR2; + srcP0 = vec_perm(srcR2, srcR3, permP0); + srcP1 = vec_perm(srcR2, srcR3, permP1); + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + } + + srcP0A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP0); + srcP0B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP0); + srcP1A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP1); + srcP1B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP1); + + srcP2A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP2); + srcP2B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP2); + srcP3A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP3); + srcP3B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP3); + + srcM1A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM1); + srcM1B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM1); + srcM2A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM2); + srcM2B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM2); + + sum1A = vec_adds(srcP0A, srcP1A); + sum1B = vec_adds(srcP0B, srcP1B); + sum2A = vec_adds(srcM1A, srcP2A); + sum2B = vec_adds(srcM1B, srcP2B); + sum3A = vec_adds(srcM2A, srcP3A); + sum3B = vec_adds(srcM2B, srcP3B); + + pp1A = vec_mladd(sum1A, v20ss, v16ss); + pp1B = vec_mladd(sum1B, v20ss, v16ss); + + pp2A = vec_mladd(sum2A, v5ss, (vector signed short)vzero); + pp2B = vec_mladd(sum2B, v5ss, (vector signed short)vzero); + + pp3A = vec_add(sum3A, pp1A); + pp3B = vec_add(sum3B, pp1B); + + psumA = vec_sub(pp3A, pp2A); + psumB = vec_sub(pp3B, pp2B); + + sumA = vec_sra(psumA, v5us); + sumB = vec_sra(psumB, v5us); + + sum = vec_packsu(sumA, sumB); + + dst1 = vec_ld(0, dst); + dst2 = vec_ld(16, dst); + vdst = vec_perm(dst1, dst2, vec_lvsl(0, dst)); + + OP_U8_ALTIVEC(fsum, sum, vdst); + + rsum = vec_perm(fsum, fsum, dstperm); + fdst1 = vec_sel(dst1, rsum, dstmask); + fdst2 = vec_sel(rsum, dst2, dstmask); + + vec_st(fdst1, 0, dst); + vec_st(fdst2, 16, dst); + + src += srcStride; + dst += dstStride; + } +POWERPC_PERF_STOP_COUNT(PREFIX_h264_qpel16_h_lowpass_num, 1); +} + +/* this code assume stride % 16 == 0 */ +static void PREFIX_h264_qpel16_v_lowpass_altivec(uint8_t * dst, uint8_t * src, int dstStride, int srcStride) { + POWERPC_PERF_DECLARE(PREFIX_h264_qpel16_v_lowpass_num, 1); + + register int i; + + const vector signed int vzero = vec_splat_s32(0); + const vector unsigned char perm = vec_lvsl(0, src); + const vector signed short v20ss = vec_sl(vec_splat_s16(5),vec_splat_u16(2)); + const vector unsigned short v5us = vec_splat_u16(5); + const vector signed short v5ss = vec_splat_s16(5); + const vector signed short v16ss = vec_sl(vec_splat_s16(1),vec_splat_u16(4)); + const vector unsigned char dstperm = vec_lvsr(0, dst); + const vector unsigned char neg1 = (const vector unsigned char)vec_splat_s8(-1); + const vector unsigned char dstmask = vec_perm((const vector unsigned char)vzero, neg1, dstperm); + + uint8_t *srcbis = src - (srcStride * 2); + + const vector unsigned char srcM2a = vec_ld(0, srcbis); + const vector unsigned char srcM2b = vec_ld(16, srcbis); + const vector unsigned char srcM2 = vec_perm(srcM2a, srcM2b, perm); +// srcbis += srcStride; + const vector unsigned char srcM1a = vec_ld(0, srcbis += srcStride); + const vector unsigned char srcM1b = vec_ld(16, srcbis); + const vector unsigned char srcM1 = vec_perm(srcM1a, srcM1b, perm); +// srcbis += srcStride; + const vector unsigned char srcP0a = vec_ld(0, srcbis += srcStride); + const vector unsigned char srcP0b = vec_ld(16, srcbis); + const vector unsigned char srcP0 = vec_perm(srcP0a, srcP0b, perm); +// srcbis += srcStride; + const vector unsigned char srcP1a = vec_ld(0, srcbis += srcStride); + const vector unsigned char srcP1b = vec_ld(16, srcbis); + const vector unsigned char srcP1 = vec_perm(srcP1a, srcP1b, perm); +// srcbis += srcStride; + const vector unsigned char srcP2a = vec_ld(0, srcbis += srcStride); + const vector unsigned char srcP2b = vec_ld(16, srcbis); + const vector unsigned char srcP2 = vec_perm(srcP2a, srcP2b, perm); +// srcbis += srcStride; + + vector signed short srcM2ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM2); + vector signed short srcM2ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM2); + vector signed short srcM1ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM1); + vector signed short srcM1ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM1); + vector signed short srcP0ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP0); + vector signed short srcP0ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP0); + vector signed short srcP1ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP1); + vector signed short srcP1ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP1); + vector signed short srcP2ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP2); + vector signed short srcP2ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP2); + + vector signed short pp1A, pp1B, pp2A, pp2B, pp3A, pp3B, + psumA, psumB, sumA, sumB, + srcP3ssA, srcP3ssB, + sum1A, sum1B, sum2A, sum2B, sum3A, sum3B; + + vector unsigned char sum, dst1, dst2, vdst, fsum, rsum, fdst1, fdst2, + srcP3a, srcP3b, srcP3; + + POWERPC_PERF_START_COUNT(PREFIX_h264_qpel16_v_lowpass_num, 1); + + for (i = 0 ; i < 16 ; i++) { + srcP3a = vec_ld(0, srcbis += srcStride); + srcP3b = vec_ld(16, srcbis); + srcP3 = vec_perm(srcP3a, srcP3b, perm); + srcP3ssA = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP3); + srcP3ssB = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP3); +// srcbis += srcStride; + + sum1A = vec_adds(srcP0ssA, srcP1ssA); + sum1B = vec_adds(srcP0ssB, srcP1ssB); + sum2A = vec_adds(srcM1ssA, srcP2ssA); + sum2B = vec_adds(srcM1ssB, srcP2ssB); + sum3A = vec_adds(srcM2ssA, srcP3ssA); + sum3B = vec_adds(srcM2ssB, srcP3ssB); + + srcM2ssA = srcM1ssA; + srcM2ssB = srcM1ssB; + srcM1ssA = srcP0ssA; + srcM1ssB = srcP0ssB; + srcP0ssA = srcP1ssA; + srcP0ssB = srcP1ssB; + srcP1ssA = srcP2ssA; + srcP1ssB = srcP2ssB; + srcP2ssA = srcP3ssA; + srcP2ssB = srcP3ssB; + + pp1A = vec_mladd(sum1A, v20ss, v16ss); + pp1B = vec_mladd(sum1B, v20ss, v16ss); + + pp2A = vec_mladd(sum2A, v5ss, (vector signed short)vzero); + pp2B = vec_mladd(sum2B, v5ss, (vector signed short)vzero); + + pp3A = vec_add(sum3A, pp1A); + pp3B = vec_add(sum3B, pp1B); + + psumA = vec_sub(pp3A, pp2A); + psumB = vec_sub(pp3B, pp2B); + + sumA = vec_sra(psumA, v5us); + sumB = vec_sra(psumB, v5us); + + sum = vec_packsu(sumA, sumB); + + dst1 = vec_ld(0, dst); + dst2 = vec_ld(16, dst); + vdst = vec_perm(dst1, dst2, vec_lvsl(0, dst)); + + OP_U8_ALTIVEC(fsum, sum, vdst); + + rsum = vec_perm(fsum, fsum, dstperm); + fdst1 = vec_sel(dst1, rsum, dstmask); + fdst2 = vec_sel(rsum, dst2, dstmask); + + vec_st(fdst1, 0, dst); + vec_st(fdst2, 16, dst); + + dst += dstStride; + } + POWERPC_PERF_STOP_COUNT(PREFIX_h264_qpel16_v_lowpass_num, 1); +} + +/* this code assume stride % 16 == 0 *and* tmp is properly aligned */ +static void PREFIX_h264_qpel16_hv_lowpass_altivec(uint8_t * dst, int16_t * tmp, uint8_t * src, int dstStride, int tmpStride, int srcStride) { + POWERPC_PERF_DECLARE(PREFIX_h264_qpel16_hv_lowpass_num, 1); + register int i; + const vector signed int vzero = vec_splat_s32(0); + const vector unsigned char permM2 = vec_lvsl(-2, src); + const vector unsigned char permM1 = vec_lvsl(-1, src); + const vector unsigned char permP0 = vec_lvsl(+0, src); + const vector unsigned char permP1 = vec_lvsl(+1, src); + const vector unsigned char permP2 = vec_lvsl(+2, src); + const vector unsigned char permP3 = vec_lvsl(+3, src); + const vector signed short v20ss = vec_sl(vec_splat_s16(5),vec_splat_u16(2)); + const vector unsigned int v10ui = vec_splat_u32(10); + const vector signed short v5ss = vec_splat_s16(5); + const vector signed short v1ss = vec_splat_s16(1); + const vector signed int v512si = vec_sl(vec_splat_s32(1),vec_splat_u32(9)); + const vector unsigned int v16ui = vec_sl(vec_splat_u32(1),vec_splat_u32(4)); + + register int align = ((((unsigned long)src) - 2) % 16); + + const vector unsigned char neg1 = (const vector unsigned char) + vec_splat_s8(-1); + + vector signed short srcP0A, srcP0B, srcP1A, srcP1B, + srcP2A, srcP2B, srcP3A, srcP3B, + srcM1A, srcM1B, srcM2A, srcM2B, + sum1A, sum1B, sum2A, sum2B, sum3A, sum3B, + pp1A, pp1B, pp2A, pp2B, psumA, psumB; + + const vector unsigned char dstperm = vec_lvsr(0, dst); + + const vector unsigned char dstmask = vec_perm((const vector unsigned char)vzero, neg1, dstperm); + + const vector unsigned char mperm = (const vector unsigned char) + AVV(0x00, 0x08, 0x01, 0x09, 0x02, 0x0A, 0x03, 0x0B, + 0x04, 0x0C, 0x05, 0x0D, 0x06, 0x0E, 0x07, 0x0F); + int16_t *tmpbis = tmp; + + vector signed short tmpM1ssA, tmpM1ssB, tmpM2ssA, tmpM2ssB, + tmpP0ssA, tmpP0ssB, tmpP1ssA, tmpP1ssB, + tmpP2ssA, tmpP2ssB; + + vector signed int pp1Ae, pp1Ao, pp1Be, pp1Bo, pp2Ae, pp2Ao, pp2Be, pp2Bo, + pp3Ae, pp3Ao, pp3Be, pp3Bo, pp1cAe, pp1cAo, pp1cBe, pp1cBo, + pp32Ae, pp32Ao, pp32Be, pp32Bo, sumAe, sumAo, sumBe, sumBo, + ssumAe, ssumAo, ssumBe, ssumBo; + vector unsigned char fsum, sumv, sum, dst1, dst2, vdst, + rsum, fdst1, fdst2; + vector signed short ssume, ssumo; + + POWERPC_PERF_START_COUNT(PREFIX_h264_qpel16_hv_lowpass_num, 1); + src -= (2 * srcStride); + for (i = 0 ; i < 21 ; i ++) { + vector unsigned char srcM2, srcM1, srcP0, srcP1, srcP2, srcP3; + vector unsigned char srcR1 = vec_ld(-2, src); + vector unsigned char srcR2 = vec_ld(14, src); + + switch (align) { + default: { + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = vec_perm(srcR1, srcR2, permP2); + srcP3 = vec_perm(srcR1, srcR2, permP3); + } break; + case 11: { + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = vec_perm(srcR1, srcR2, permP2); + srcP3 = srcR2; + } break; + case 12: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = vec_perm(srcR1, srcR2, permP1); + srcP2 = srcR2; + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 13: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = vec_perm(srcR1, srcR2, permP0); + srcP1 = srcR2; + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 14: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = vec_perm(srcR1, srcR2, permM1); + srcP0 = srcR2; + srcP1 = vec_perm(srcR2, srcR3, permP1); + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + case 15: { + vector unsigned char srcR3 = vec_ld(30, src); + srcM2 = vec_perm(srcR1, srcR2, permM2); + srcM1 = srcR2; + srcP0 = vec_perm(srcR2, srcR3, permP0); + srcP1 = vec_perm(srcR2, srcR3, permP1); + srcP2 = vec_perm(srcR2, srcR3, permP2); + srcP3 = vec_perm(srcR2, srcR3, permP3); + } break; + } + + srcP0A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP0); + srcP0B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP0); + srcP1A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP1); + srcP1B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP1); + + srcP2A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP2); + srcP2B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP2); + srcP3A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcP3); + srcP3B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcP3); + + srcM1A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM1); + srcM1B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM1); + srcM2A = (vector signed short) + vec_mergeh((vector unsigned char)vzero, srcM2); + srcM2B = (vector signed short) + vec_mergel((vector unsigned char)vzero, srcM2); + + sum1A = vec_adds(srcP0A, srcP1A); + sum1B = vec_adds(srcP0B, srcP1B); + sum2A = vec_adds(srcM1A, srcP2A); + sum2B = vec_adds(srcM1B, srcP2B); + sum3A = vec_adds(srcM2A, srcP3A); + sum3B = vec_adds(srcM2B, srcP3B); + + pp1A = vec_mladd(sum1A, v20ss, sum3A); + pp1B = vec_mladd(sum1B, v20ss, sum3B); + + pp2A = vec_mladd(sum2A, v5ss, (vector signed short)vzero); + pp2B = vec_mladd(sum2B, v5ss, (vector signed short)vzero); + + psumA = vec_sub(pp1A, pp2A); + psumB = vec_sub(pp1B, pp2B); + + vec_st(psumA, 0, tmp); + vec_st(psumB, 16, tmp); + + src += srcStride; + tmp += tmpStride; /* int16_t*, and stride is 16, so it's OK here */ + } + + tmpM2ssA = vec_ld(0, tmpbis); + tmpM2ssB = vec_ld(16, tmpbis); + tmpbis += tmpStride; + tmpM1ssA = vec_ld(0, tmpbis); + tmpM1ssB = vec_ld(16, tmpbis); + tmpbis += tmpStride; + tmpP0ssA = vec_ld(0, tmpbis); + tmpP0ssB = vec_ld(16, tmpbis); + tmpbis += tmpStride; + tmpP1ssA = vec_ld(0, tmpbis); + tmpP1ssB = vec_ld(16, tmpbis); + tmpbis += tmpStride; + tmpP2ssA = vec_ld(0, tmpbis); + tmpP2ssB = vec_ld(16, tmpbis); + tmpbis += tmpStride; + + for (i = 0 ; i < 16 ; i++) { + const vector signed short tmpP3ssA = vec_ld(0, tmpbis); + const vector signed short tmpP3ssB = vec_ld(16, tmpbis); + + const vector signed short sum1A = vec_adds(tmpP0ssA, tmpP1ssA); + const vector signed short sum1B = vec_adds(tmpP0ssB, tmpP1ssB); + const vector signed short sum2A = vec_adds(tmpM1ssA, tmpP2ssA); + const vector signed short sum2B = vec_adds(tmpM1ssB, tmpP2ssB); + const vector signed short sum3A = vec_adds(tmpM2ssA, tmpP3ssA); + const vector signed short sum3B = vec_adds(tmpM2ssB, tmpP3ssB); + + tmpbis += tmpStride; + + tmpM2ssA = tmpM1ssA; + tmpM2ssB = tmpM1ssB; + tmpM1ssA = tmpP0ssA; + tmpM1ssB = tmpP0ssB; + tmpP0ssA = tmpP1ssA; + tmpP0ssB = tmpP1ssB; + tmpP1ssA = tmpP2ssA; + tmpP1ssB = tmpP2ssB; + tmpP2ssA = tmpP3ssA; + tmpP2ssB = tmpP3ssB; + + pp1Ae = vec_mule(sum1A, v20ss); + pp1Ao = vec_mulo(sum1A, v20ss); + pp1Be = vec_mule(sum1B, v20ss); + pp1Bo = vec_mulo(sum1B, v20ss); + + pp2Ae = vec_mule(sum2A, v5ss); + pp2Ao = vec_mulo(sum2A, v5ss); + pp2Be = vec_mule(sum2B, v5ss); + pp2Bo = vec_mulo(sum2B, v5ss); + + pp3Ae = vec_sra((vector signed int)sum3A, v16ui); + pp3Ao = vec_mulo(sum3A, v1ss); + pp3Be = vec_sra((vector signed int)sum3B, v16ui); + pp3Bo = vec_mulo(sum3B, v1ss); + + pp1cAe = vec_add(pp1Ae, v512si); + pp1cAo = vec_add(pp1Ao, v512si); + pp1cBe = vec_add(pp1Be, v512si); + pp1cBo = vec_add(pp1Bo, v512si); + + pp32Ae = vec_sub(pp3Ae, pp2Ae); + pp32Ao = vec_sub(pp3Ao, pp2Ao); + pp32Be = vec_sub(pp3Be, pp2Be); + pp32Bo = vec_sub(pp3Bo, pp2Bo); + + sumAe = vec_add(pp1cAe, pp32Ae); + sumAo = vec_add(pp1cAo, pp32Ao); + sumBe = vec_add(pp1cBe, pp32Be); + sumBo = vec_add(pp1cBo, pp32Bo); + + ssumAe = vec_sra(sumAe, v10ui); + ssumAo = vec_sra(sumAo, v10ui); + ssumBe = vec_sra(sumBe, v10ui); + ssumBo = vec_sra(sumBo, v10ui); + + ssume = vec_packs(ssumAe, ssumBe); + ssumo = vec_packs(ssumAo, ssumBo); + + sumv = vec_packsu(ssume, ssumo); + sum = vec_perm(sumv, sumv, mperm); + + dst1 = vec_ld(0, dst); + dst2 = vec_ld(16, dst); + vdst = vec_perm(dst1, dst2, vec_lvsl(0, dst)); + + OP_U8_ALTIVEC(fsum, sum, vdst); + + rsum = vec_perm(fsum, fsum, dstperm); + fdst1 = vec_sel(dst1, rsum, dstmask); + fdst2 = vec_sel(rsum, dst2, dstmask); + + vec_st(fdst1, 0, dst); + vec_st(fdst2, 16, dst); + + dst += dstStride; + } + POWERPC_PERF_STOP_COUNT(PREFIX_h264_qpel16_hv_lowpass_num, 1); +} diff --git a/src/libffmpeg/libavcodec/ppc/mathops.h b/src/libffmpeg/libavcodec/ppc/mathops.h new file mode 100644 index 000000000..6af23f246 --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/mathops.h @@ -0,0 +1,33 @@ +/* + * simple math operations + * Copyright (c) 2001, 2002 Fabrice Bellard. + * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#if defined(ARCH_POWERPC_405) +/* signed 16x16 -> 32 multiply add accumulate */ +# define MAC16(rt, ra, rb) \ + asm ("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb)); + +/* signed 16x16 -> 32 multiply */ +# define MUL16(ra, rb) \ + ({ int __rt; + asm ("mullhw %0, %1, %2" : "=r" (__rt) : "r" (ra), "r" (rb)); + __rt; }) +#endif diff --git a/src/libffmpeg/libavcodec/ppc/snow_altivec.c b/src/libffmpeg/libavcodec/ppc/snow_altivec.c new file mode 100644 index 000000000..b15672ffe --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/snow_altivec.c @@ -0,0 +1,788 @@ +/* + * Altivec optimized snow DSP utils + * Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + * + * + */ + +#include "../dsputil.h" + +#include "gcc_fixes.h" +#include "dsputil_altivec.h" +#include "../snow.h" + +#undef NDEBUG +#include <assert.h> + + + +//FIXME remove this replication +#define slice_buffer_get_line(slice_buf, line_num) ((slice_buf)->line[line_num] ? (slice_buf)->line[line_num] : slice_buffer_load_line((slice_buf), (line_num))) + +static DWTELEM * slice_buffer_load_line(slice_buffer * buf, int line) +{ + int offset; + DWTELEM * buffer; + +// av_log(NULL, AV_LOG_DEBUG, "Cache hit: %d\n", line); + + assert(buf->data_stack_top >= 0); +// assert(!buf->line[line]); + if (buf->line[line]) + return buf->line[line]; + + offset = buf->line_width * line; + buffer = buf->data_stack[buf->data_stack_top]; + buf->data_stack_top--; + buf->line[line] = buffer; + +// av_log(NULL, AV_LOG_DEBUG, "slice_buffer_load_line: line: %d remaining: %d\n", line, buf->data_stack_top + 1); + + return buffer; +} + + +//altivec code + +void ff_snow_horizontal_compose97i_altivec(DWTELEM *b, int width) +{ + const int w2= (width+1)>>1; + DECLARE_ALIGNED_16(DWTELEM, temp[(width>>1)]); + const int w_l= (width>>1); + const int w_r= w2 - 1; + int i; + vector signed int t1, t2, x, y, tmp1, tmp2; + vector signed int *vbuf, *vtmp; + vector unsigned char align; + + + + { // Lift 0 + DWTELEM * const ref = b + w2 - 1; + DWTELEM b_0 = b[0]; + vbuf = (vector signed int *)b; + + tmp1 = vec_ld (0, ref); + align = vec_lvsl (0, ref); + tmp2 = vec_ld (15, ref); + t1= vec_perm(tmp1, tmp2, align); + + i = 0; + + for (i=0; i<w_l-15; i+=16) { +#if 0 + b[i+0] = b[i+0] - ((3 * (ref[i+0] + ref[i+1]) + 4) >> 3); + b[i+1] = b[i+1] - ((3 * (ref[i+1] + ref[i+2]) + 4) >> 3); + b[i+2] = b[i+2] - ((3 * (ref[i+2] + ref[i+3]) + 4) >> 3); + b[i+3] = b[i+3] - ((3 * (ref[i+3] + ref[i+4]) + 4) >> 3); +#else + + tmp1 = vec_ld (0, ref+4+i); + tmp2 = vec_ld (15, ref+4+i); + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_add(vec_add(y,y),y); + + tmp1 = vec_ld (0, ref+8+i); + + y = vec_add(y, vec_splat_s32(4)); + y = vec_sra(y, vec_splat_u32(3)); + + tmp2 = vec_ld (15, ref+8+i); + + *vbuf = vec_sub(*vbuf, y); + + t1=t2; + + vbuf++; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_add(vec_add(y,y),y); + + tmp1 = vec_ld (0, ref+12+i); + + y = vec_add(y, vec_splat_s32(4)); + y = vec_sra(y, vec_splat_u32(3)); + + tmp2 = vec_ld (15, ref+12+i); + + *vbuf = vec_sub(*vbuf, y); + + t1=t2; + + vbuf++; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_add(vec_add(y,y),y); + + tmp1 = vec_ld (0, ref+16+i); + + y = vec_add(y, vec_splat_s32(4)); + y = vec_sra(y, vec_splat_u32(3)); + + tmp2 = vec_ld (15, ref+16+i); + + *vbuf = vec_sub(*vbuf, y); + + t1=t2; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_add(vec_add(y,y),y); + + vbuf++; + + y = vec_add(y, vec_splat_s32(4)); + y = vec_sra(y, vec_splat_u32(3)); + *vbuf = vec_sub(*vbuf, y); + + t1=t2; + + vbuf++; + +#endif + } + + snow_horizontal_compose_lift_lead_out(i, b, b, ref, width, w_l, 0, W_DM, W_DO, W_DS); + b[0] = b_0 - ((W_DM * 2 * ref[1]+W_DO)>>W_DS); + } + + { // Lift 1 + DWTELEM * const dst = b+w2; + + i = 0; + for(; (((long)&dst[i]) & 0xF) && i<w_r; i++){ + dst[i] = dst[i] - (b[i] + b[i + 1]); + } + + align = vec_lvsl(0, b+i); + tmp1 = vec_ld(0, b+i); + vbuf = (vector signed int*) (dst + i); + tmp2 = vec_ld(15, b+i); + + t1 = vec_perm(tmp1, tmp2, align); + + for (; i<w_r-3; i+=4) { + +#if 0 + dst[i] = dst[i] - (b[i] + b[i + 1]); + dst[i+1] = dst[i+1] - (b[i+1] + b[i + 2]); + dst[i+2] = dst[i+2] - (b[i+2] + b[i + 3]); + dst[i+3] = dst[i+3] - (b[i+3] + b[i + 4]); +#else + + tmp1 = vec_ld(0, b+4+i); + tmp2 = vec_ld(15, b+4+i); + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1, vec_sld(t1,t2,4)); + *vbuf = vec_sub (*vbuf, y); + + vbuf++; + + t1 = t2; + +#endif + + } + + snow_horizontal_compose_lift_lead_out(i, dst, dst, b, width, w_r, 1, W_CM, W_CO, W_CS); + } + + { // Lift 2 + DWTELEM * const ref = b+w2 - 1; + DWTELEM b_0 = b[0]; + vbuf= (vector signed int *) b; + + tmp1 = vec_ld (0, ref); + align = vec_lvsl (0, ref); + tmp2 = vec_ld (15, ref); + t1= vec_perm(tmp1, tmp2, align); + + i = 0; + for (; i<w_l-15; i+=16) { +#if 0 + b[i] = b[i] - (((8 -(ref[i] + ref[i+1])) - (b[i] <<2)) >> 4); + b[i+1] = b[i+1] - (((8 -(ref[i+1] + ref[i+2])) - (b[i+1]<<2)) >> 4); + b[i+2] = b[i+2] - (((8 -(ref[i+2] + ref[i+3])) - (b[i+2]<<2)) >> 4); + b[i+3] = b[i+3] - (((8 -(ref[i+3] + ref[i+4])) - (b[i+3]<<2)) >> 4); +#else + tmp1 = vec_ld (0, ref+4+i); + tmp2 = vec_ld (15, ref+4+i); + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_sub(vec_splat_s32(8),y); + + tmp1 = vec_ld (0, ref+8+i); + + x = vec_sl(*vbuf,vec_splat_u32(2)); + y = vec_sra(vec_sub(y,x),vec_splat_u32(4)); + + tmp2 = vec_ld (15, ref+8+i); + + *vbuf = vec_sub( *vbuf, y); + + t1 = t2; + + vbuf++; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_sub(vec_splat_s32(8),y); + + tmp1 = vec_ld (0, ref+12+i); + + x = vec_sl(*vbuf,vec_splat_u32(2)); + y = vec_sra(vec_sub(y,x),vec_splat_u32(4)); + + tmp2 = vec_ld (15, ref+12+i); + + *vbuf = vec_sub( *vbuf, y); + + t1 = t2; + + vbuf++; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_sub(vec_splat_s32(8),y); + + tmp1 = vec_ld (0, ref+16+i); + + x = vec_sl(*vbuf,vec_splat_u32(2)); + y = vec_sra(vec_sub(y,x),vec_splat_u32(4)); + + tmp2 = vec_ld (15, ref+16+i); + + *vbuf = vec_sub( *vbuf, y); + + t1 = t2; + + vbuf++; + + t2 = vec_perm(tmp1, tmp2, align); + + y = vec_add(t1,vec_sld(t1,t2,4)); + y = vec_sub(vec_splat_s32(8),y); + + t1 = t2; + + x = vec_sl(*vbuf,vec_splat_u32(2)); + y = vec_sra(vec_sub(y,x),vec_splat_u32(4)); + *vbuf = vec_sub( *vbuf, y); + + vbuf++; + +#endif + } + + snow_horizontal_compose_liftS_lead_out(i, b, b, ref, width, w_l); + b[0] = b_0 - (((-2 * ref[1] + W_BO) - 4 * b_0) >> W_BS); + } + + { // Lift 3 + DWTELEM * const src = b+w2; + + vbuf = (vector signed int *)b; + vtmp = (vector signed int *)temp; + + i = 0; + align = vec_lvsl(0, src); + + for (; i<w_r-3; i+=4) { +#if 0 + temp[i] = src[i] - ((-3*(b[i] + b[i+1]))>>1); + temp[i+1] = src[i+1] - ((-3*(b[i+1] + b[i+2]))>>1); + temp[i+2] = src[i+2] - ((-3*(b[i+2] + b[i+3]))>>1); + temp[i+3] = src[i+3] - ((-3*(b[i+3] + b[i+4]))>>1); +#else + tmp1 = vec_ld(0,src+i); + t1 = vec_add(vbuf[0],vec_sld(vbuf[0],vbuf[1],4)); + tmp2 = vec_ld(15,src+i); + t1 = vec_sub(vec_splat_s32(0),t1); //bad! + t1 = vec_add(t1,vec_add(t1,t1)); + t2 = vec_perm(tmp1 ,tmp2 ,align); + t1 = vec_sra(t1,vec_splat_u32(1)); + vbuf++; + *vtmp = vec_sub(t2,t1); + vtmp++; + +#endif + + } + + snow_horizontal_compose_lift_lead_out(i, temp, src, b, width, w_r, 1, -3, 0, 1); + } + + { + //Interleave + int a; + vector signed int *t = (vector signed int *)temp, + *v = (vector signed int *)b; + + snow_interleave_line_header(&i, width, b, temp); + + for (; (i & 0xE) != 0xE; i-=2){ + b[i+1] = temp[i>>1]; + b[i] = b[i>>1]; + } + for (i-=14; i>=0; i-=16){ + a=i/4; + + v[a+3]=vec_mergel(v[(a>>1)+1],t[(a>>1)+1]); + v[a+2]=vec_mergeh(v[(a>>1)+1],t[(a>>1)+1]); + v[a+1]=vec_mergel(v[a>>1],t[a>>1]); + v[a]=vec_mergeh(v[a>>1],t[a>>1]); + + } + + } +} + +void ff_snow_vertical_compose97i_altivec(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, DWTELEM *b3, DWTELEM *b4, DWTELEM *b5, int width) +{ + int i, w4 = width/4; + vector signed int *v0, *v1,*v2,*v3,*v4,*v5; + vector signed int t1, t2; + + v0=(vector signed int *)b0; + v1=(vector signed int *)b1; + v2=(vector signed int *)b2; + v3=(vector signed int *)b3; + v4=(vector signed int *)b4; + v5=(vector signed int *)b5; + + for (i=0; i< w4;i++) + { + + #if 0 + b4[i] -= (3*(b3[i] + b5[i])+4)>>3; + b3[i] -= ((b2[i] + b4[i])); + b2[i] += ((b1[i] + b3[i])+4*b2[i]+8)>>4; + b1[i] += (3*(b0[i] + b2[i]))>>1; + #else + t1 = vec_add(v3[i], v5[i]); + t2 = vec_add(t1, vec_add(t1,t1)); + t1 = vec_add(t2, vec_splat_s32(4)); + v4[i] = vec_sub(v4[i], vec_sra(t1,vec_splat_u32(3))); + + v3[i] = vec_sub(v3[i], vec_add(v2[i], v4[i])); + + t1 = vec_add(vec_splat_s32(8), vec_add(v1[i], v3[i])); + t2 = vec_sl(v2[i], vec_splat_u32(2)); + v2[i] = vec_add(v2[i], vec_sra(vec_add(t1,t2),vec_splat_u32(4))); + t1 = vec_add(v0[i], v2[i]); + t2 = vec_add(t1, vec_add(t1,t1)); + v1[i] = vec_add(v1[i], vec_sra(t2,vec_splat_u32(1))); + + #endif + } + + for(i*=4; i < width; i++) + { + b4[i] -= (W_DM*(b3[i] + b5[i])+W_DO)>>W_DS; + b3[i] -= (W_CM*(b2[i] + b4[i])+W_CO)>>W_CS; + b2[i] += (W_BM*(b1[i] + b3[i])+4*b2[i]+W_BO)>>W_BS; + b1[i] += (W_AM*(b0[i] + b2[i])+W_AO)>>W_AS; + } +} + +#define LOAD_BLOCKS \ + tmp1 = vec_ld(0, &block[3][y*src_stride]);\ + align = vec_lvsl(0, &block[3][y*src_stride]);\ + tmp2 = vec_ld(15, &block[3][y*src_stride]);\ +\ + b3 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, &block[2][y*src_stride]);\ + align = vec_lvsl(0, &block[2][y*src_stride]);\ + tmp2 = vec_ld(15, &block[2][y*src_stride]);\ +\ + b2 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, &block[1][y*src_stride]);\ + align = vec_lvsl(0, &block[1][y*src_stride]);\ + tmp2 = vec_ld(15, &block[1][y*src_stride]);\ +\ + b1 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, &block[0][y*src_stride]);\ + align = vec_lvsl(0, &block[0][y*src_stride]);\ + tmp2 = vec_ld(15, &block[0][y*src_stride]);\ +\ + b0 = vec_perm(tmp1,tmp2,align); + +#define LOAD_OBMCS \ + tmp1 = vec_ld(0, obmc1);\ + align = vec_lvsl(0, obmc1);\ + tmp2 = vec_ld(15, obmc1);\ +\ + ob1 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, obmc2);\ + align = vec_lvsl(0, obmc2);\ + tmp2 = vec_ld(15, obmc2);\ +\ + ob2 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, obmc3);\ + align = vec_lvsl(0, obmc3);\ + tmp2 = vec_ld(15, obmc3);\ +\ + ob3 = vec_perm(tmp1,tmp2,align);\ +\ + tmp1 = vec_ld(0, obmc4);\ + align = vec_lvsl(0, obmc4);\ + tmp2 = vec_ld(15, obmc4);\ +\ + ob4 = vec_perm(tmp1,tmp2,align); + +/* interleave logic + * h1 <- [ a,b,a,b, a,b,a,b, a,b,a,b, a,b,a,b ] + * h2 <- [ c,d,c,d, c,d,c,d, c,d,c,d, c,d,c,d ] + * h <- [ a,b,c,d, a,b,c,d, a,b,c,d, a,b,c,d ] + */ + +#define STEPS_0_1\ + h1 = (vector unsigned short)\ + vec_mergeh(ob1, ob2);\ +\ + h2 = (vector unsigned short)\ + vec_mergeh(ob3, ob4);\ +\ + ih = (vector unsigned char)\ + vec_mergeh(h1,h2);\ +\ + l1 = (vector unsigned short) vec_mergeh(b3, b2);\ +\ + ih1 = (vector unsigned char) vec_mergel(h1, h2);\ +\ + l2 = (vector unsigned short) vec_mergeh(b1, b0);\ +\ + il = (vector unsigned char) vec_mergeh(l1, l2);\ +\ + v[0] = (vector signed int) vec_msum(ih, il, vec_splat_u32(0));\ +\ + il1 = (vector unsigned char) vec_mergel(l1, l2);\ +\ + v[1] = (vector signed int) vec_msum(ih1, il1, vec_splat_u32(0)); + +#define FINAL_STEP_SCALAR\ + for(x=0; x<b_w; x++)\ + if(add){\ + vbuf[x] += dst[x + src_x];\ + vbuf[x] = (vbuf[x] + (1<<(FRAC_BITS-1))) >> FRAC_BITS;\ + if(vbuf[x]&(~255)) vbuf[x]= ~(vbuf[x]>>31);\ + dst8[x + y*src_stride] = vbuf[x];\ + }else{\ + dst[x + src_x] -= vbuf[x];\ + } + +static void inner_add_yblock_bw_8_obmc_16_altivec(uint8_t *obmc, + const int obmc_stride, + uint8_t * * block, int b_w, + int b_h, int src_x, int src_y, + int src_stride, slice_buffer * sb, + int add, uint8_t * dst8) +{ + int y, x; + DWTELEM * dst; + vector unsigned short h1, h2, l1, l2; + vector unsigned char ih, il, ih1, il1, tmp1, tmp2, align; + vector unsigned char b0,b1,b2,b3; + vector unsigned char ob1,ob2,ob3,ob4; + + DECLARE_ALIGNED_16(int, vbuf[16]); + vector signed int *v = (vector signed int *)vbuf, *d; + + for(y=0; y<b_h; y++){ + //FIXME ugly missue of obmc_stride + + uint8_t *obmc1= obmc + y*obmc_stride; + uint8_t *obmc2= obmc1+ (obmc_stride>>1); + uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); + uint8_t *obmc4= obmc3+ (obmc_stride>>1); + + dst = slice_buffer_get_line(sb, src_y + y); + d = (vector signed int *)(dst + src_x); + +//FIXME i could avoid some loads! + + // load blocks + LOAD_BLOCKS + + // load obmcs + LOAD_OBMCS + + // steps 0 1 + STEPS_0_1 + + FINAL_STEP_SCALAR + + } + +} + +#define STEPS_2_3\ + h1 = (vector unsigned short) vec_mergel(ob1, ob2);\ +\ + h2 = (vector unsigned short) vec_mergel(ob3, ob4);\ +\ + ih = (vector unsigned char) vec_mergeh(h1,h2);\ +\ + l1 = (vector unsigned short) vec_mergel(b3, b2);\ +\ + l2 = (vector unsigned short) vec_mergel(b1, b0);\ +\ + ih1 = (vector unsigned char) vec_mergel(h1,h2);\ +\ + il = (vector unsigned char) vec_mergeh(l1,l2);\ +\ + v[2] = (vector signed int) vec_msum(ih, il, vec_splat_u32(0));\ +\ + il1 = (vector unsigned char) vec_mergel(l1,l2);\ +\ + v[3] = (vector signed int) vec_msum(ih1, il1, vec_splat_u32(0)); + + +static void inner_add_yblock_bw_16_obmc_32_altivec(uint8_t *obmc, + const int obmc_stride, + uint8_t * * block, int b_w, + int b_h, int src_x, int src_y, + int src_stride, slice_buffer * sb, + int add, uint8_t * dst8) +{ + int y, x; + DWTELEM * dst; + vector unsigned short h1, h2, l1, l2; + vector unsigned char ih, il, ih1, il1, tmp1, tmp2, align; + vector unsigned char b0,b1,b2,b3; + vector unsigned char ob1,ob2,ob3,ob4; + DECLARE_ALIGNED_16(int, vbuf[b_w]); + vector signed int *v = (vector signed int *)vbuf, *d; + + for(y=0; y<b_h; y++){ + //FIXME ugly missue of obmc_stride + + uint8_t *obmc1= obmc + y*obmc_stride; + uint8_t *obmc2= obmc1+ (obmc_stride>>1); + uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); + uint8_t *obmc4= obmc3+ (obmc_stride>>1); + + dst = slice_buffer_get_line(sb, src_y + y); + d = (vector signed int *)(dst + src_x); + + // load blocks + LOAD_BLOCKS + + // load obmcs + LOAD_OBMCS + + // steps 0 1 2 3 + STEPS_0_1 + + STEPS_2_3 + + FINAL_STEP_SCALAR + + } +} + +#define FINAL_STEP_VEC \ +\ + if(add)\ + {\ + for(x=0; x<b_w/4; x++)\ + {\ + v[x] = vec_add(v[x], d[x]);\ + v[x] = vec_sra(vec_add(v[x],\ + vec_sl( vec_splat_s32(1),\ + vec_splat_u32(7))),\ + vec_splat_u32(8));\ +\ + mask = (vector bool int) vec_sl((vector signed int)\ + vec_cmpeq(v[x],v[x]),vec_splat_u32(8));\ + mask = (vector bool int) vec_and(v[x],vec_nor(mask,mask));\ +\ + mask = (vector bool int)\ + vec_cmpeq((vector signed int)mask,\ + (vector signed int)vec_splat_u32(0));\ +\ + vs = vec_sra(v[x],vec_splat_u32(8));\ + vs = vec_sra(v[x],vec_splat_u32(8));\ + vs = vec_sra(v[x],vec_splat_u32(15));\ +\ + vs = vec_nor(vs,vs);\ +\ + v[x]= vec_sel(v[x],vs,mask);\ + }\ +\ + for(x=0; x<b_w; x++)\ + dst8[x + y*src_stride] = vbuf[x];\ +\ + }\ + else\ + for(x=0; x<b_w/4; x++)\ + d[x] = vec_sub(d[x], v[x]); + +static void inner_add_yblock_a_bw_8_obmc_16_altivec(uint8_t *obmc, + const int obmc_stride, + uint8_t * * block, int b_w, + int b_h, int src_x, int src_y, + int src_stride, slice_buffer * sb, + int add, uint8_t * dst8) +{ + int y, x; + DWTELEM * dst; + vector bool int mask; + vector signed int vs; + vector unsigned short h1, h2, l1, l2; + vector unsigned char ih, il, ih1, il1, tmp1, tmp2, align; + vector unsigned char b0,b1,b2,b3; + vector unsigned char ob1,ob2,ob3,ob4; + + DECLARE_ALIGNED_16(int, vbuf[16]); + vector signed int *v = (vector signed int *)vbuf, *d; + + for(y=0; y<b_h; y++){ + //FIXME ugly missue of obmc_stride + + uint8_t *obmc1= obmc + y*obmc_stride; + uint8_t *obmc2= obmc1+ (obmc_stride>>1); + uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); + uint8_t *obmc4= obmc3+ (obmc_stride>>1); + + dst = slice_buffer_get_line(sb, src_y + y); + d = (vector signed int *)(dst + src_x); + +//FIXME i could avoid some loads! + + // load blocks + LOAD_BLOCKS + + // load obmcs + LOAD_OBMCS + + // steps 0 1 + STEPS_0_1 + + FINAL_STEP_VEC + + } + +} + +static void inner_add_yblock_a_bw_16_obmc_32_altivec(uint8_t *obmc, + const int obmc_stride, + uint8_t * * block, int b_w, + int b_h, int src_x, int src_y, + int src_stride, slice_buffer * sb, + int add, uint8_t * dst8) +{ + int y, x; + DWTELEM * dst; + vector bool int mask; + vector signed int vs; + vector unsigned short h1, h2, l1, l2; + vector unsigned char ih, il, ih1, il1, tmp1, tmp2, align; + vector unsigned char b0,b1,b2,b3; + vector unsigned char ob1,ob2,ob3,ob4; + DECLARE_ALIGNED_16(int, vbuf[b_w]); + vector signed int *v = (vector signed int *)vbuf, *d; + + for(y=0; y<b_h; y++){ + //FIXME ugly missue of obmc_stride + + uint8_t *obmc1= obmc + y*obmc_stride; + uint8_t *obmc2= obmc1+ (obmc_stride>>1); + uint8_t *obmc3= obmc1+ obmc_stride*(obmc_stride>>1); + uint8_t *obmc4= obmc3+ (obmc_stride>>1); + + dst = slice_buffer_get_line(sb, src_y + y); + d = (vector signed int *)(dst + src_x); + + // load blocks + LOAD_BLOCKS + + // load obmcs + LOAD_OBMCS + + // steps 0 1 2 3 + STEPS_0_1 + + STEPS_2_3 + + FINAL_STEP_VEC + + } +} + + +void ff_snow_inner_add_yblock_altivec(uint8_t *obmc, const int obmc_stride, + uint8_t * * block, int b_w, int b_h, + int src_x, int src_y, int src_stride, + slice_buffer * sb, int add, + uint8_t * dst8) +{ + if (src_x&15) { + if (b_w == 16) + inner_add_yblock_bw_16_obmc_32_altivec(obmc, obmc_stride, block, + b_w, b_h, src_x, src_y, + src_stride, sb, add, dst8); + else if (b_w == 8) + inner_add_yblock_bw_8_obmc_16_altivec(obmc, obmc_stride, block, + b_w, b_h, src_x, src_y, + src_stride, sb, add, dst8); + else + ff_snow_inner_add_yblock(obmc, obmc_stride, block, b_w, b_h, src_x, + src_y, src_stride, sb, add, dst8); + } else { + if (b_w == 16) + inner_add_yblock_a_bw_16_obmc_32_altivec(obmc, obmc_stride, block, + b_w, b_h, src_x, src_y, + src_stride, sb, add, dst8); + else if (b_w == 8) + inner_add_yblock_a_bw_8_obmc_16_altivec(obmc, obmc_stride, block, + b_w, b_h, src_x, src_y, + src_stride, sb, add, dst8); + else + ff_snow_inner_add_yblock(obmc, obmc_stride, block, b_w, b_h, src_x, + src_y, src_stride, sb, add, dst8); + } +} + + +void snow_init_altivec(DSPContext* c, AVCodecContext *avctx) +{ + c->horizontal_compose97i = ff_snow_horizontal_compose97i_altivec; + c->vertical_compose97i = ff_snow_vertical_compose97i_altivec; + c->inner_add_yblock = ff_snow_inner_add_yblock_altivec; +} diff --git a/src/libffmpeg/libavcodec/ppc/types_altivec.h b/src/libffmpeg/libavcodec/ppc/types_altivec.h new file mode 100644 index 000000000..f29026e04 --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/types_altivec.h @@ -0,0 +1,41 @@ +/* + * Copyright (c) 2006 Guillaume Poirier <gpoirier@mplayerhq.hu> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/*********************************************************************** + * Vector types + **********************************************************************/ +#define vec_u8_t vector unsigned char +#define vec_s8_t vector signed char +#define vec_u16_t vector unsigned short +#define vec_s16_t vector signed short +#define vec_u32_t vector unsigned int +#define vec_s32_t vector signed int + +/*********************************************************************** + * Null vector + **********************************************************************/ +#define LOAD_ZERO const vec_u8_t zerov = vec_splat_u8( 0 ) + +#define zero_u8v (vec_u8_t) zerov +#define zero_s8v (vec_s8_t) zerov +#define zero_u16v (vec_u16_t) zerov +#define zero_s16v (vec_s16_t) zerov +#define zero_u32v (vec_u32_t) zerov +#define zero_s32v (vec_s32_t) zerov diff --git a/src/libffmpeg/libavcodec/ppc/vc1dsp_altivec.c b/src/libffmpeg/libavcodec/ppc/vc1dsp_altivec.c new file mode 100644 index 000000000..114c9d41f --- /dev/null +++ b/src/libffmpeg/libavcodec/ppc/vc1dsp_altivec.c @@ -0,0 +1,338 @@ +/* + * VC-1 and WMV3 decoder - DSP functions AltiVec-optimized + * Copyright (c) 2006 Konstantin Shishkov + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + * + */ + +#include "../dsputil.h" + +#include "gcc_fixes.h" + +#include "dsputil_altivec.h" + +// main steps of 8x8 transform +#define STEP8(s0, s1, s2, s3, s4, s5, s6, s7, vec_rnd) \ +do { \ + t0 = vec_sl(vec_add(s0, s4), vec_2); \ + t0 = vec_add(vec_sl(t0, vec_1), t0); \ + t0 = vec_add(t0, vec_rnd); \ + t1 = vec_sl(vec_sub(s0, s4), vec_2); \ + t1 = vec_add(vec_sl(t1, vec_1), t1); \ + t1 = vec_add(t1, vec_rnd); \ + t2 = vec_add(vec_sl(s6, vec_2), vec_sl(s6, vec_1)); \ + t2 = vec_add(t2, vec_sl(s2, vec_4)); \ + t3 = vec_add(vec_sl(s2, vec_2), vec_sl(s2, vec_1)); \ + t3 = vec_sub(t3, vec_sl(s6, vec_4)); \ + t4 = vec_add(t0, t2); \ + t5 = vec_add(t1, t3); \ + t6 = vec_sub(t1, t3); \ + t7 = vec_sub(t0, t2); \ +\ + t0 = vec_sl(vec_add(s1, s3), vec_4); \ + t0 = vec_add(t0, vec_sl(s5, vec_3)); \ + t0 = vec_add(t0, vec_sl(s7, vec_2)); \ + t0 = vec_add(t0, vec_sub(s5, s3)); \ +\ + t1 = vec_sl(vec_sub(s1, s5), vec_4); \ + t1 = vec_sub(t1, vec_sl(s7, vec_3)); \ + t1 = vec_sub(t1, vec_sl(s3, vec_2)); \ + t1 = vec_sub(t1, vec_add(s1, s7)); \ +\ + t2 = vec_sl(vec_sub(s7, s3), vec_4); \ + t2 = vec_add(t2, vec_sl(s1, vec_3)); \ + t2 = vec_add(t2, vec_sl(s5, vec_2)); \ + t2 = vec_add(t2, vec_sub(s1, s7)); \ +\ + t3 = vec_sl(vec_sub(s5, s7), vec_4); \ + t3 = vec_sub(t3, vec_sl(s3, vec_3)); \ + t3 = vec_add(t3, vec_sl(s1, vec_2)); \ + t3 = vec_sub(t3, vec_add(s3, s5)); \ +\ + s0 = vec_add(t4, t0); \ + s1 = vec_add(t5, t1); \ + s2 = vec_add(t6, t2); \ + s3 = vec_add(t7, t3); \ + s4 = vec_sub(t7, t3); \ + s5 = vec_sub(t6, t2); \ + s6 = vec_sub(t5, t1); \ + s7 = vec_sub(t4, t0); \ +}while(0) + +#define SHIFT_HOR8(s0, s1, s2, s3, s4, s5, s6, s7) \ +do { \ + s0 = vec_sra(s0, vec_3); \ + s1 = vec_sra(s1, vec_3); \ + s2 = vec_sra(s2, vec_3); \ + s3 = vec_sra(s3, vec_3); \ + s4 = vec_sra(s4, vec_3); \ + s5 = vec_sra(s5, vec_3); \ + s6 = vec_sra(s6, vec_3); \ + s7 = vec_sra(s7, vec_3); \ +}while(0) + +#define SHIFT_VERT8(s0, s1, s2, s3, s4, s5, s6, s7) \ +do { \ + s0 = vec_sra(s0, vec_7); \ + s1 = vec_sra(s1, vec_7); \ + s2 = vec_sra(s2, vec_7); \ + s3 = vec_sra(s3, vec_7); \ + s4 = vec_sra(vec_add(s4, vec_1s), vec_7); \ + s5 = vec_sra(vec_add(s5, vec_1s), vec_7); \ + s6 = vec_sra(vec_add(s6, vec_1s), vec_7); \ + s7 = vec_sra(vec_add(s7, vec_1s), vec_7); \ +}while(0) + +/* main steps of 4x4 transform */ +#define STEP4(s0, s1, s2, s3, vec_rnd) \ +do { \ + t1 = vec_add(vec_sl(s0, vec_4), s0); \ + t1 = vec_add(t1, vec_rnd); \ + t2 = vec_add(vec_sl(s2, vec_4), s2); \ + t0 = vec_add(t1, t2); \ + t1 = vec_sub(t1, t2); \ + t3 = vec_sl(vec_sub(s3, s1), vec_1); \ + t3 = vec_add(t3, vec_sl(t3, vec_2)); \ + t2 = vec_add(t3, vec_sl(s1, vec_5)); \ + t3 = vec_add(t3, vec_sl(s3, vec_3)); \ + t3 = vec_add(t3, vec_sl(s3, vec_2)); \ + s0 = vec_add(t0, t2); \ + s1 = vec_sub(t1, t3); \ + s2 = vec_add(t1, t3); \ + s3 = vec_sub(t0, t2); \ +}while (0) + +#define SHIFT_HOR4(s0, s1, s2, s3) \ + s0 = vec_sra(s0, vec_3); \ + s1 = vec_sra(s1, vec_3); \ + s2 = vec_sra(s2, vec_3); \ + s3 = vec_sra(s3, vec_3); + +#define SHIFT_VERT4(s0, s1, s2, s3) \ + s0 = vec_sra(s0, vec_7); \ + s1 = vec_sra(s1, vec_7); \ + s2 = vec_sra(s2, vec_7); \ + s3 = vec_sra(s3, vec_7); + +/** Do inverse transform on 8x8 block +*/ +static void vc1_inv_trans_8x8_altivec(DCTELEM block[64]) +{ + vector signed short src0, src1, src2, src3, src4, src5, src6, src7; + vector signed int s0, s1, s2, s3, s4, s5, s6, s7; + vector signed int s8, s9, sA, sB, sC, sD, sE, sF; + vector signed int t0, t1, t2, t3, t4, t5, t6, t7; + const vector signed int vec_64 = vec_sl(vec_splat_s32(4), vec_splat_u32(4)); + const vector unsigned int vec_7 = vec_splat_u32(7); + const vector unsigned int vec_5 = vec_splat_u32(5); + const vector unsigned int vec_4 = vec_splat_u32(4); + const vector signed int vec_4s = vec_splat_s32(4); + const vector unsigned int vec_3 = vec_splat_u32(3); + const vector unsigned int vec_2 = vec_splat_u32(2); + const vector signed int vec_1s = vec_splat_s32(1); + const vector unsigned int vec_1 = vec_splat_u32(1); + + + src0 = vec_ld( 0, block); + src1 = vec_ld( 16, block); + src2 = vec_ld( 32, block); + src3 = vec_ld( 48, block); + src4 = vec_ld( 64, block); + src5 = vec_ld( 80, block); + src6 = vec_ld( 96, block); + src7 = vec_ld(112, block); + + TRANSPOSE8(src0, src1, src2, src3, src4, src5, src6, src7); + s0 = vec_unpackl(src0); + s1 = vec_unpackl(src1); + s2 = vec_unpackl(src2); + s3 = vec_unpackl(src3); + s4 = vec_unpackl(src4); + s5 = vec_unpackl(src5); + s6 = vec_unpackl(src6); + s7 = vec_unpackl(src7); + s8 = vec_unpackh(src0); + s9 = vec_unpackh(src1); + sA = vec_unpackh(src2); + sB = vec_unpackh(src3); + sC = vec_unpackh(src4); + sD = vec_unpackh(src5); + sE = vec_unpackh(src6); + sF = vec_unpackh(src7); + STEP8(s0, s1, s2, s3, s4, s5, s6, s7, vec_4s); + SHIFT_HOR8(s0, s1, s2, s3, s4, s5, s6, s7); + STEP8(s8, s9, sA, sB, sC, sD, sE, sF, vec_4s); + SHIFT_HOR8(s8, s9, sA, sB, sC, sD, sE, sF); + src0 = vec_pack(s8, s0); + src1 = vec_pack(s9, s1); + src2 = vec_pack(sA, s2); + src3 = vec_pack(sB, s3); + src4 = vec_pack(sC, s4); + src5 = vec_pack(sD, s5); + src6 = vec_pack(sE, s6); + src7 = vec_pack(sF, s7); + TRANSPOSE8(src0, src1, src2, src3, src4, src5, src6, src7); + + s0 = vec_unpackl(src0); + s1 = vec_unpackl(src1); + s2 = vec_unpackl(src2); + s3 = vec_unpackl(src3); + s4 = vec_unpackl(src4); + s5 = vec_unpackl(src5); + s6 = vec_unpackl(src6); + s7 = vec_unpackl(src7); + s8 = vec_unpackh(src0); + s9 = vec_unpackh(src1); + sA = vec_unpackh(src2); + sB = vec_unpackh(src3); + sC = vec_unpackh(src4); + sD = vec_unpackh(src5); + sE = vec_unpackh(src6); + sF = vec_unpackh(src7); + STEP8(s0, s1, s2, s3, s4, s5, s6, s7, vec_64); + SHIFT_VERT8(s0, s1, s2, s3, s4, s5, s6, s7); + STEP8(s8, s9, sA, sB, sC, sD, sE, sF, vec_64); + SHIFT_VERT8(s8, s9, sA, sB, sC, sD, sE, sF); + src0 = vec_pack(s8, s0); + src1 = vec_pack(s9, s1); + src2 = vec_pack(sA, s2); + src3 = vec_pack(sB, s3); + src4 = vec_pack(sC, s4); + src5 = vec_pack(sD, s5); + src6 = vec_pack(sE, s6); + src7 = vec_pack(sF, s7); + + vec_st(src0, 0, block); + vec_st(src1, 16, block); + vec_st(src2, 32, block); + vec_st(src3, 48, block); + vec_st(src4, 64, block); + vec_st(src5, 80, block); + vec_st(src6, 96, block); + vec_st(src7,112, block); +} + +/** Do inverse transform on 8x4 part of block +*/ +static void vc1_inv_trans_8x4_altivec(DCTELEM block[64], int n) +{ + vector signed short src0, src1, src2, src3, src4, src5, src6, src7; + vector signed int s0, s1, s2, s3, s4, s5, s6, s7; + vector signed int s8, s9, sA, sB, sC, sD, sE, sF; + vector signed int t0, t1, t2, t3, t4, t5, t6, t7; + const vector signed int vec_64 = vec_sl(vec_splat_s32(4), vec_splat_u32(4)); + const vector unsigned int vec_7 = vec_splat_u32(7); + const vector unsigned int vec_5 = vec_splat_u32(5); + const vector unsigned int vec_4 = vec_splat_u32(4); + const vector signed int vec_4s = vec_splat_s32(4); + const vector unsigned int vec_3 = vec_splat_u32(3); + const vector unsigned int vec_2 = vec_splat_u32(2); + const vector unsigned int vec_1 = vec_splat_u32(1); + + src0 = vec_ld( 0, block); + src1 = vec_ld( 16, block); + src2 = vec_ld( 32, block); + src3 = vec_ld( 48, block); + src4 = vec_ld( 64, block); + src5 = vec_ld( 80, block); + src6 = vec_ld( 96, block); + src7 = vec_ld(112, block); + + TRANSPOSE8(src0, src1, src2, src3, src4, src5, src6, src7); + s0 = vec_unpackl(src0); + s1 = vec_unpackl(src1); + s2 = vec_unpackl(src2); + s3 = vec_unpackl(src3); + s4 = vec_unpackl(src4); + s5 = vec_unpackl(src5); + s6 = vec_unpackl(src6); + s7 = vec_unpackl(src7); + s8 = vec_unpackh(src0); + s9 = vec_unpackh(src1); + sA = vec_unpackh(src2); + sB = vec_unpackh(src3); + sC = vec_unpackh(src4); + sD = vec_unpackh(src5); + sE = vec_unpackh(src6); + sF = vec_unpackh(src7); + STEP8(s0, s1, s2, s3, s4, s5, s6, s7, vec_4s); + SHIFT_HOR8(s0, s1, s2, s3, s4, s5, s6, s7); + STEP8(s8, s9, sA, sB, sC, sD, sE, sF, vec_4s); + SHIFT_HOR8(s8, s9, sA, sB, sC, sD, sE, sF); + src0 = vec_pack(s8, s0); + src1 = vec_pack(s9, s1); + src2 = vec_pack(sA, s2); + src3 = vec_pack(sB, s3); + src4 = vec_pack(sC, s4); + src5 = vec_pack(sD, s5); + src6 = vec_pack(sE, s6); + src7 = vec_pack(sF, s7); + TRANSPOSE8(src0, src1, src2, src3, src4, src5, src6, src7); + + if(!n){ // upper half of block + s0 = vec_unpackh(src0); + s1 = vec_unpackh(src1); + s2 = vec_unpackh(src2); + s3 = vec_unpackh(src3); + s8 = vec_unpackl(src0); + s9 = vec_unpackl(src1); + sA = vec_unpackl(src2); + sB = vec_unpackl(src3); + STEP4(s0, s1, s2, s3, vec_64); + SHIFT_VERT4(s0, s1, s2, s3); + STEP4(s8, s9, sA, sB, vec_64); + SHIFT_VERT4(s8, s9, sA, sB); + src0 = vec_pack(s0, s8); + src1 = vec_pack(s1, s9); + src2 = vec_pack(s2, sA); + src3 = vec_pack(s3, sB); + + vec_st(src0, 0, block); + vec_st(src1, 16, block); + vec_st(src2, 32, block); + vec_st(src3, 48, block); + } else { //lower half of block + s0 = vec_unpackh(src4); + s1 = vec_unpackh(src5); + s2 = vec_unpackh(src6); + s3 = vec_unpackh(src7); + s8 = vec_unpackl(src4); + s9 = vec_unpackl(src5); + sA = vec_unpackl(src6); + sB = vec_unpackl(src7); + STEP4(s0, s1, s2, s3, vec_64); + SHIFT_VERT4(s0, s1, s2, s3); + STEP4(s8, s9, sA, sB, vec_64); + SHIFT_VERT4(s8, s9, sA, sB); + src4 = vec_pack(s0, s8); + src5 = vec_pack(s1, s9); + src6 = vec_pack(s2, sA); + src7 = vec_pack(s3, sB); + + vec_st(src4, 64, block); + vec_st(src5, 80, block); + vec_st(src6, 96, block); + vec_st(src7,112, block); + } +} + + +void vc1dsp_init_altivec(DSPContext* dsp, AVCodecContext *avctx) { + dsp->vc1_inv_trans_8x8 = vc1_inv_trans_8x8_altivec; + dsp->vc1_inv_trans_8x4 = vc1_inv_trans_8x4_altivec; +} diff --git a/src/libffmpeg/libavcodec/ratecontrol.h b/src/libffmpeg/libavcodec/ratecontrol.h new file mode 100644 index 000000000..c428923a5 --- /dev/null +++ b/src/libffmpeg/libavcodec/ratecontrol.h @@ -0,0 +1,103 @@ +/* + * Ratecontrol + * Copyright (c) 2000, 2001, 2002 Fabrice Bellard. + * Copyright (c) 2002-2004 Michael Niedermayer + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVCODEC_RATECONTROL_H +#define AVCODEC_RATECONTROL_H + +/** + * @file ratecontrol.h + * ratecontrol header. + */ + +#include "eval.h" + +typedef struct Predictor{ + double coeff; + double count; + double decay; +} Predictor; + +typedef struct RateControlEntry{ + int pict_type; + float qscale; + int mv_bits; + int i_tex_bits; + int p_tex_bits; + int misc_bits; + int header_bits; + uint64_t expected_bits; + int new_pict_type; + float new_qscale; + int mc_mb_var_sum; + int mb_var_sum; + int i_count; + int skip_count; + int f_code; + int b_code; +}RateControlEntry; + +/** + * rate control context. + */ +typedef struct RateControlContext{ + FILE *stats_file; + int num_entries; ///< number of RateControlEntries + RateControlEntry *entry; + double buffer_index; ///< amount of bits in the video/audio buffer + Predictor pred[5]; + double short_term_qsum; ///< sum of recent qscales + double short_term_qcount; ///< count of recent qscales + double pass1_rc_eq_output_sum;///< sum of the output of the rc equation, this is used for normalization + double pass1_wanted_bits; ///< bits which should have been outputed by the pass1 code (including complexity init) + double last_qscale; + double last_qscale_for[5]; ///< last qscale for a specific pict type, used for max_diff & ipb factor stuff + int last_mc_mb_var_sum; + int last_mb_var_sum; + uint64_t i_cplx_sum[5]; + uint64_t p_cplx_sum[5]; + uint64_t mv_bits_sum[5]; + uint64_t qscale_sum[5]; + int frame_count[5]; + int last_non_b_pict_type; + + void *non_lavc_opaque; ///< context for non lavc rc code (for example xvid) + float dry_run_qscale; ///< for xvid rc + int last_picture_number; ///< for xvid rc + AVEvalExpr * rc_eq_eval; +}RateControlContext; + +struct MpegEncContext; + +/* rate control */ +int ff_rate_control_init(struct MpegEncContext *s); +float ff_rate_estimate_qscale(struct MpegEncContext *s, int dry_run); +void ff_write_pass1_stats(struct MpegEncContext *s); +void ff_rate_control_uninit(struct MpegEncContext *s); +int ff_vbv_update(struct MpegEncContext *s, int frame_size); +void ff_get_2pass_fcode(struct MpegEncContext *s); + +int ff_xvid_rate_control_init(struct MpegEncContext *s); +void ff_xvid_rate_control_uninit(struct MpegEncContext *s); +float ff_xvid_rate_estimate_qscale(struct MpegEncContext *s, int dry_run); + +#endif /* AVCODEC_RATECONTROL_H */ + diff --git a/src/libffmpeg/libavcodec/sparc/.hgignore b/src/libffmpeg/libavcodec/sparc/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavcodec/sparc/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavcodec/vorbis_data.c b/src/libffmpeg/libavcodec/vorbis_data.c new file mode 100644 index 000000000..5dc9c5f00 --- /dev/null +++ b/src/libffmpeg/libavcodec/vorbis_data.c @@ -0,0 +1,2155 @@ +/* + * copyright (c) 2005 Denes Balatoni ( dbalatoni programozo hu ) + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "vorbis.h" + +static const float vwin64[32] = { + 0.0009460463F, 0.0085006468F, 0.0235352254F, 0.0458950567F, + 0.0753351908F, 0.1115073077F, 0.1539457973F, 0.2020557475F, + 0.2551056759F, 0.3122276645F, 0.3724270287F, 0.4346027792F, + 0.4975789974F, 0.5601459521F, 0.6211085051F, 0.6793382689F, + 0.7338252629F, 0.7837245849F, 0.8283939355F, 0.8674186656F, + 0.9006222429F, 0.9280614787F, 0.9500073081F, 0.9669131782F, + 0.9793740220F, 0.9880792941F, 0.9937636139F, 0.9971582668F, + 0.9989462667F, 0.9997230082F, 0.9999638688F, 0.9999995525F, +}; + +static const float vwin128[64] = { + 0.0002365472F, 0.0021280687F, 0.0059065254F, 0.0115626550F, + 0.0190823442F, 0.0284463735F, 0.0396300935F, 0.0526030430F, + 0.0673285281F, 0.0837631763F, 0.1018564887F, 0.1215504095F, + 0.1427789367F, 0.1654677960F, 0.1895342001F, 0.2148867160F, + 0.2414252576F, 0.2690412240F, 0.2976177952F, 0.3270303960F, + 0.3571473350F, 0.3878306189F, 0.4189369387F, 0.4503188188F, + 0.4818259135F, 0.5133064334F, 0.5446086751F, 0.5755826278F, + 0.6060816248F, 0.6359640047F, 0.6650947483F, 0.6933470543F, + 0.7206038179F, 0.7467589810F, 0.7717187213F, 0.7954024542F, + 0.8177436264F, 0.8386902831F, 0.8582053981F, 0.8762669622F, + 0.8928678298F, 0.9080153310F, 0.9217306608F, 0.9340480615F, + 0.9450138200F, 0.9546851041F, 0.9631286621F, 0.9704194171F, + 0.9766389810F, 0.9818741197F, 0.9862151938F, 0.9897546035F, + 0.9925852598F, 0.9947991032F, 0.9964856900F, 0.9977308602F, + 0.9986155015F, 0.9992144193F, 0.9995953200F, 0.9998179155F, + 0.9999331503F, 0.9999825563F, 0.9999977357F, 0.9999999720F, +}; + +static const float vwin256[128] = { + 0.0000591390F, 0.0005321979F, 0.0014780301F, 0.0028960636F, + 0.0047854363F, 0.0071449926F, 0.0099732775F, 0.0132685298F, + 0.0170286741F, 0.0212513119F, 0.0259337111F, 0.0310727950F, + 0.0366651302F, 0.0427069140F, 0.0491939614F, 0.0561216907F, + 0.0634851102F, 0.0712788035F, 0.0794969160F, 0.0881331402F, + 0.0971807028F, 0.1066323515F, 0.1164803426F, 0.1267164297F, + 0.1373318534F, 0.1483173323F, 0.1596630553F, 0.1713586755F, + 0.1833933062F, 0.1957555184F, 0.2084333404F, 0.2214142599F, + 0.2346852280F, 0.2482326664F, 0.2620424757F, 0.2761000481F, + 0.2903902813F, 0.3048975959F, 0.3196059553F, 0.3344988887F, + 0.3495595160F, 0.3647705766F, 0.3801144597F, 0.3955732382F, + 0.4111287047F, 0.4267624093F, 0.4424557009F, 0.4581897696F, + 0.4739456913F, 0.4897044744F, 0.5054471075F, 0.5211546088F, + 0.5368080763F, 0.5523887395F, 0.5678780103F, 0.5832575361F, + 0.5985092508F, 0.6136154277F, 0.6285587300F, 0.6433222619F, + 0.6578896175F, 0.6722449294F, 0.6863729144F, 0.7002589187F, + 0.7138889597F, 0.7272497662F, 0.7403288154F, 0.7531143679F, + 0.7655954985F, 0.7777621249F, 0.7896050322F, 0.8011158947F, + 0.8122872932F, 0.8231127294F, 0.8335866365F, 0.8437043850F, + 0.8534622861F, 0.8628575905F, 0.8718884835F, 0.8805540765F, + 0.8888543947F, 0.8967903616F, 0.9043637797F, 0.9115773078F, + 0.9184344360F, 0.9249394562F, 0.9310974312F, 0.9369141608F, + 0.9423961446F, 0.9475505439F, 0.9523851406F, 0.9569082947F, + 0.9611289005F, 0.9650563408F, 0.9687004405F, 0.9720714191F, + 0.9751798427F, 0.9780365753F, 0.9806527301F, 0.9830396204F, + 0.9852087111F, 0.9871715701F, 0.9889398207F, 0.9905250941F, + 0.9919389832F, 0.9931929973F, 0.9942985174F, 0.9952667537F, + 0.9961087037F, 0.9968351119F, 0.9974564312F, 0.9979827858F, + 0.9984239359F, 0.9987892441F, 0.9990876435F, 0.9993276081F, + 0.9995171241F, 0.9996636648F, 0.9997741654F, 0.9998550016F, + 0.9999119692F, 0.9999502656F, 0.9999744742F, 0.9999885497F, + 0.9999958064F, 0.9999989077F, 0.9999998584F, 0.9999999983F, +}; + +static const float vwin512[256] = { + 0.0000147849F, 0.0001330607F, 0.0003695946F, 0.0007243509F, + 0.0011972759F, 0.0017882983F, 0.0024973285F, 0.0033242588F, + 0.0042689632F, 0.0053312973F, 0.0065110982F, 0.0078081841F, + 0.0092223540F, 0.0107533880F, 0.0124010466F, 0.0141650703F, + 0.0160451800F, 0.0180410758F, 0.0201524373F, 0.0223789233F, + 0.0247201710F, 0.0271757958F, 0.0297453914F, 0.0324285286F, + 0.0352247556F, 0.0381335972F, 0.0411545545F, 0.0442871045F, + 0.0475306997F, 0.0508847676F, 0.0543487103F, 0.0579219038F, + 0.0616036982F, 0.0653934164F, 0.0692903546F, 0.0732937809F, + 0.0774029356F, 0.0816170305F, 0.0859352485F, 0.0903567428F, + 0.0948806375F, 0.0995060259F, 0.1042319712F, 0.1090575056F, + 0.1139816300F, 0.1190033137F, 0.1241214941F, 0.1293350764F, + 0.1346429333F, 0.1400439046F, 0.1455367974F, 0.1511203852F, + 0.1567934083F, 0.1625545735F, 0.1684025537F, 0.1743359881F, + 0.1803534820F, 0.1864536069F, 0.1926349000F, 0.1988958650F, + 0.2052349715F, 0.2116506555F, 0.2181413191F, 0.2247053313F, + 0.2313410275F, 0.2380467105F, 0.2448206500F, 0.2516610835F, + 0.2585662164F, 0.2655342226F, 0.2725632448F, 0.2796513950F, + 0.2867967551F, 0.2939973773F, 0.3012512852F, 0.3085564739F, + 0.3159109111F, 0.3233125375F, 0.3307592680F, 0.3382489922F, + 0.3457795756F, 0.3533488602F, 0.3609546657F, 0.3685947904F, + 0.3762670121F, 0.3839690896F, 0.3916987634F, 0.3994537572F, + 0.4072317788F, 0.4150305215F, 0.4228476653F, 0.4306808783F, + 0.4385278181F, 0.4463861329F, 0.4542534630F, 0.4621274424F, + 0.4700057001F, 0.4778858615F, 0.4857655502F, 0.4936423891F, + 0.5015140023F, 0.5093780165F, 0.5172320626F, 0.5250737772F, + 0.5329008043F, 0.5407107971F, 0.5485014192F, 0.5562703465F, + 0.5640152688F, 0.5717338914F, 0.5794239366F, 0.5870831457F, + 0.5947092801F, 0.6023001235F, 0.6098534829F, 0.6173671907F, + 0.6248391059F, 0.6322671161F, 0.6396491384F, 0.6469831217F, + 0.6542670475F, 0.6614989319F, 0.6686768267F, 0.6757988210F, + 0.6828630426F, 0.6898676592F, 0.6968108799F, 0.7036909564F, + 0.7105061843F, 0.7172549043F, 0.7239355032F, 0.7305464154F, + 0.7370861235F, 0.7435531598F, 0.7499461068F, 0.7562635986F, + 0.7625043214F, 0.7686670148F, 0.7747504721F, 0.7807535410F, + 0.7866751247F, 0.7925141825F, 0.7982697296F, 0.8039408387F, + 0.8095266395F, 0.8150263196F, 0.8204391248F, 0.8257643590F, + 0.8310013848F, 0.8361496236F, 0.8412085555F, 0.8461777194F, + 0.8510567129F, 0.8558451924F, 0.8605428730F, 0.8651495278F, + 0.8696649882F, 0.8740891432F, 0.8784219392F, 0.8826633797F, + 0.8868135244F, 0.8908724888F, 0.8948404441F, 0.8987176157F, + 0.9025042831F, 0.9062007791F, 0.9098074886F, 0.9133248482F, + 0.9167533451F, 0.9200935163F, 0.9233459472F, 0.9265112712F, + 0.9295901680F, 0.9325833632F, 0.9354916263F, 0.9383157705F, + 0.9410566504F, 0.9437151618F, 0.9462922398F, 0.9487888576F, + 0.9512060252F, 0.9535447882F, 0.9558062262F, 0.9579914516F, + 0.9601016078F, 0.9621378683F, 0.9641014348F, 0.9659935361F, + 0.9678154261F, 0.9695683830F, 0.9712537071F, 0.9728727198F, + 0.9744267618F, 0.9759171916F, 0.9773453842F, 0.9787127293F, + 0.9800206298F, 0.9812705006F, 0.9824637665F, 0.9836018613F, + 0.9846862258F, 0.9857183066F, 0.9866995544F, 0.9876314227F, + 0.9885153662F, 0.9893528393F, 0.9901452948F, 0.9908941823F, + 0.9916009470F, 0.9922670279F, 0.9928938570F, 0.9934828574F, + 0.9940354423F, 0.9945530133F, 0.9950369595F, 0.9954886562F, + 0.9959094633F, 0.9963007242F, 0.9966637649F, 0.9969998925F, + 0.9973103939F, 0.9975965351F, 0.9978595598F, 0.9981006885F, + 0.9983211172F, 0.9985220166F, 0.9987045311F, 0.9988697776F, + 0.9990188449F, 0.9991527924F, 0.9992726499F, 0.9993794157F, + 0.9994740570F, 0.9995575079F, 0.9996306699F, 0.9996944099F, + 0.9997495605F, 0.9997969190F, 0.9998372465F, 0.9998712678F, + 0.9998996704F, 0.9999231041F, 0.9999421807F, 0.9999574732F, + 0.9999695157F, 0.9999788026F, 0.9999857885F, 0.9999908879F, + 0.9999944746F, 0.9999968817F, 0.9999984010F, 0.9999992833F, + 0.9999997377F, 0.9999999317F, 0.9999999911F, 0.9999999999F, +}; + +static const float vwin1024[512] = { + 0.0000036962F, 0.0000332659F, 0.0000924041F, 0.0001811086F, + 0.0002993761F, 0.0004472021F, 0.0006245811F, 0.0008315063F, + 0.0010679699F, 0.0013339631F, 0.0016294757F, 0.0019544965F, + 0.0023090133F, 0.0026930125F, 0.0031064797F, 0.0035493989F, + 0.0040217533F, 0.0045235250F, 0.0050546946F, 0.0056152418F, + 0.0062051451F, 0.0068243817F, 0.0074729278F, 0.0081507582F, + 0.0088578466F, 0.0095941655F, 0.0103596863F, 0.0111543789F, + 0.0119782122F, 0.0128311538F, 0.0137131701F, 0.0146242260F, + 0.0155642855F, 0.0165333111F, 0.0175312640F, 0.0185581042F, + 0.0196137903F, 0.0206982797F, 0.0218115284F, 0.0229534910F, + 0.0241241208F, 0.0253233698F, 0.0265511886F, 0.0278075263F, + 0.0290923308F, 0.0304055484F, 0.0317471241F, 0.0331170013F, + 0.0345151222F, 0.0359414274F, 0.0373958560F, 0.0388783456F, + 0.0403888325F, 0.0419272511F, 0.0434935347F, 0.0450876148F, + 0.0467094213F, 0.0483588828F, 0.0500359261F, 0.0517404765F, + 0.0534724575F, 0.0552317913F, 0.0570183983F, 0.0588321971F, + 0.0606731048F, 0.0625410369F, 0.0644359070F, 0.0663576272F, + 0.0683061077F, 0.0702812571F, 0.0722829821F, 0.0743111878F, + 0.0763657775F, 0.0784466526F, 0.0805537129F, 0.0826868561F, + 0.0848459782F, 0.0870309736F, 0.0892417345F, 0.0914781514F, + 0.0937401128F, 0.0960275056F, 0.0983402145F, 0.1006781223F, + 0.1030411101F, 0.1054290568F, 0.1078418397F, 0.1102793336F, + 0.1127414119F, 0.1152279457F, 0.1177388042F, 0.1202738544F, + 0.1228329618F, 0.1254159892F, 0.1280227980F, 0.1306532471F, + 0.1333071937F, 0.1359844927F, 0.1386849970F, 0.1414085575F, + 0.1441550230F, 0.1469242403F, 0.1497160539F, 0.1525303063F, + 0.1553668381F, 0.1582254875F, 0.1611060909F, 0.1640084822F, + 0.1669324936F, 0.1698779549F, 0.1728446939F, 0.1758325362F, + 0.1788413055F, 0.1818708232F, 0.1849209084F, 0.1879913785F, + 0.1910820485F, 0.1941927312F, 0.1973232376F, 0.2004733764F, + 0.2036429541F, 0.2068317752F, 0.2100396421F, 0.2132663552F, + 0.2165117125F, 0.2197755102F, 0.2230575422F, 0.2263576007F, + 0.2296754753F, 0.2330109540F, 0.2363638225F, 0.2397338646F, + 0.2431208619F, 0.2465245941F, 0.2499448389F, 0.2533813719F, + 0.2568339669F, 0.2603023956F, 0.2637864277F, 0.2672858312F, + 0.2708003718F, 0.2743298135F, 0.2778739186F, 0.2814324472F, + 0.2850051576F, 0.2885918065F, 0.2921921485F, 0.2958059366F, + 0.2994329219F, 0.3030728538F, 0.3067254799F, 0.3103905462F, + 0.3140677969F, 0.3177569747F, 0.3214578205F, 0.3251700736F, + 0.3288934718F, 0.3326277513F, 0.3363726468F, 0.3401278914F, + 0.3438932168F, 0.3476683533F, 0.3514530297F, 0.3552469734F, + 0.3590499106F, 0.3628615659F, 0.3666816630F, 0.3705099239F, + 0.3743460698F, 0.3781898204F, 0.3820408945F, 0.3858990095F, + 0.3897638820F, 0.3936352274F, 0.3975127601F, 0.4013961936F, + 0.4052852405F, 0.4091796123F, 0.4130790198F, 0.4169831732F, + 0.4208917815F, 0.4248045534F, 0.4287211965F, 0.4326414181F, + 0.4365649248F, 0.4404914225F, 0.4444206167F, 0.4483522125F, + 0.4522859146F, 0.4562214270F, 0.4601584538F, 0.4640966984F, + 0.4680358644F, 0.4719756548F, 0.4759157726F, 0.4798559209F, + 0.4837958024F, 0.4877351199F, 0.4916735765F, 0.4956108751F, + 0.4995467188F, 0.5034808109F, 0.5074128550F, 0.5113425550F, + 0.5152696149F, 0.5191937395F, 0.5231146336F, 0.5270320028F, + 0.5309455530F, 0.5348549910F, 0.5387600239F, 0.5426603597F, + 0.5465557070F, 0.5504457754F, 0.5543302752F, 0.5582089175F, + 0.5620814145F, 0.5659474793F, 0.5698068262F, 0.5736591704F, + 0.5775042283F, 0.5813417176F, 0.5851713571F, 0.5889928670F, + 0.5928059689F, 0.5966103856F, 0.6004058415F, 0.6041920626F, + 0.6079687761F, 0.6117357113F, 0.6154925986F, 0.6192391705F, + 0.6229751612F, 0.6267003064F, 0.6304143441F, 0.6341170137F, + 0.6378080569F, 0.6414872173F, 0.6451542405F, 0.6488088741F, + 0.6524508681F, 0.6560799742F, 0.6596959469F, 0.6632985424F, + 0.6668875197F, 0.6704626398F, 0.6740236662F, 0.6775703649F, + 0.6811025043F, 0.6846198554F, 0.6881221916F, 0.6916092892F, + 0.6950809269F, 0.6985368861F, 0.7019769510F, 0.7054009085F, + 0.7088085484F, 0.7121996632F, 0.7155740484F, 0.7189315023F, + 0.7222718263F, 0.7255948245F, 0.7289003043F, 0.7321880760F, + 0.7354579530F, 0.7387097518F, 0.7419432921F, 0.7451583966F, + 0.7483548915F, 0.7515326059F, 0.7546913723F, 0.7578310265F, + 0.7609514077F, 0.7640523581F, 0.7671337237F, 0.7701953535F, + 0.7732371001F, 0.7762588195F, 0.7792603711F, 0.7822416178F, + 0.7852024259F, 0.7881426654F, 0.7910622097F, 0.7939609356F, + 0.7968387237F, 0.7996954579F, 0.8025310261F, 0.8053453193F, + 0.8081382324F, 0.8109096638F, 0.8136595156F, 0.8163876936F, + 0.8190941071F, 0.8217786690F, 0.8244412960F, 0.8270819086F, + 0.8297004305F, 0.8322967896F, 0.8348709171F, 0.8374227481F, + 0.8399522213F, 0.8424592789F, 0.8449438672F, 0.8474059356F, + 0.8498454378F, 0.8522623306F, 0.8546565748F, 0.8570281348F, + 0.8593769787F, 0.8617030779F, 0.8640064080F, 0.8662869477F, + 0.8685446796F, 0.8707795899F, 0.8729916682F, 0.8751809079F, + 0.8773473059F, 0.8794908626F, 0.8816115819F, 0.8837094713F, + 0.8857845418F, 0.8878368079F, 0.8898662874F, 0.8918730019F, + 0.8938569760F, 0.8958182380F, 0.8977568194F, 0.8996727552F, + 0.9015660837F, 0.9034368465F, 0.9052850885F, 0.9071108577F, + 0.9089142057F, 0.9106951869F, 0.9124538591F, 0.9141902832F, + 0.9159045233F, 0.9175966464F, 0.9192667228F, 0.9209148257F, + 0.9225410313F, 0.9241454187F, 0.9257280701F, 0.9272890704F, + 0.9288285075F, 0.9303464720F, 0.9318430576F, 0.9333183603F, + 0.9347724792F, 0.9362055158F, 0.9376175745F, 0.9390087622F, + 0.9403791881F, 0.9417289644F, 0.9430582055F, 0.9443670283F, + 0.9456555521F, 0.9469238986F, 0.9481721917F, 0.9494005577F, + 0.9506091252F, 0.9517980248F, 0.9529673894F, 0.9541173540F, + 0.9552480557F, 0.9563596334F, 0.9574522282F, 0.9585259830F, + 0.9595810428F, 0.9606175542F, 0.9616356656F, 0.9626355274F, + 0.9636172915F, 0.9645811114F, 0.9655271425F, 0.9664555414F, + 0.9673664664F, 0.9682600774F, 0.9691365355F, 0.9699960034F, + 0.9708386448F, 0.9716646250F, 0.9724741103F, 0.9732672685F, + 0.9740442683F, 0.9748052795F, 0.9755504729F, 0.9762800205F, + 0.9769940950F, 0.9776928703F, 0.9783765210F, 0.9790452223F, + 0.9796991504F, 0.9803384823F, 0.9809633954F, 0.9815740679F, + 0.9821706784F, 0.9827534063F, 0.9833224312F, 0.9838779332F, + 0.9844200928F, 0.9849490910F, 0.9854651087F, 0.9859683274F, + 0.9864589286F, 0.9869370940F, 0.9874030054F, 0.9878568447F, + 0.9882987937F, 0.9887290343F, 0.9891477481F, 0.9895551169F, + 0.9899513220F, 0.9903365446F, 0.9907109658F, 0.9910747662F, + 0.9914281260F, 0.9917712252F, 0.9921042433F, 0.9924273593F, + 0.9927407516F, 0.9930445982F, 0.9933390763F, 0.9936243626F, + 0.9939006331F, 0.9941680631F, 0.9944268269F, 0.9946770982F, + 0.9949190498F, 0.9951528537F, 0.9953786808F, 0.9955967011F, + 0.9958070836F, 0.9960099963F, 0.9962056061F, 0.9963940787F, + 0.9965755786F, 0.9967502693F, 0.9969183129F, 0.9970798704F, + 0.9972351013F, 0.9973841640F, 0.9975272151F, 0.9976644103F, + 0.9977959036F, 0.9979218476F, 0.9980423932F, 0.9981576901F, + 0.9982678862F, 0.9983731278F, 0.9984735596F, 0.9985693247F, + 0.9986605645F, 0.9987474186F, 0.9988300248F, 0.9989085193F, + 0.9989830364F, 0.9990537085F, 0.9991206662F, 0.9991840382F, + 0.9992439513F, 0.9993005303F, 0.9993538982F, 0.9994041757F, + 0.9994514817F, 0.9994959330F, 0.9995376444F, 0.9995767286F, + 0.9996132960F, 0.9996474550F, 0.9996793121F, 0.9997089710F, + 0.9997365339F, 0.9997621003F, 0.9997857677F, 0.9998076311F, + 0.9998277836F, 0.9998463156F, 0.9998633155F, 0.9998788692F, + 0.9998930603F, 0.9999059701F, 0.9999176774F, 0.9999282586F, + 0.9999377880F, 0.9999463370F, 0.9999539749F, 0.9999607685F, + 0.9999667820F, 0.9999720773F, 0.9999767136F, 0.9999807479F, + 0.9999842344F, 0.9999872249F, 0.9999897688F, 0.9999919127F, + 0.9999937009F, 0.9999951749F, 0.9999963738F, 0.9999973342F, + 0.9999980900F, 0.9999986724F, 0.9999991103F, 0.9999994297F, + 0.9999996543F, 0.9999998049F, 0.9999999000F, 0.9999999552F, + 0.9999999836F, 0.9999999957F, 0.9999999994F, 1.0000000000F, +}; + +static const float vwin2048[1024] = { + 0.0000009241F, 0.0000083165F, 0.0000231014F, 0.0000452785F, + 0.0000748476F, 0.0001118085F, 0.0001561608F, 0.0002079041F, + 0.0002670379F, 0.0003335617F, 0.0004074748F, 0.0004887765F, + 0.0005774661F, 0.0006735427F, 0.0007770054F, 0.0008878533F, + 0.0010060853F, 0.0011317002F, 0.0012646969F, 0.0014050742F, + 0.0015528307F, 0.0017079650F, 0.0018704756F, 0.0020403610F, + 0.0022176196F, 0.0024022497F, 0.0025942495F, 0.0027936173F, + 0.0030003511F, 0.0032144490F, 0.0034359088F, 0.0036647286F, + 0.0039009061F, 0.0041444391F, 0.0043953253F, 0.0046535621F, + 0.0049191472F, 0.0051920781F, 0.0054723520F, 0.0057599664F, + 0.0060549184F, 0.0063572052F, 0.0066668239F, 0.0069837715F, + 0.0073080449F, 0.0076396410F, 0.0079785566F, 0.0083247884F, + 0.0086783330F, 0.0090391871F, 0.0094073470F, 0.0097828092F, + 0.0101655700F, 0.0105556258F, 0.0109529726F, 0.0113576065F, + 0.0117695237F, 0.0121887200F, 0.0126151913F, 0.0130489335F, + 0.0134899422F, 0.0139382130F, 0.0143937415F, 0.0148565233F, + 0.0153265536F, 0.0158038279F, 0.0162883413F, 0.0167800889F, + 0.0172790660F, 0.0177852675F, 0.0182986882F, 0.0188193231F, + 0.0193471668F, 0.0198822141F, 0.0204244594F, 0.0209738974F, + 0.0215305225F, 0.0220943289F, 0.0226653109F, 0.0232434627F, + 0.0238287784F, 0.0244212519F, 0.0250208772F, 0.0256276481F, + 0.0262415582F, 0.0268626014F, 0.0274907711F, 0.0281260608F, + 0.0287684638F, 0.0294179736F, 0.0300745833F, 0.0307382859F, + 0.0314090747F, 0.0320869424F, 0.0327718819F, 0.0334638860F, + 0.0341629474F, 0.0348690586F, 0.0355822122F, 0.0363024004F, + 0.0370296157F, 0.0377638502F, 0.0385050960F, 0.0392533451F, + 0.0400085896F, 0.0407708211F, 0.0415400315F, 0.0423162123F, + 0.0430993552F, 0.0438894515F, 0.0446864926F, 0.0454904698F, + 0.0463013742F, 0.0471191969F, 0.0479439288F, 0.0487755607F, + 0.0496140836F, 0.0504594879F, 0.0513117642F, 0.0521709031F, + 0.0530368949F, 0.0539097297F, 0.0547893979F, 0.0556758894F, + 0.0565691941F, 0.0574693019F, 0.0583762026F, 0.0592898858F, + 0.0602103410F, 0.0611375576F, 0.0620715250F, 0.0630122324F, + 0.0639596688F, 0.0649138234F, 0.0658746848F, 0.0668422421F, + 0.0678164838F, 0.0687973985F, 0.0697849746F, 0.0707792005F, + 0.0717800645F, 0.0727875547F, 0.0738016591F, 0.0748223656F, + 0.0758496620F, 0.0768835359F, 0.0779239751F, 0.0789709668F, + 0.0800244985F, 0.0810845574F, 0.0821511306F, 0.0832242052F, + 0.0843037679F, 0.0853898056F, 0.0864823050F, 0.0875812525F, + 0.0886866347F, 0.0897984378F, 0.0909166480F, 0.0920412513F, + 0.0931722338F, 0.0943095813F, 0.0954532795F, 0.0966033140F, + 0.0977596702F, 0.0989223336F, 0.1000912894F, 0.1012665227F, + 0.1024480185F, 0.1036357616F, 0.1048297369F, 0.1060299290F, + 0.1072363224F, 0.1084489014F, 0.1096676504F, 0.1108925534F, + 0.1121235946F, 0.1133607577F, 0.1146040267F, 0.1158533850F, + 0.1171088163F, 0.1183703040F, 0.1196378312F, 0.1209113812F, + 0.1221909370F, 0.1234764815F, 0.1247679974F, 0.1260654674F, + 0.1273688740F, 0.1286781995F, 0.1299934263F, 0.1313145365F, + 0.1326415121F, 0.1339743349F, 0.1353129866F, 0.1366574490F, + 0.1380077035F, 0.1393637315F, 0.1407255141F, 0.1420930325F, + 0.1434662677F, 0.1448452004F, 0.1462298115F, 0.1476200814F, + 0.1490159906F, 0.1504175195F, 0.1518246482F, 0.1532373569F, + 0.1546556253F, 0.1560794333F, 0.1575087606F, 0.1589435866F, + 0.1603838909F, 0.1618296526F, 0.1632808509F, 0.1647374648F, + 0.1661994731F, 0.1676668546F, 0.1691395880F, 0.1706176516F, + 0.1721010238F, 0.1735896829F, 0.1750836068F, 0.1765827736F, + 0.1780871610F, 0.1795967468F, 0.1811115084F, 0.1826314234F, + 0.1841564689F, 0.1856866221F, 0.1872218600F, 0.1887621595F, + 0.1903074974F, 0.1918578503F, 0.1934131947F, 0.1949735068F, + 0.1965387630F, 0.1981089393F, 0.1996840117F, 0.2012639560F, + 0.2028487479F, 0.2044383630F, 0.2060327766F, 0.2076319642F, + 0.2092359007F, 0.2108445614F, 0.2124579211F, 0.2140759545F, + 0.2156986364F, 0.2173259411F, 0.2189578432F, 0.2205943168F, + 0.2222353361F, 0.2238808751F, 0.2255309076F, 0.2271854073F, + 0.2288443480F, 0.2305077030F, 0.2321754457F, 0.2338475493F, + 0.2355239869F, 0.2372047315F, 0.2388897560F, 0.2405790329F, + 0.2422725350F, 0.2439702347F, 0.2456721043F, 0.2473781159F, + 0.2490882418F, 0.2508024539F, 0.2525207240F, 0.2542430237F, + 0.2559693248F, 0.2576995986F, 0.2594338166F, 0.2611719498F, + 0.2629139695F, 0.2646598466F, 0.2664095520F, 0.2681630564F, + 0.2699203304F, 0.2716813445F, 0.2734460691F, 0.2752144744F, + 0.2769865307F, 0.2787622079F, 0.2805414760F, 0.2823243047F, + 0.2841106637F, 0.2859005227F, 0.2876938509F, 0.2894906179F, + 0.2912907928F, 0.2930943447F, 0.2949012426F, 0.2967114554F, + 0.2985249520F, 0.3003417009F, 0.3021616708F, 0.3039848301F, + 0.3058111471F, 0.3076405901F, 0.3094731273F, 0.3113087266F, + 0.3131473560F, 0.3149889833F, 0.3168335762F, 0.3186811024F, + 0.3205315294F, 0.3223848245F, 0.3242409552F, 0.3260998886F, + 0.3279615918F, 0.3298260319F, 0.3316931758F, 0.3335629903F, + 0.3354354423F, 0.3373104982F, 0.3391881247F, 0.3410682882F, + 0.3429509551F, 0.3448360917F, 0.3467236642F, 0.3486136387F, + 0.3505059811F, 0.3524006575F, 0.3542976336F, 0.3561968753F, + 0.3580983482F, 0.3600020179F, 0.3619078499F, 0.3638158096F, + 0.3657258625F, 0.3676379737F, 0.3695521086F, 0.3714682321F, + 0.3733863094F, 0.3753063055F, 0.3772281852F, 0.3791519134F, + 0.3810774548F, 0.3830047742F, 0.3849338362F, 0.3868646053F, + 0.3887970459F, 0.3907311227F, 0.3926667998F, 0.3946040417F, + 0.3965428125F, 0.3984830765F, 0.4004247978F, 0.4023679403F, + 0.4043124683F, 0.4062583455F, 0.4082055359F, 0.4101540034F, + 0.4121037117F, 0.4140546246F, 0.4160067058F, 0.4179599190F, + 0.4199142277F, 0.4218695956F, 0.4238259861F, 0.4257833627F, + 0.4277416888F, 0.4297009279F, 0.4316610433F, 0.4336219983F, + 0.4355837562F, 0.4375462803F, 0.4395095337F, 0.4414734797F, + 0.4434380815F, 0.4454033021F, 0.4473691046F, 0.4493354521F, + 0.4513023078F, 0.4532696345F, 0.4552373954F, 0.4572055533F, + 0.4591740713F, 0.4611429123F, 0.4631120393F, 0.4650814151F, + 0.4670510028F, 0.4690207650F, 0.4709906649F, 0.4729606651F, + 0.4749307287F, 0.4769008185F, 0.4788708972F, 0.4808409279F, + 0.4828108732F, 0.4847806962F, 0.4867503597F, 0.4887198264F, + 0.4906890593F, 0.4926580213F, 0.4946266753F, 0.4965949840F, + 0.4985629105F, 0.5005304176F, 0.5024974683F, 0.5044640255F, + 0.5064300522F, 0.5083955114F, 0.5103603659F, 0.5123245790F, + 0.5142881136F, 0.5162509328F, 0.5182129997F, 0.5201742774F, + 0.5221347290F, 0.5240943178F, 0.5260530070F, 0.5280107598F, + 0.5299675395F, 0.5319233095F, 0.5338780330F, 0.5358316736F, + 0.5377841946F, 0.5397355596F, 0.5416857320F, 0.5436346755F, + 0.5455823538F, 0.5475287304F, 0.5494737691F, 0.5514174337F, + 0.5533596881F, 0.5553004962F, 0.5572398218F, 0.5591776291F, + 0.5611138821F, 0.5630485449F, 0.5649815818F, 0.5669129570F, + 0.5688426349F, 0.5707705799F, 0.5726967564F, 0.5746211290F, + 0.5765436624F, 0.5784643212F, 0.5803830702F, 0.5822998743F, + 0.5842146984F, 0.5861275076F, 0.5880382669F, 0.5899469416F, + 0.5918534968F, 0.5937578981F, 0.5956601107F, 0.5975601004F, + 0.5994578326F, 0.6013532732F, 0.6032463880F, 0.6051371429F, + 0.6070255039F, 0.6089114372F, 0.6107949090F, 0.6126758856F, + 0.6145543334F, 0.6164302191F, 0.6183035092F, 0.6201741706F, + 0.6220421700F, 0.6239074745F, 0.6257700513F, 0.6276298674F, + 0.6294868903F, 0.6313410873F, 0.6331924262F, 0.6350408745F, + 0.6368864001F, 0.6387289710F, 0.6405685552F, 0.6424051209F, + 0.6442386364F, 0.6460690702F, 0.6478963910F, 0.6497205673F, + 0.6515415682F, 0.6533593625F, 0.6551739194F, 0.6569852082F, + 0.6587931984F, 0.6605978593F, 0.6623991609F, 0.6641970728F, + 0.6659915652F, 0.6677826081F, 0.6695701718F, 0.6713542268F, + 0.6731347437F, 0.6749116932F, 0.6766850461F, 0.6784547736F, + 0.6802208469F, 0.6819832374F, 0.6837419164F, 0.6854968559F, + 0.6872480275F, 0.6889954034F, 0.6907389556F, 0.6924786566F, + 0.6942144788F, 0.6959463950F, 0.6976743780F, 0.6993984008F, + 0.7011184365F, 0.7028344587F, 0.7045464407F, 0.7062543564F, + 0.7079581796F, 0.7096578844F, 0.7113534450F, 0.7130448359F, + 0.7147320316F, 0.7164150070F, 0.7180937371F, 0.7197681970F, + 0.7214383620F, 0.7231042077F, 0.7247657098F, 0.7264228443F, + 0.7280755871F, 0.7297239147F, 0.7313678035F, 0.7330072301F, + 0.7346421715F, 0.7362726046F, 0.7378985069F, 0.7395198556F, + 0.7411366285F, 0.7427488034F, 0.7443563584F, 0.7459592717F, + 0.7475575218F, 0.7491510873F, 0.7507399471F, 0.7523240803F, + 0.7539034661F, 0.7554780839F, 0.7570479136F, 0.7586129349F, + 0.7601731279F, 0.7617284730F, 0.7632789506F, 0.7648245416F, + 0.7663652267F, 0.7679009872F, 0.7694318044F, 0.7709576599F, + 0.7724785354F, 0.7739944130F, 0.7755052749F, 0.7770111035F, + 0.7785118815F, 0.7800075916F, 0.7814982170F, 0.7829837410F, + 0.7844641472F, 0.7859394191F, 0.7874095408F, 0.7888744965F, + 0.7903342706F, 0.7917888476F, 0.7932382124F, 0.7946823501F, + 0.7961212460F, 0.7975548855F, 0.7989832544F, 0.8004063386F, + 0.8018241244F, 0.8032365981F, 0.8046437463F, 0.8060455560F, + 0.8074420141F, 0.8088331080F, 0.8102188253F, 0.8115991536F, + 0.8129740810F, 0.8143435957F, 0.8157076861F, 0.8170663409F, + 0.8184195489F, 0.8197672994F, 0.8211095817F, 0.8224463853F, + 0.8237777001F, 0.8251035161F, 0.8264238235F, 0.8277386129F, + 0.8290478750F, 0.8303516008F, 0.8316497814F, 0.8329424083F, + 0.8342294731F, 0.8355109677F, 0.8367868841F, 0.8380572148F, + 0.8393219523F, 0.8405810893F, 0.8418346190F, 0.8430825345F, + 0.8443248294F, 0.8455614974F, 0.8467925323F, 0.8480179285F, + 0.8492376802F, 0.8504517822F, 0.8516602292F, 0.8528630164F, + 0.8540601391F, 0.8552515928F, 0.8564373733F, 0.8576174766F, + 0.8587918990F, 0.8599606368F, 0.8611236868F, 0.8622810460F, + 0.8634327113F, 0.8645786802F, 0.8657189504F, 0.8668535195F, + 0.8679823857F, 0.8691055472F, 0.8702230025F, 0.8713347503F, + 0.8724407896F, 0.8735411194F, 0.8746357394F, 0.8757246489F, + 0.8768078479F, 0.8778853364F, 0.8789571146F, 0.8800231832F, + 0.8810835427F, 0.8821381942F, 0.8831871387F, 0.8842303777F, + 0.8852679127F, 0.8862997456F, 0.8873258784F, 0.8883463132F, + 0.8893610527F, 0.8903700994F, 0.8913734562F, 0.8923711263F, + 0.8933631129F, 0.8943494196F, 0.8953300500F, 0.8963050083F, + 0.8972742985F, 0.8982379249F, 0.8991958922F, 0.9001482052F, + 0.9010948688F, 0.9020358883F, 0.9029712690F, 0.9039010165F, + 0.9048251367F, 0.9057436357F, 0.9066565195F, 0.9075637946F, + 0.9084654678F, 0.9093615456F, 0.9102520353F, 0.9111369440F, + 0.9120162792F, 0.9128900484F, 0.9137582595F, 0.9146209204F, + 0.9154780394F, 0.9163296248F, 0.9171756853F, 0.9180162296F, + 0.9188512667F, 0.9196808057F, 0.9205048559F, 0.9213234270F, + 0.9221365285F, 0.9229441704F, 0.9237463629F, 0.9245431160F, + 0.9253344404F, 0.9261203465F, 0.9269008453F, 0.9276759477F, + 0.9284456648F, 0.9292100080F, 0.9299689889F, 0.9307226190F, + 0.9314709103F, 0.9322138747F, 0.9329515245F, 0.9336838721F, + 0.9344109300F, 0.9351327108F, 0.9358492275F, 0.9365604931F, + 0.9372665208F, 0.9379673239F, 0.9386629160F, 0.9393533107F, + 0.9400385220F, 0.9407185637F, 0.9413934501F, 0.9420631954F, + 0.9427278141F, 0.9433873208F, 0.9440417304F, 0.9446910576F, + 0.9453353176F, 0.9459745255F, 0.9466086968F, 0.9472378469F, + 0.9478619915F, 0.9484811463F, 0.9490953274F, 0.9497045506F, + 0.9503088323F, 0.9509081888F, 0.9515026365F, 0.9520921921F, + 0.9526768723F, 0.9532566940F, 0.9538316742F, 0.9544018300F, + 0.9549671786F, 0.9555277375F, 0.9560835241F, 0.9566345562F, + 0.9571808513F, 0.9577224275F, 0.9582593027F, 0.9587914949F, + 0.9593190225F, 0.9598419038F, 0.9603601571F, 0.9608738012F, + 0.9613828546F, 0.9618873361F, 0.9623872646F, 0.9628826591F, + 0.9633735388F, 0.9638599227F, 0.9643418303F, 0.9648192808F, + 0.9652922939F, 0.9657608890F, 0.9662250860F, 0.9666849046F, + 0.9671403646F, 0.9675914861F, 0.9680382891F, 0.9684807937F, + 0.9689190202F, 0.9693529890F, 0.9697827203F, 0.9702082347F, + 0.9706295529F, 0.9710466953F, 0.9714596828F, 0.9718685362F, + 0.9722732762F, 0.9726739240F, 0.9730705005F, 0.9734630267F, + 0.9738515239F, 0.9742360134F, 0.9746165163F, 0.9749930540F, + 0.9753656481F, 0.9757343198F, 0.9760990909F, 0.9764599829F, + 0.9768170175F, 0.9771702164F, 0.9775196013F, 0.9778651941F, + 0.9782070167F, 0.9785450909F, 0.9788794388F, 0.9792100824F, + 0.9795370437F, 0.9798603449F, 0.9801800080F, 0.9804960554F, + 0.9808085092F, 0.9811173916F, 0.9814227251F, 0.9817245318F, + 0.9820228343F, 0.9823176549F, 0.9826090160F, 0.9828969402F, + 0.9831814498F, 0.9834625674F, 0.9837403156F, 0.9840147169F, + 0.9842857939F, 0.9845535692F, 0.9848180654F, 0.9850793052F, + 0.9853373113F, 0.9855921062F, 0.9858437127F, 0.9860921535F, + 0.9863374512F, 0.9865796287F, 0.9868187085F, 0.9870547136F, + 0.9872876664F, 0.9875175899F, 0.9877445067F, 0.9879684396F, + 0.9881894112F, 0.9884074444F, 0.9886225619F, 0.9888347863F, + 0.9890441404F, 0.9892506468F, 0.9894543284F, 0.9896552077F, + 0.9898533074F, 0.9900486502F, 0.9902412587F, 0.9904311555F, + 0.9906183633F, 0.9908029045F, 0.9909848019F, 0.9911640779F, + 0.9913407550F, 0.9915148557F, 0.9916864025F, 0.9918554179F, + 0.9920219241F, 0.9921859437F, 0.9923474989F, 0.9925066120F, + 0.9926633054F, 0.9928176012F, 0.9929695218F, 0.9931190891F, + 0.9932663254F, 0.9934112527F, 0.9935538932F, 0.9936942686F, + 0.9938324012F, 0.9939683126F, 0.9941020248F, 0.9942335597F, + 0.9943629388F, 0.9944901841F, 0.9946153170F, 0.9947383593F, + 0.9948593325F, 0.9949782579F, 0.9950951572F, 0.9952100516F, + 0.9953229625F, 0.9954339111F, 0.9955429186F, 0.9956500062F, + 0.9957551948F, 0.9958585056F, 0.9959599593F, 0.9960595769F, + 0.9961573792F, 0.9962533869F, 0.9963476206F, 0.9964401009F, + 0.9965308483F, 0.9966198833F, 0.9967072261F, 0.9967928971F, + 0.9968769164F, 0.9969593041F, 0.9970400804F, 0.9971192651F, + 0.9971968781F, 0.9972729391F, 0.9973474680F, 0.9974204842F, + 0.9974920074F, 0.9975620569F, 0.9976306521F, 0.9976978122F, + 0.9977635565F, 0.9978279039F, 0.9978908736F, 0.9979524842F, + 0.9980127547F, 0.9980717037F, 0.9981293499F, 0.9981857116F, + 0.9982408073F, 0.9982946554F, 0.9983472739F, 0.9983986810F, + 0.9984488947F, 0.9984979328F, 0.9985458132F, 0.9985925534F, + 0.9986381711F, 0.9986826838F, 0.9987261086F, 0.9987684630F, + 0.9988097640F, 0.9988500286F, 0.9988892738F, 0.9989275163F, + 0.9989647727F, 0.9990010597F, 0.9990363938F, 0.9990707911F, + 0.9991042679F, 0.9991368404F, 0.9991685244F, 0.9991993358F, + 0.9992292905F, 0.9992584038F, 0.9992866914F, 0.9993141686F, + 0.9993408506F, 0.9993667526F, 0.9993918895F, 0.9994162761F, + 0.9994399273F, 0.9994628576F, 0.9994850815F, 0.9995066133F, + 0.9995274672F, 0.9995476574F, 0.9995671978F, 0.9995861021F, + 0.9996043841F, 0.9996220573F, 0.9996391352F, 0.9996556310F, + 0.9996715579F, 0.9996869288F, 0.9997017568F, 0.9997160543F, + 0.9997298342F, 0.9997431088F, 0.9997558905F, 0.9997681914F, + 0.9997800236F, 0.9997913990F, 0.9998023292F, 0.9998128261F, + 0.9998229009F, 0.9998325650F, 0.9998418296F, 0.9998507058F, + 0.9998592044F, 0.9998673362F, 0.9998751117F, 0.9998825415F, + 0.9998896358F, 0.9998964047F, 0.9999028584F, 0.9999090066F, + 0.9999148590F, 0.9999204253F, 0.9999257148F, 0.9999307368F, + 0.9999355003F, 0.9999400144F, 0.9999442878F, 0.9999483293F, + 0.9999521472F, 0.9999557499F, 0.9999591457F, 0.9999623426F, + 0.9999653483F, 0.9999681708F, 0.9999708175F, 0.9999732959F, + 0.9999756132F, 0.9999777765F, 0.9999797928F, 0.9999816688F, + 0.9999834113F, 0.9999850266F, 0.9999865211F, 0.9999879009F, + 0.9999891721F, 0.9999903405F, 0.9999914118F, 0.9999923914F, + 0.9999932849F, 0.9999940972F, 0.9999948336F, 0.9999954989F, + 0.9999960978F, 0.9999966349F, 0.9999971146F, 0.9999975411F, + 0.9999979185F, 0.9999982507F, 0.9999985414F, 0.9999987944F, + 0.9999990129F, 0.9999992003F, 0.9999993596F, 0.9999994939F, + 0.9999996059F, 0.9999996981F, 0.9999997732F, 0.9999998333F, + 0.9999998805F, 0.9999999170F, 0.9999999444F, 0.9999999643F, + 0.9999999784F, 0.9999999878F, 0.9999999937F, 0.9999999972F, + 0.9999999990F, 0.9999999997F, 1.0000000000F, 1.0000000000F, +}; + +static const float vwin4096[2048] = { + 0.0000002310F, 0.0000020791F, 0.0000057754F, 0.0000113197F, + 0.0000187121F, 0.0000279526F, 0.0000390412F, 0.0000519777F, + 0.0000667623F, 0.0000833949F, 0.0001018753F, 0.0001222036F, + 0.0001443798F, 0.0001684037F, 0.0001942754F, 0.0002219947F, + 0.0002515616F, 0.0002829761F, 0.0003162380F, 0.0003513472F, + 0.0003883038F, 0.0004271076F, 0.0004677584F, 0.0005102563F, + 0.0005546011F, 0.0006007928F, 0.0006488311F, 0.0006987160F, + 0.0007504474F, 0.0008040251F, 0.0008594490F, 0.0009167191F, + 0.0009758351F, 0.0010367969F, 0.0010996044F, 0.0011642574F, + 0.0012307558F, 0.0012990994F, 0.0013692880F, 0.0014413216F, + 0.0015151998F, 0.0015909226F, 0.0016684898F, 0.0017479011F, + 0.0018291565F, 0.0019122556F, 0.0019971983F, 0.0020839845F, + 0.0021726138F, 0.0022630861F, 0.0023554012F, 0.0024495588F, + 0.0025455588F, 0.0026434008F, 0.0027430847F, 0.0028446103F, + 0.0029479772F, 0.0030531853F, 0.0031602342F, 0.0032691238F, + 0.0033798538F, 0.0034924239F, 0.0036068338F, 0.0037230833F, + 0.0038411721F, 0.0039610999F, 0.0040828664F, 0.0042064714F, + 0.0043319145F, 0.0044591954F, 0.0045883139F, 0.0047192696F, + 0.0048520622F, 0.0049866914F, 0.0051231569F, 0.0052614583F, + 0.0054015953F, 0.0055435676F, 0.0056873748F, 0.0058330166F, + 0.0059804926F, 0.0061298026F, 0.0062809460F, 0.0064339226F, + 0.0065887320F, 0.0067453738F, 0.0069038476F, 0.0070641531F, + 0.0072262899F, 0.0073902575F, 0.0075560556F, 0.0077236838F, + 0.0078931417F, 0.0080644288F, 0.0082375447F, 0.0084124891F, + 0.0085892615F, 0.0087678614F, 0.0089482885F, 0.0091305422F, + 0.0093146223F, 0.0095005281F, 0.0096882592F, 0.0098778153F, + 0.0100691958F, 0.0102624002F, 0.0104574281F, 0.0106542791F, + 0.0108529525F, 0.0110534480F, 0.0112557651F, 0.0114599032F, + 0.0116658618F, 0.0118736405F, 0.0120832387F, 0.0122946560F, + 0.0125078917F, 0.0127229454F, 0.0129398166F, 0.0131585046F, + 0.0133790090F, 0.0136013292F, 0.0138254647F, 0.0140514149F, + 0.0142791792F, 0.0145087572F, 0.0147401481F, 0.0149733515F, + 0.0152083667F, 0.0154451932F, 0.0156838304F, 0.0159242777F, + 0.0161665345F, 0.0164106001F, 0.0166564741F, 0.0169041557F, + 0.0171536443F, 0.0174049393F, 0.0176580401F, 0.0179129461F, + 0.0181696565F, 0.0184281708F, 0.0186884883F, 0.0189506084F, + 0.0192145303F, 0.0194802535F, 0.0197477772F, 0.0200171008F, + 0.0202882236F, 0.0205611449F, 0.0208358639F, 0.0211123801F, + 0.0213906927F, 0.0216708011F, 0.0219527043F, 0.0222364019F, + 0.0225218930F, 0.0228091769F, 0.0230982529F, 0.0233891203F, + 0.0236817782F, 0.0239762259F, 0.0242724628F, 0.0245704880F, + 0.0248703007F, 0.0251719002F, 0.0254752858F, 0.0257804565F, + 0.0260874117F, 0.0263961506F, 0.0267066722F, 0.0270189760F, + 0.0273330609F, 0.0276489263F, 0.0279665712F, 0.0282859949F, + 0.0286071966F, 0.0289301753F, 0.0292549303F, 0.0295814607F, + 0.0299097656F, 0.0302398442F, 0.0305716957F, 0.0309053191F, + 0.0312407135F, 0.0315778782F, 0.0319168122F, 0.0322575145F, + 0.0325999844F, 0.0329442209F, 0.0332902231F, 0.0336379900F, + 0.0339875208F, 0.0343388146F, 0.0346918703F, 0.0350466871F, + 0.0354032640F, 0.0357616000F, 0.0361216943F, 0.0364835458F, + 0.0368471535F, 0.0372125166F, 0.0375796339F, 0.0379485046F, + 0.0383191276F, 0.0386915020F, 0.0390656267F, 0.0394415008F, + 0.0398191231F, 0.0401984927F, 0.0405796086F, 0.0409624698F, + 0.0413470751F, 0.0417334235F, 0.0421215141F, 0.0425113457F, + 0.0429029172F, 0.0432962277F, 0.0436912760F, 0.0440880610F, + 0.0444865817F, 0.0448868370F, 0.0452888257F, 0.0456925468F, + 0.0460979992F, 0.0465051816F, 0.0469140931F, 0.0473247325F, + 0.0477370986F, 0.0481511902F, 0.0485670064F, 0.0489845458F, + 0.0494038074F, 0.0498247899F, 0.0502474922F, 0.0506719131F, + 0.0510980514F, 0.0515259060F, 0.0519554756F, 0.0523867590F, + 0.0528197550F, 0.0532544624F, 0.0536908800F, 0.0541290066F, + 0.0545688408F, 0.0550103815F, 0.0554536274F, 0.0558985772F, + 0.0563452297F, 0.0567935837F, 0.0572436377F, 0.0576953907F, + 0.0581488412F, 0.0586039880F, 0.0590608297F, 0.0595193651F, + 0.0599795929F, 0.0604415117F, 0.0609051202F, 0.0613704170F, + 0.0618374009F, 0.0623060704F, 0.0627764243F, 0.0632484611F, + 0.0637221795F, 0.0641975781F, 0.0646746555F, 0.0651534104F, + 0.0656338413F, 0.0661159469F, 0.0665997257F, 0.0670851763F, + 0.0675722973F, 0.0680610873F, 0.0685515448F, 0.0690436684F, + 0.0695374567F, 0.0700329081F, 0.0705300213F, 0.0710287947F, + 0.0715292269F, 0.0720313163F, 0.0725350616F, 0.0730404612F, + 0.0735475136F, 0.0740562172F, 0.0745665707F, 0.0750785723F, + 0.0755922207F, 0.0761075143F, 0.0766244515F, 0.0771430307F, + 0.0776632505F, 0.0781851092F, 0.0787086052F, 0.0792337371F, + 0.0797605032F, 0.0802889018F, 0.0808189315F, 0.0813505905F, + 0.0818838773F, 0.0824187903F, 0.0829553277F, 0.0834934881F, + 0.0840332697F, 0.0845746708F, 0.0851176899F, 0.0856623252F, + 0.0862085751F, 0.0867564379F, 0.0873059119F, 0.0878569954F, + 0.0884096867F, 0.0889639840F, 0.0895198858F, 0.0900773902F, + 0.0906364955F, 0.0911972000F, 0.0917595019F, 0.0923233995F, + 0.0928888909F, 0.0934559745F, 0.0940246485F, 0.0945949110F, + 0.0951667604F, 0.0957401946F, 0.0963152121F, 0.0968918109F, + 0.0974699893F, 0.0980497454F, 0.0986310773F, 0.0992139832F, + 0.0997984614F, 0.1003845098F, 0.1009721267F, 0.1015613101F, + 0.1021520582F, 0.1027443692F, 0.1033382410F, 0.1039336718F, + 0.1045306597F, 0.1051292027F, 0.1057292990F, 0.1063309466F, + 0.1069341435F, 0.1075388878F, 0.1081451776F, 0.1087530108F, + 0.1093623856F, 0.1099732998F, 0.1105857516F, 0.1111997389F, + 0.1118152597F, 0.1124323121F, 0.1130508939F, 0.1136710032F, + 0.1142926379F, 0.1149157960F, 0.1155404755F, 0.1161666742F, + 0.1167943901F, 0.1174236211F, 0.1180543652F, 0.1186866202F, + 0.1193203841F, 0.1199556548F, 0.1205924300F, 0.1212307078F, + 0.1218704860F, 0.1225117624F, 0.1231545349F, 0.1237988013F, + 0.1244445596F, 0.1250918074F, 0.1257405427F, 0.1263907632F, + 0.1270424667F, 0.1276956512F, 0.1283503142F, 0.1290064537F, + 0.1296640674F, 0.1303231530F, 0.1309837084F, 0.1316457312F, + 0.1323092193F, 0.1329741703F, 0.1336405820F, 0.1343084520F, + 0.1349777782F, 0.1356485582F, 0.1363207897F, 0.1369944704F, + 0.1376695979F, 0.1383461700F, 0.1390241842F, 0.1397036384F, + 0.1403845300F, 0.1410668567F, 0.1417506162F, 0.1424358061F, + 0.1431224240F, 0.1438104674F, 0.1444999341F, 0.1451908216F, + 0.1458831274F, 0.1465768492F, 0.1472719844F, 0.1479685308F, + 0.1486664857F, 0.1493658468F, 0.1500666115F, 0.1507687775F, + 0.1514723422F, 0.1521773031F, 0.1528836577F, 0.1535914035F, + 0.1543005380F, 0.1550110587F, 0.1557229631F, 0.1564362485F, + 0.1571509124F, 0.1578669524F, 0.1585843657F, 0.1593031499F, + 0.1600233024F, 0.1607448205F, 0.1614677017F, 0.1621919433F, + 0.1629175428F, 0.1636444975F, 0.1643728047F, 0.1651024619F, + 0.1658334665F, 0.1665658156F, 0.1672995067F, 0.1680345371F, + 0.1687709041F, 0.1695086050F, 0.1702476372F, 0.1709879978F, + 0.1717296843F, 0.1724726938F, 0.1732170237F, 0.1739626711F, + 0.1747096335F, 0.1754579079F, 0.1762074916F, 0.1769583819F, + 0.1777105760F, 0.1784640710F, 0.1792188642F, 0.1799749529F, + 0.1807323340F, 0.1814910049F, 0.1822509628F, 0.1830122046F, + 0.1837747277F, 0.1845385292F, 0.1853036062F, 0.1860699558F, + 0.1868375751F, 0.1876064613F, 0.1883766114F, 0.1891480226F, + 0.1899206919F, 0.1906946164F, 0.1914697932F, 0.1922462194F, + 0.1930238919F, 0.1938028079F, 0.1945829643F, 0.1953643583F, + 0.1961469868F, 0.1969308468F, 0.1977159353F, 0.1985022494F, + 0.1992897859F, 0.2000785420F, 0.2008685145F, 0.2016597005F, + 0.2024520968F, 0.2032457005F, 0.2040405084F, 0.2048365175F, + 0.2056337247F, 0.2064321269F, 0.2072317211F, 0.2080325041F, + 0.2088344727F, 0.2096376240F, 0.2104419547F, 0.2112474618F, + 0.2120541420F, 0.2128619923F, 0.2136710094F, 0.2144811902F, + 0.2152925315F, 0.2161050301F, 0.2169186829F, 0.2177334866F, + 0.2185494381F, 0.2193665340F, 0.2201847712F, 0.2210041465F, + 0.2218246565F, 0.2226462981F, 0.2234690680F, 0.2242929629F, + 0.2251179796F, 0.2259441147F, 0.2267713650F, 0.2275997272F, + 0.2284291979F, 0.2292597739F, 0.2300914518F, 0.2309242283F, + 0.2317581001F, 0.2325930638F, 0.2334291160F, 0.2342662534F, + 0.2351044727F, 0.2359437703F, 0.2367841431F, 0.2376255875F, + 0.2384681001F, 0.2393116776F, 0.2401563165F, 0.2410020134F, + 0.2418487649F, 0.2426965675F, 0.2435454178F, 0.2443953122F, + 0.2452462474F, 0.2460982199F, 0.2469512262F, 0.2478052628F, + 0.2486603262F, 0.2495164129F, 0.2503735194F, 0.2512316421F, + 0.2520907776F, 0.2529509222F, 0.2538120726F, 0.2546742250F, + 0.2555373760F, 0.2564015219F, 0.2572666593F, 0.2581327845F, + 0.2589998939F, 0.2598679840F, 0.2607370510F, 0.2616070916F, + 0.2624781019F, 0.2633500783F, 0.2642230173F, 0.2650969152F, + 0.2659717684F, 0.2668475731F, 0.2677243257F, 0.2686020226F, + 0.2694806601F, 0.2703602344F, 0.2712407419F, 0.2721221789F, + 0.2730045417F, 0.2738878265F, 0.2747720297F, 0.2756571474F, + 0.2765431760F, 0.2774301117F, 0.2783179508F, 0.2792066895F, + 0.2800963240F, 0.2809868505F, 0.2818782654F, 0.2827705647F, + 0.2836637447F, 0.2845578016F, 0.2854527315F, 0.2863485307F, + 0.2872451953F, 0.2881427215F, 0.2890411055F, 0.2899403433F, + 0.2908404312F, 0.2917413654F, 0.2926431418F, 0.2935457567F, + 0.2944492061F, 0.2953534863F, 0.2962585932F, 0.2971645230F, + 0.2980712717F, 0.2989788356F, 0.2998872105F, 0.3007963927F, + 0.3017063781F, 0.3026171629F, 0.3035287430F, 0.3044411145F, + 0.3053542736F, 0.3062682161F, 0.3071829381F, 0.3080984356F, + 0.3090147047F, 0.3099317413F, 0.3108495414F, 0.3117681011F, + 0.3126874163F, 0.3136074830F, 0.3145282972F, 0.3154498548F, + 0.3163721517F, 0.3172951841F, 0.3182189477F, 0.3191434385F, + 0.3200686525F, 0.3209945856F, 0.3219212336F, 0.3228485927F, + 0.3237766585F, 0.3247054271F, 0.3256348943F, 0.3265650560F, + 0.3274959081F, 0.3284274465F, 0.3293596671F, 0.3302925657F, + 0.3312261382F, 0.3321603804F, 0.3330952882F, 0.3340308574F, + 0.3349670838F, 0.3359039634F, 0.3368414919F, 0.3377796651F, + 0.3387184789F, 0.3396579290F, 0.3405980113F, 0.3415387216F, + 0.3424800556F, 0.3434220091F, 0.3443645779F, 0.3453077578F, + 0.3462515446F, 0.3471959340F, 0.3481409217F, 0.3490865036F, + 0.3500326754F, 0.3509794328F, 0.3519267715F, 0.3528746873F, + 0.3538231759F, 0.3547722330F, 0.3557218544F, 0.3566720357F, + 0.3576227727F, 0.3585740610F, 0.3595258964F, 0.3604782745F, + 0.3614311910F, 0.3623846417F, 0.3633386221F, 0.3642931280F, + 0.3652481549F, 0.3662036987F, 0.3671597548F, 0.3681163191F, + 0.3690733870F, 0.3700309544F, 0.3709890167F, 0.3719475696F, + 0.3729066089F, 0.3738661299F, 0.3748261285F, 0.3757866002F, + 0.3767475406F, 0.3777089453F, 0.3786708100F, 0.3796331302F, + 0.3805959014F, 0.3815591194F, 0.3825227796F, 0.3834868777F, + 0.3844514093F, 0.3854163698F, 0.3863817549F, 0.3873475601F, + 0.3883137810F, 0.3892804131F, 0.3902474521F, 0.3912148933F, + 0.3921827325F, 0.3931509650F, 0.3941195865F, 0.3950885925F, + 0.3960579785F, 0.3970277400F, 0.3979978725F, 0.3989683716F, + 0.3999392328F, 0.4009104516F, 0.4018820234F, 0.4028539438F, + 0.4038262084F, 0.4047988125F, 0.4057717516F, 0.4067450214F, + 0.4077186172F, 0.4086925345F, 0.4096667688F, 0.4106413155F, + 0.4116161703F, 0.4125913284F, 0.4135667854F, 0.4145425368F, + 0.4155185780F, 0.4164949044F, 0.4174715116F, 0.4184483949F, + 0.4194255498F, 0.4204029718F, 0.4213806563F, 0.4223585987F, + 0.4233367946F, 0.4243152392F, 0.4252939281F, 0.4262728566F, + 0.4272520202F, 0.4282314144F, 0.4292110345F, 0.4301908760F, + 0.4311709343F, 0.4321512047F, 0.4331316828F, 0.4341123639F, + 0.4350932435F, 0.4360743168F, 0.4370555794F, 0.4380370267F, + 0.4390186540F, 0.4400004567F, 0.4409824303F, 0.4419645701F, + 0.4429468716F, 0.4439293300F, 0.4449119409F, 0.4458946996F, + 0.4468776014F, 0.4478606418F, 0.4488438162F, 0.4498271199F, + 0.4508105483F, 0.4517940967F, 0.4527777607F, 0.4537615355F, + 0.4547454165F, 0.4557293991F, 0.4567134786F, 0.4576976505F, + 0.4586819101F, 0.4596662527F, 0.4606506738F, 0.4616351687F, + 0.4626197328F, 0.4636043614F, 0.4645890499F, 0.4655737936F, + 0.4665585880F, 0.4675434284F, 0.4685283101F, 0.4695132286F, + 0.4704981791F, 0.4714831570F, 0.4724681577F, 0.4734531766F, + 0.4744382089F, 0.4754232501F, 0.4764082956F, 0.4773933406F, + 0.4783783806F, 0.4793634108F, 0.4803484267F, 0.4813334237F, + 0.4823183969F, 0.4833033419F, 0.4842882540F, 0.4852731285F, + 0.4862579608F, 0.4872427462F, 0.4882274802F, 0.4892121580F, + 0.4901967751F, 0.4911813267F, 0.4921658083F, 0.4931502151F, + 0.4941345427F, 0.4951187863F, 0.4961029412F, 0.4970870029F, + 0.4980709667F, 0.4990548280F, 0.5000385822F, 0.5010222245F, + 0.5020057505F, 0.5029891553F, 0.5039724345F, 0.5049555834F, + 0.5059385973F, 0.5069214716F, 0.5079042018F, 0.5088867831F, + 0.5098692110F, 0.5108514808F, 0.5118335879F, 0.5128155277F, + 0.5137972956F, 0.5147788869F, 0.5157602971F, 0.5167415215F, + 0.5177225555F, 0.5187033945F, 0.5196840339F, 0.5206644692F, + 0.5216446956F, 0.5226247086F, 0.5236045035F, 0.5245840759F, + 0.5255634211F, 0.5265425344F, 0.5275214114F, 0.5285000474F, + 0.5294784378F, 0.5304565781F, 0.5314344637F, 0.5324120899F, + 0.5333894522F, 0.5343665461F, 0.5353433670F, 0.5363199102F, + 0.5372961713F, 0.5382721457F, 0.5392478287F, 0.5402232159F, + 0.5411983027F, 0.5421730845F, 0.5431475569F, 0.5441217151F, + 0.5450955548F, 0.5460690714F, 0.5470422602F, 0.5480151169F, + 0.5489876368F, 0.5499598155F, 0.5509316484F, 0.5519031310F, + 0.5528742587F, 0.5538450271F, 0.5548154317F, 0.5557854680F, + 0.5567551314F, 0.5577244174F, 0.5586933216F, 0.5596618395F, + 0.5606299665F, 0.5615976983F, 0.5625650302F, 0.5635319580F, + 0.5644984770F, 0.5654645828F, 0.5664302709F, 0.5673955370F, + 0.5683603765F, 0.5693247850F, 0.5702887580F, 0.5712522912F, + 0.5722153800F, 0.5731780200F, 0.5741402069F, 0.5751019362F, + 0.5760632034F, 0.5770240042F, 0.5779843341F, 0.5789441889F, + 0.5799035639F, 0.5808624549F, 0.5818208575F, 0.5827787673F, + 0.5837361800F, 0.5846930910F, 0.5856494961F, 0.5866053910F, + 0.5875607712F, 0.5885156324F, 0.5894699703F, 0.5904237804F, + 0.5913770586F, 0.5923298004F, 0.5932820016F, 0.5942336578F, + 0.5951847646F, 0.5961353179F, 0.5970853132F, 0.5980347464F, + 0.5989836131F, 0.5999319090F, 0.6008796298F, 0.6018267713F, + 0.6027733292F, 0.6037192993F, 0.6046646773F, 0.6056094589F, + 0.6065536400F, 0.6074972162F, 0.6084401833F, 0.6093825372F, + 0.6103242736F, 0.6112653884F, 0.6122058772F, 0.6131457359F, + 0.6140849604F, 0.6150235464F, 0.6159614897F, 0.6168987862F, + 0.6178354318F, 0.6187714223F, 0.6197067535F, 0.6206414213F, + 0.6215754215F, 0.6225087501F, 0.6234414028F, 0.6243733757F, + 0.6253046646F, 0.6262352654F, 0.6271651739F, 0.6280943862F, + 0.6290228982F, 0.6299507057F, 0.6308778048F, 0.6318041913F, + 0.6327298612F, 0.6336548105F, 0.6345790352F, 0.6355025312F, + 0.6364252945F, 0.6373473211F, 0.6382686070F, 0.6391891483F, + 0.6401089409F, 0.6410279808F, 0.6419462642F, 0.6428637869F, + 0.6437805452F, 0.6446965350F, 0.6456117524F, 0.6465261935F, + 0.6474398544F, 0.6483527311F, 0.6492648197F, 0.6501761165F, + 0.6510866174F, 0.6519963186F, 0.6529052162F, 0.6538133064F, + 0.6547205854F, 0.6556270492F, 0.6565326941F, 0.6574375162F, + 0.6583415117F, 0.6592446769F, 0.6601470079F, 0.6610485009F, + 0.6619491521F, 0.6628489578F, 0.6637479143F, 0.6646460177F, + 0.6655432643F, 0.6664396505F, 0.6673351724F, 0.6682298264F, + 0.6691236087F, 0.6700165157F, 0.6709085436F, 0.6717996889F, + 0.6726899478F, 0.6735793167F, 0.6744677918F, 0.6753553697F, + 0.6762420466F, 0.6771278190F, 0.6780126832F, 0.6788966357F, + 0.6797796728F, 0.6806617909F, 0.6815429866F, 0.6824232562F, + 0.6833025961F, 0.6841810030F, 0.6850584731F, 0.6859350031F, + 0.6868105894F, 0.6876852284F, 0.6885589168F, 0.6894316510F, + 0.6903034275F, 0.6911742430F, 0.6920440939F, 0.6929129769F, + 0.6937808884F, 0.6946478251F, 0.6955137837F, 0.6963787606F, + 0.6972427525F, 0.6981057560F, 0.6989677678F, 0.6998287845F, + 0.7006888028F, 0.7015478194F, 0.7024058309F, 0.7032628340F, + 0.7041188254F, 0.7049738019F, 0.7058277601F, 0.7066806969F, + 0.7075326089F, 0.7083834929F, 0.7092333457F, 0.7100821640F, + 0.7109299447F, 0.7117766846F, 0.7126223804F, 0.7134670291F, + 0.7143106273F, 0.7151531721F, 0.7159946602F, 0.7168350885F, + 0.7176744539F, 0.7185127534F, 0.7193499837F, 0.7201861418F, + 0.7210212247F, 0.7218552293F, 0.7226881526F, 0.7235199914F, + 0.7243507428F, 0.7251804039F, 0.7260089715F, 0.7268364426F, + 0.7276628144F, 0.7284880839F, 0.7293122481F, 0.7301353040F, + 0.7309572487F, 0.7317780794F, 0.7325977930F, 0.7334163868F, + 0.7342338579F, 0.7350502033F, 0.7358654202F, 0.7366795059F, + 0.7374924573F, 0.7383042718F, 0.7391149465F, 0.7399244787F, + 0.7407328655F, 0.7415401041F, 0.7423461920F, 0.7431511261F, + 0.7439549040F, 0.7447575227F, 0.7455589797F, 0.7463592723F, + 0.7471583976F, 0.7479563532F, 0.7487531363F, 0.7495487443F, + 0.7503431745F, 0.7511364244F, 0.7519284913F, 0.7527193726F, + 0.7535090658F, 0.7542975683F, 0.7550848776F, 0.7558709910F, + 0.7566559062F, 0.7574396205F, 0.7582221314F, 0.7590034366F, + 0.7597835334F, 0.7605624194F, 0.7613400923F, 0.7621165495F, + 0.7628917886F, 0.7636658072F, 0.7644386030F, 0.7652101735F, + 0.7659805164F, 0.7667496292F, 0.7675175098F, 0.7682841556F, + 0.7690495645F, 0.7698137341F, 0.7705766622F, 0.7713383463F, + 0.7720987844F, 0.7728579741F, 0.7736159132F, 0.7743725994F, + 0.7751280306F, 0.7758822046F, 0.7766351192F, 0.7773867722F, + 0.7781371614F, 0.7788862848F, 0.7796341401F, 0.7803807253F, + 0.7811260383F, 0.7818700769F, 0.7826128392F, 0.7833543230F, + 0.7840945263F, 0.7848334471F, 0.7855710833F, 0.7863074330F, + 0.7870424941F, 0.7877762647F, 0.7885087428F, 0.7892399264F, + 0.7899698137F, 0.7906984026F, 0.7914256914F, 0.7921516780F, + 0.7928763607F, 0.7935997375F, 0.7943218065F, 0.7950425661F, + 0.7957620142F, 0.7964801492F, 0.7971969692F, 0.7979124724F, + 0.7986266570F, 0.7993395214F, 0.8000510638F, 0.8007612823F, + 0.8014701754F, 0.8021777413F, 0.8028839784F, 0.8035888849F, + 0.8042924592F, 0.8049946997F, 0.8056956048F, 0.8063951727F, + 0.8070934020F, 0.8077902910F, 0.8084858381F, 0.8091800419F, + 0.8098729007F, 0.8105644130F, 0.8112545774F, 0.8119433922F, + 0.8126308561F, 0.8133169676F, 0.8140017251F, 0.8146851272F, + 0.8153671726F, 0.8160478598F, 0.8167271874F, 0.8174051539F, + 0.8180817582F, 0.8187569986F, 0.8194308741F, 0.8201033831F, + 0.8207745244F, 0.8214442966F, 0.8221126986F, 0.8227797290F, + 0.8234453865F, 0.8241096700F, 0.8247725781F, 0.8254341097F, + 0.8260942636F, 0.8267530385F, 0.8274104334F, 0.8280664470F, + 0.8287210782F, 0.8293743259F, 0.8300261889F, 0.8306766662F, + 0.8313257566F, 0.8319734591F, 0.8326197727F, 0.8332646963F, + 0.8339082288F, 0.8345503692F, 0.8351911167F, 0.8358304700F, + 0.8364684284F, 0.8371049907F, 0.8377401562F, 0.8383739238F, + 0.8390062927F, 0.8396372618F, 0.8402668305F, 0.8408949977F, + 0.8415217626F, 0.8421471245F, 0.8427710823F, 0.8433936354F, + 0.8440147830F, 0.8446345242F, 0.8452528582F, 0.8458697844F, + 0.8464853020F, 0.8470994102F, 0.8477121084F, 0.8483233958F, + 0.8489332718F, 0.8495417356F, 0.8501487866F, 0.8507544243F, + 0.8513586479F, 0.8519614568F, 0.8525628505F, 0.8531628283F, + 0.8537613897F, 0.8543585341F, 0.8549542611F, 0.8555485699F, + 0.8561414603F, 0.8567329315F, 0.8573229832F, 0.8579116149F, + 0.8584988262F, 0.8590846165F, 0.8596689855F, 0.8602519327F, + 0.8608334577F, 0.8614135603F, 0.8619922399F, 0.8625694962F, + 0.8631453289F, 0.8637197377F, 0.8642927222F, 0.8648642821F, + 0.8654344172F, 0.8660031272F, 0.8665704118F, 0.8671362708F, + 0.8677007039F, 0.8682637109F, 0.8688252917F, 0.8693854460F, + 0.8699441737F, 0.8705014745F, 0.8710573485F, 0.8716117953F, + 0.8721648150F, 0.8727164073F, 0.8732665723F, 0.8738153098F, + 0.8743626197F, 0.8749085021F, 0.8754529569F, 0.8759959840F, + 0.8765375835F, 0.8770777553F, 0.8776164996F, 0.8781538162F, + 0.8786897054F, 0.8792241670F, 0.8797572013F, 0.8802888082F, + 0.8808189880F, 0.8813477407F, 0.8818750664F, 0.8824009653F, + 0.8829254375F, 0.8834484833F, 0.8839701028F, 0.8844902961F, + 0.8850090636F, 0.8855264054F, 0.8860423218F, 0.8865568131F, + 0.8870698794F, 0.8875815212F, 0.8880917386F, 0.8886005319F, + 0.8891079016F, 0.8896138479F, 0.8901183712F, 0.8906214719F, + 0.8911231503F, 0.8916234067F, 0.8921222417F, 0.8926196556F, + 0.8931156489F, 0.8936102219F, 0.8941033752F, 0.8945951092F, + 0.8950854244F, 0.8955743212F, 0.8960618003F, 0.8965478621F, + 0.8970325071F, 0.8975157359F, 0.8979975490F, 0.8984779471F, + 0.8989569307F, 0.8994345004F, 0.8999106568F, 0.9003854005F, + 0.9008587323F, 0.9013306526F, 0.9018011623F, 0.9022702619F, + 0.9027379521F, 0.9032042337F, 0.9036691074F, 0.9041325739F, + 0.9045946339F, 0.9050552882F, 0.9055145376F, 0.9059723828F, + 0.9064288246F, 0.9068838638F, 0.9073375013F, 0.9077897379F, + 0.9082405743F, 0.9086900115F, 0.9091380503F, 0.9095846917F, + 0.9100299364F, 0.9104737854F, 0.9109162397F, 0.9113573001F, + 0.9117969675F, 0.9122352430F, 0.9126721275F, 0.9131076219F, + 0.9135417273F, 0.9139744447F, 0.9144057750F, 0.9148357194F, + 0.9152642787F, 0.9156914542F, 0.9161172468F, 0.9165416576F, + 0.9169646877F, 0.9173863382F, 0.9178066102F, 0.9182255048F, + 0.9186430232F, 0.9190591665F, 0.9194739359F, 0.9198873324F, + 0.9202993574F, 0.9207100120F, 0.9211192973F, 0.9215272147F, + 0.9219337653F, 0.9223389504F, 0.9227427713F, 0.9231452290F, + 0.9235463251F, 0.9239460607F, 0.9243444371F, 0.9247414557F, + 0.9251371177F, 0.9255314245F, 0.9259243774F, 0.9263159778F, + 0.9267062270F, 0.9270951264F, 0.9274826774F, 0.9278688814F, + 0.9282537398F, 0.9286372540F, 0.9290194254F, 0.9294002555F, + 0.9297797458F, 0.9301578976F, 0.9305347125F, 0.9309101919F, + 0.9312843373F, 0.9316571503F, 0.9320286323F, 0.9323987849F, + 0.9327676097F, 0.9331351080F, 0.9335012816F, 0.9338661320F, + 0.9342296607F, 0.9345918694F, 0.9349527596F, 0.9353123330F, + 0.9356705911F, 0.9360275357F, 0.9363831683F, 0.9367374905F, + 0.9370905042F, 0.9374422108F, 0.9377926122F, 0.9381417099F, + 0.9384895057F, 0.9388360014F, 0.9391811985F, 0.9395250989F, + 0.9398677043F, 0.9402090165F, 0.9405490371F, 0.9408877680F, + 0.9412252110F, 0.9415613678F, 0.9418962402F, 0.9422298301F, + 0.9425621392F, 0.9428931695F, 0.9432229226F, 0.9435514005F, + 0.9438786050F, 0.9442045381F, 0.9445292014F, 0.9448525971F, + 0.9451747268F, 0.9454955926F, 0.9458151963F, 0.9461335399F, + 0.9464506253F, 0.9467664545F, 0.9470810293F, 0.9473943517F, + 0.9477064238F, 0.9480172474F, 0.9483268246F, 0.9486351573F, + 0.9489422475F, 0.9492480973F, 0.9495527087F, 0.9498560837F, + 0.9501582243F, 0.9504591325F, 0.9507588105F, 0.9510572603F, + 0.9513544839F, 0.9516504834F, 0.9519452609F, 0.9522388186F, + 0.9525311584F, 0.9528222826F, 0.9531121932F, 0.9534008923F, + 0.9536883821F, 0.9539746647F, 0.9542597424F, 0.9545436171F, + 0.9548262912F, 0.9551077667F, 0.9553880459F, 0.9556671309F, + 0.9559450239F, 0.9562217272F, 0.9564972429F, 0.9567715733F, + 0.9570447206F, 0.9573166871F, 0.9575874749F, 0.9578570863F, + 0.9581255236F, 0.9583927890F, 0.9586588849F, 0.9589238134F, + 0.9591875769F, 0.9594501777F, 0.9597116180F, 0.9599719003F, + 0.9602310267F, 0.9604889995F, 0.9607458213F, 0.9610014942F, + 0.9612560206F, 0.9615094028F, 0.9617616433F, 0.9620127443F, + 0.9622627083F, 0.9625115376F, 0.9627592345F, 0.9630058016F, + 0.9632512411F, 0.9634955555F, 0.9637387471F, 0.9639808185F, + 0.9642217720F, 0.9644616100F, 0.9647003349F, 0.9649379493F, + 0.9651744556F, 0.9654098561F, 0.9656441534F, 0.9658773499F, + 0.9661094480F, 0.9663404504F, 0.9665703593F, 0.9667991774F, + 0.9670269071F, 0.9672535509F, 0.9674791114F, 0.9677035909F, + 0.9679269921F, 0.9681493174F, 0.9683705694F, 0.9685907506F, + 0.9688098636F, 0.9690279108F, 0.9692448948F, 0.9694608182F, + 0.9696756836F, 0.9698894934F, 0.9701022503F, 0.9703139569F, + 0.9705246156F, 0.9707342291F, 0.9709428000F, 0.9711503309F, + 0.9713568243F, 0.9715622829F, 0.9717667093F, 0.9719701060F, + 0.9721724757F, 0.9723738210F, 0.9725741446F, 0.9727734490F, + 0.9729717369F, 0.9731690109F, 0.9733652737F, 0.9735605279F, + 0.9737547762F, 0.9739480212F, 0.9741402656F, 0.9743315120F, + 0.9745217631F, 0.9747110216F, 0.9748992901F, 0.9750865714F, + 0.9752728681F, 0.9754581829F, 0.9756425184F, 0.9758258775F, + 0.9760082627F, 0.9761896768F, 0.9763701224F, 0.9765496024F, + 0.9767281193F, 0.9769056760F, 0.9770822751F, 0.9772579193F, + 0.9774326114F, 0.9776063542F, 0.9777791502F, 0.9779510023F, + 0.9781219133F, 0.9782918858F, 0.9784609226F, 0.9786290264F, + 0.9787962000F, 0.9789624461F, 0.9791277676F, 0.9792921671F, + 0.9794556474F, 0.9796182113F, 0.9797798615F, 0.9799406009F, + 0.9801004321F, 0.9802593580F, 0.9804173813F, 0.9805745049F, + 0.9807307314F, 0.9808860637F, 0.9810405046F, 0.9811940568F, + 0.9813467232F, 0.9814985065F, 0.9816494095F, 0.9817994351F, + 0.9819485860F, 0.9820968650F, 0.9822442750F, 0.9823908186F, + 0.9825364988F, 0.9826813184F, 0.9828252801F, 0.9829683868F, + 0.9831106413F, 0.9832520463F, 0.9833926048F, 0.9835323195F, + 0.9836711932F, 0.9838092288F, 0.9839464291F, 0.9840827969F, + 0.9842183351F, 0.9843530464F, 0.9844869337F, 0.9846199998F, + 0.9847522475F, 0.9848836798F, 0.9850142993F, 0.9851441090F, + 0.9852731117F, 0.9854013101F, 0.9855287073F, 0.9856553058F, + 0.9857811087F, 0.9859061188F, 0.9860303388F, 0.9861537717F, + 0.9862764202F, 0.9863982872F, 0.9865193756F, 0.9866396882F, + 0.9867592277F, 0.9868779972F, 0.9869959993F, 0.9871132370F, + 0.9872297131F, 0.9873454304F, 0.9874603918F, 0.9875746001F, + 0.9876880581F, 0.9878007688F, 0.9879127348F, 0.9880239592F, + 0.9881344447F, 0.9882441941F, 0.9883532104F, 0.9884614962F, + 0.9885690546F, 0.9886758883F, 0.9887820001F, 0.9888873930F, + 0.9889920697F, 0.9890960331F, 0.9891992859F, 0.9893018312F, + 0.9894036716F, 0.9895048100F, 0.9896052493F, 0.9897049923F, + 0.9898040418F, 0.9899024006F, 0.9900000717F, 0.9900970577F, + 0.9901933616F, 0.9902889862F, 0.9903839343F, 0.9904782087F, + 0.9905718122F, 0.9906647477F, 0.9907570180F, 0.9908486259F, + 0.9909395742F, 0.9910298658F, 0.9911195034F, 0.9912084899F, + 0.9912968281F, 0.9913845208F, 0.9914715708F, 0.9915579810F, + 0.9916437540F, 0.9917288928F, 0.9918134001F, 0.9918972788F, + 0.9919805316F, 0.9920631613F, 0.9921451707F, 0.9922265626F, + 0.9923073399F, 0.9923875052F, 0.9924670615F, 0.9925460114F, + 0.9926243577F, 0.9927021033F, 0.9927792508F, 0.9928558032F, + 0.9929317631F, 0.9930071333F, 0.9930819167F, 0.9931561158F, + 0.9932297337F, 0.9933027728F, 0.9933752362F, 0.9934471264F, + 0.9935184462F, 0.9935891985F, 0.9936593859F, 0.9937290112F, + 0.9937980771F, 0.9938665864F, 0.9939345418F, 0.9940019460F, + 0.9940688018F, 0.9941351118F, 0.9942008789F, 0.9942661057F, + 0.9943307950F, 0.9943949494F, 0.9944585717F, 0.9945216645F, + 0.9945842307F, 0.9946462728F, 0.9947077936F, 0.9947687957F, + 0.9948292820F, 0.9948892550F, 0.9949487174F, 0.9950076719F, + 0.9950661212F, 0.9951240679F, 0.9951815148F, 0.9952384645F, + 0.9952949196F, 0.9953508828F, 0.9954063568F, 0.9954613442F, + 0.9955158476F, 0.9955698697F, 0.9956234132F, 0.9956764806F, + 0.9957290746F, 0.9957811978F, 0.9958328528F, 0.9958840423F, + 0.9959347688F, 0.9959850351F, 0.9960348435F, 0.9960841969F, + 0.9961330977F, 0.9961815486F, 0.9962295521F, 0.9962771108F, + 0.9963242274F, 0.9963709043F, 0.9964171441F, 0.9964629494F, + 0.9965083228F, 0.9965532668F, 0.9965977840F, 0.9966418768F, + 0.9966855479F, 0.9967287998F, 0.9967716350F, 0.9968140559F, + 0.9968560653F, 0.9968976655F, 0.9969388591F, 0.9969796485F, + 0.9970200363F, 0.9970600250F, 0.9970996170F, 0.9971388149F, + 0.9971776211F, 0.9972160380F, 0.9972540683F, 0.9972917142F, + 0.9973289783F, 0.9973658631F, 0.9974023709F, 0.9974385042F, + 0.9974742655F, 0.9975096571F, 0.9975446816F, 0.9975793413F, + 0.9976136386F, 0.9976475759F, 0.9976811557F, 0.9977143803F, + 0.9977472521F, 0.9977797736F, 0.9978119470F, 0.9978437748F, + 0.9978752593F, 0.9979064029F, 0.9979372079F, 0.9979676768F, + 0.9979978117F, 0.9980276151F, 0.9980570893F, 0.9980862367F, + 0.9981150595F, 0.9981435600F, 0.9981717406F, 0.9981996035F, + 0.9982271511F, 0.9982543856F, 0.9982813093F, 0.9983079246F, + 0.9983342336F, 0.9983602386F, 0.9983859418F, 0.9984113456F, + 0.9984364522F, 0.9984612638F, 0.9984857825F, 0.9985100108F, + 0.9985339507F, 0.9985576044F, 0.9985809743F, 0.9986040624F, + 0.9986268710F, 0.9986494022F, 0.9986716583F, 0.9986936413F, + 0.9987153535F, 0.9987367969F, 0.9987579738F, 0.9987788864F, + 0.9987995366F, 0.9988199267F, 0.9988400587F, 0.9988599348F, + 0.9988795572F, 0.9988989278F, 0.9989180487F, 0.9989369222F, + 0.9989555501F, 0.9989739347F, 0.9989920780F, 0.9990099820F, + 0.9990276487F, 0.9990450803F, 0.9990622787F, 0.9990792460F, + 0.9990959841F, 0.9991124952F, 0.9991287812F, 0.9991448440F, + 0.9991606858F, 0.9991763084F, 0.9991917139F, 0.9992069042F, + 0.9992218813F, 0.9992366471F, 0.9992512035F, 0.9992655525F, + 0.9992796961F, 0.9992936361F, 0.9993073744F, 0.9993209131F, + 0.9993342538F, 0.9993473987F, 0.9993603494F, 0.9993731080F, + 0.9993856762F, 0.9993980559F, 0.9994102490F, 0.9994222573F, + 0.9994340827F, 0.9994457269F, 0.9994571918F, 0.9994684793F, + 0.9994795910F, 0.9994905288F, 0.9995012945F, 0.9995118898F, + 0.9995223165F, 0.9995325765F, 0.9995426713F, 0.9995526029F, + 0.9995623728F, 0.9995719829F, 0.9995814349F, 0.9995907304F, + 0.9995998712F, 0.9996088590F, 0.9996176954F, 0.9996263821F, + 0.9996349208F, 0.9996433132F, 0.9996515609F, 0.9996596656F, + 0.9996676288F, 0.9996754522F, 0.9996831375F, 0.9996906862F, + 0.9996981000F, 0.9997053804F, 0.9997125290F, 0.9997195474F, + 0.9997264371F, 0.9997331998F, 0.9997398369F, 0.9997463500F, + 0.9997527406F, 0.9997590103F, 0.9997651606F, 0.9997711930F, + 0.9997771089F, 0.9997829098F, 0.9997885973F, 0.9997941728F, + 0.9997996378F, 0.9998049936F, 0.9998102419F, 0.9998153839F, + 0.9998204211F, 0.9998253550F, 0.9998301868F, 0.9998349182F, + 0.9998395503F, 0.9998440847F, 0.9998485226F, 0.9998528654F, + 0.9998571146F, 0.9998612713F, 0.9998653370F, 0.9998693130F, + 0.9998732007F, 0.9998770012F, 0.9998807159F, 0.9998843461F, + 0.9998878931F, 0.9998913581F, 0.9998947424F, 0.9998980473F, + 0.9999012740F, 0.9999044237F, 0.9999074976F, 0.9999104971F, + 0.9999134231F, 0.9999162771F, 0.9999190601F, 0.9999217733F, + 0.9999244179F, 0.9999269950F, 0.9999295058F, 0.9999319515F, + 0.9999343332F, 0.9999366519F, 0.9999389088F, 0.9999411050F, + 0.9999432416F, 0.9999453196F, 0.9999473402F, 0.9999493044F, + 0.9999512132F, 0.9999530677F, 0.9999548690F, 0.9999566180F, + 0.9999583157F, 0.9999599633F, 0.9999615616F, 0.9999631116F, + 0.9999646144F, 0.9999660709F, 0.9999674820F, 0.9999688487F, + 0.9999701719F, 0.9999714526F, 0.9999726917F, 0.9999738900F, + 0.9999750486F, 0.9999761682F, 0.9999772497F, 0.9999782941F, + 0.9999793021F, 0.9999802747F, 0.9999812126F, 0.9999821167F, + 0.9999829878F, 0.9999838268F, 0.9999846343F, 0.9999854113F, + 0.9999861584F, 0.9999868765F, 0.9999875664F, 0.9999882287F, + 0.9999888642F, 0.9999894736F, 0.9999900577F, 0.9999906172F, + 0.9999911528F, 0.9999916651F, 0.9999921548F, 0.9999926227F, + 0.9999930693F, 0.9999934954F, 0.9999939015F, 0.9999942883F, + 0.9999946564F, 0.9999950064F, 0.9999953390F, 0.9999956547F, + 0.9999959541F, 0.9999962377F, 0.9999965062F, 0.9999967601F, + 0.9999969998F, 0.9999972260F, 0.9999974392F, 0.9999976399F, + 0.9999978285F, 0.9999980056F, 0.9999981716F, 0.9999983271F, + 0.9999984724F, 0.9999986081F, 0.9999987345F, 0.9999988521F, + 0.9999989613F, 0.9999990625F, 0.9999991562F, 0.9999992426F, + 0.9999993223F, 0.9999993954F, 0.9999994625F, 0.9999995239F, + 0.9999995798F, 0.9999996307F, 0.9999996768F, 0.9999997184F, + 0.9999997559F, 0.9999997895F, 0.9999998195F, 0.9999998462F, + 0.9999998698F, 0.9999998906F, 0.9999999088F, 0.9999999246F, + 0.9999999383F, 0.9999999500F, 0.9999999600F, 0.9999999684F, + 0.9999999754F, 0.9999999811F, 0.9999999858F, 0.9999999896F, + 0.9999999925F, 0.9999999948F, 0.9999999965F, 0.9999999978F, + 0.9999999986F, 0.9999999992F, 0.9999999996F, 0.9999999998F, + 0.9999999999F, 1.0000000000F, 1.0000000000F, 1.0000000000F, +}; + +static const float vwin8192[4096] = { + 0.0000000578F, 0.0000005198F, 0.0000014438F, 0.0000028299F, + 0.0000046780F, 0.0000069882F, 0.0000097604F, 0.0000129945F, + 0.0000166908F, 0.0000208490F, 0.0000254692F, 0.0000305515F, + 0.0000360958F, 0.0000421021F, 0.0000485704F, 0.0000555006F, + 0.0000628929F, 0.0000707472F, 0.0000790635F, 0.0000878417F, + 0.0000970820F, 0.0001067842F, 0.0001169483F, 0.0001275744F, + 0.0001386625F, 0.0001502126F, 0.0001622245F, 0.0001746984F, + 0.0001876343F, 0.0002010320F, 0.0002148917F, 0.0002292132F, + 0.0002439967F, 0.0002592421F, 0.0002749493F, 0.0002911184F, + 0.0003077493F, 0.0003248421F, 0.0003423967F, 0.0003604132F, + 0.0003788915F, 0.0003978316F, 0.0004172335F, 0.0004370971F, + 0.0004574226F, 0.0004782098F, 0.0004994587F, 0.0005211694F, + 0.0005433418F, 0.0005659759F, 0.0005890717F, 0.0006126292F, + 0.0006366484F, 0.0006611292F, 0.0006860716F, 0.0007114757F, + 0.0007373414F, 0.0007636687F, 0.0007904576F, 0.0008177080F, + 0.0008454200F, 0.0008735935F, 0.0009022285F, 0.0009313250F, + 0.0009608830F, 0.0009909025F, 0.0010213834F, 0.0010523257F, + 0.0010837295F, 0.0011155946F, 0.0011479211F, 0.0011807090F, + 0.0012139582F, 0.0012476687F, 0.0012818405F, 0.0013164736F, + 0.0013515679F, 0.0013871235F, 0.0014231402F, 0.0014596182F, + 0.0014965573F, 0.0015339576F, 0.0015718190F, 0.0016101415F, + 0.0016489251F, 0.0016881698F, 0.0017278754F, 0.0017680421F, + 0.0018086698F, 0.0018497584F, 0.0018913080F, 0.0019333185F, + 0.0019757898F, 0.0020187221F, 0.0020621151F, 0.0021059690F, + 0.0021502837F, 0.0021950591F, 0.0022402953F, 0.0022859921F, + 0.0023321497F, 0.0023787679F, 0.0024258467F, 0.0024733861F, + 0.0025213861F, 0.0025698466F, 0.0026187676F, 0.0026681491F, + 0.0027179911F, 0.0027682935F, 0.0028190562F, 0.0028702794F, + 0.0029219628F, 0.0029741066F, 0.0030267107F, 0.0030797749F, + 0.0031332994F, 0.0031872841F, 0.0032417289F, 0.0032966338F, + 0.0033519988F, 0.0034078238F, 0.0034641089F, 0.0035208539F, + 0.0035780589F, 0.0036357237F, 0.0036938485F, 0.0037524331F, + 0.0038114775F, 0.0038709817F, 0.0039309456F, 0.0039913692F, + 0.0040522524F, 0.0041135953F, 0.0041753978F, 0.0042376599F, + 0.0043003814F, 0.0043635624F, 0.0044272029F, 0.0044913028F, + 0.0045558620F, 0.0046208806F, 0.0046863585F, 0.0047522955F, + 0.0048186919F, 0.0048855473F, 0.0049528619F, 0.0050206356F, + 0.0050888684F, 0.0051575601F, 0.0052267108F, 0.0052963204F, + 0.0053663890F, 0.0054369163F, 0.0055079025F, 0.0055793474F, + 0.0056512510F, 0.0057236133F, 0.0057964342F, 0.0058697137F, + 0.0059434517F, 0.0060176482F, 0.0060923032F, 0.0061674166F, + 0.0062429883F, 0.0063190183F, 0.0063955066F, 0.0064724532F, + 0.0065498579F, 0.0066277207F, 0.0067060416F, 0.0067848205F, + 0.0068640575F, 0.0069437523F, 0.0070239051F, 0.0071045157F, + 0.0071855840F, 0.0072671102F, 0.0073490940F, 0.0074315355F, + 0.0075144345F, 0.0075977911F, 0.0076816052F, 0.0077658768F, + 0.0078506057F, 0.0079357920F, 0.0080214355F, 0.0081075363F, + 0.0081940943F, 0.0082811094F, 0.0083685816F, 0.0084565108F, + 0.0085448970F, 0.0086337401F, 0.0087230401F, 0.0088127969F, + 0.0089030104F, 0.0089936807F, 0.0090848076F, 0.0091763911F, + 0.0092684311F, 0.0093609276F, 0.0094538805F, 0.0095472898F, + 0.0096411554F, 0.0097354772F, 0.0098302552F, 0.0099254894F, + 0.0100211796F, 0.0101173259F, 0.0102139281F, 0.0103109863F, + 0.0104085002F, 0.0105064700F, 0.0106048955F, 0.0107037766F, + 0.0108031133F, 0.0109029056F, 0.0110031534F, 0.0111038565F, + 0.0112050151F, 0.0113066289F, 0.0114086980F, 0.0115112222F, + 0.0116142015F, 0.0117176359F, 0.0118215252F, 0.0119258695F, + 0.0120306686F, 0.0121359225F, 0.0122416312F, 0.0123477944F, + 0.0124544123F, 0.0125614847F, 0.0126690116F, 0.0127769928F, + 0.0128854284F, 0.0129943182F, 0.0131036623F, 0.0132134604F, + 0.0133237126F, 0.0134344188F, 0.0135455790F, 0.0136571929F, + 0.0137692607F, 0.0138817821F, 0.0139947572F, 0.0141081859F, + 0.0142220681F, 0.0143364037F, 0.0144511927F, 0.0145664350F, + 0.0146821304F, 0.0147982791F, 0.0149148808F, 0.0150319355F, + 0.0151494431F, 0.0152674036F, 0.0153858168F, 0.0155046828F, + 0.0156240014F, 0.0157437726F, 0.0158639962F, 0.0159846723F, + 0.0161058007F, 0.0162273814F, 0.0163494142F, 0.0164718991F, + 0.0165948361F, 0.0167182250F, 0.0168420658F, 0.0169663584F, + 0.0170911027F, 0.0172162987F, 0.0173419462F, 0.0174680452F, + 0.0175945956F, 0.0177215974F, 0.0178490504F, 0.0179769545F, + 0.0181053098F, 0.0182341160F, 0.0183633732F, 0.0184930812F, + 0.0186232399F, 0.0187538494F, 0.0188849094F, 0.0190164200F, + 0.0191483809F, 0.0192807923F, 0.0194136539F, 0.0195469656F, + 0.0196807275F, 0.0198149394F, 0.0199496012F, 0.0200847128F, + 0.0202202742F, 0.0203562853F, 0.0204927460F, 0.0206296561F, + 0.0207670157F, 0.0209048245F, 0.0210430826F, 0.0211817899F, + 0.0213209462F, 0.0214605515F, 0.0216006057F, 0.0217411086F, + 0.0218820603F, 0.0220234605F, 0.0221653093F, 0.0223076066F, + 0.0224503521F, 0.0225935459F, 0.0227371879F, 0.0228812779F, + 0.0230258160F, 0.0231708018F, 0.0233162355F, 0.0234621169F, + 0.0236084459F, 0.0237552224F, 0.0239024462F, 0.0240501175F, + 0.0241982359F, 0.0243468015F, 0.0244958141F, 0.0246452736F, + 0.0247951800F, 0.0249455331F, 0.0250963329F, 0.0252475792F, + 0.0253992720F, 0.0255514111F, 0.0257039965F, 0.0258570281F, + 0.0260105057F, 0.0261644293F, 0.0263187987F, 0.0264736139F, + 0.0266288747F, 0.0267845811F, 0.0269407330F, 0.0270973302F, + 0.0272543727F, 0.0274118604F, 0.0275697930F, 0.0277281707F, + 0.0278869932F, 0.0280462604F, 0.0282059723F, 0.0283661287F, + 0.0285267295F, 0.0286877747F, 0.0288492641F, 0.0290111976F, + 0.0291735751F, 0.0293363965F, 0.0294996617F, 0.0296633706F, + 0.0298275231F, 0.0299921190F, 0.0301571583F, 0.0303226409F, + 0.0304885667F, 0.0306549354F, 0.0308217472F, 0.0309890017F, + 0.0311566989F, 0.0313248388F, 0.0314934211F, 0.0316624459F, + 0.0318319128F, 0.0320018220F, 0.0321721732F, 0.0323429663F, + 0.0325142013F, 0.0326858779F, 0.0328579962F, 0.0330305559F, + 0.0332035570F, 0.0333769994F, 0.0335508829F, 0.0337252074F, + 0.0338999728F, 0.0340751790F, 0.0342508259F, 0.0344269134F, + 0.0346034412F, 0.0347804094F, 0.0349578178F, 0.0351356663F, + 0.0353139548F, 0.0354926831F, 0.0356718511F, 0.0358514588F, + 0.0360315059F, 0.0362119924F, 0.0363929182F, 0.0365742831F, + 0.0367560870F, 0.0369383297F, 0.0371210113F, 0.0373041315F, + 0.0374876902F, 0.0376716873F, 0.0378561226F, 0.0380409961F, + 0.0382263077F, 0.0384120571F, 0.0385982443F, 0.0387848691F, + 0.0389719315F, 0.0391594313F, 0.0393473683F, 0.0395357425F, + 0.0397245537F, 0.0399138017F, 0.0401034866F, 0.0402936080F, + 0.0404841660F, 0.0406751603F, 0.0408665909F, 0.0410584576F, + 0.0412507603F, 0.0414434988F, 0.0416366731F, 0.0418302829F, + 0.0420243282F, 0.0422188088F, 0.0424137246F, 0.0426090755F, + 0.0428048613F, 0.0430010819F, 0.0431977371F, 0.0433948269F, + 0.0435923511F, 0.0437903095F, 0.0439887020F, 0.0441875285F, + 0.0443867889F, 0.0445864830F, 0.0447866106F, 0.0449871717F, + 0.0451881661F, 0.0453895936F, 0.0455914542F, 0.0457937477F, + 0.0459964738F, 0.0461996326F, 0.0464032239F, 0.0466072475F, + 0.0468117032F, 0.0470165910F, 0.0472219107F, 0.0474276622F, + 0.0476338452F, 0.0478404597F, 0.0480475056F, 0.0482549827F, + 0.0484628907F, 0.0486712297F, 0.0488799994F, 0.0490891998F, + 0.0492988306F, 0.0495088917F, 0.0497193830F, 0.0499303043F, + 0.0501416554F, 0.0503534363F, 0.0505656468F, 0.0507782867F, + 0.0509913559F, 0.0512048542F, 0.0514187815F, 0.0516331376F, + 0.0518479225F, 0.0520631358F, 0.0522787775F, 0.0524948475F, + 0.0527113455F, 0.0529282715F, 0.0531456252F, 0.0533634066F, + 0.0535816154F, 0.0538002515F, 0.0540193148F, 0.0542388051F, + 0.0544587222F, 0.0546790660F, 0.0548998364F, 0.0551210331F, + 0.0553426561F, 0.0555647051F, 0.0557871801F, 0.0560100807F, + 0.0562334070F, 0.0564571587F, 0.0566813357F, 0.0569059378F, + 0.0571309649F, 0.0573564168F, 0.0575822933F, 0.0578085942F, + 0.0580353195F, 0.0582624689F, 0.0584900423F, 0.0587180396F, + 0.0589464605F, 0.0591753049F, 0.0594045726F, 0.0596342635F, + 0.0598643774F, 0.0600949141F, 0.0603258735F, 0.0605572555F, + 0.0607890597F, 0.0610212862F, 0.0612539346F, 0.0614870049F, + 0.0617204968F, 0.0619544103F, 0.0621887451F, 0.0624235010F, + 0.0626586780F, 0.0628942758F, 0.0631302942F, 0.0633667331F, + 0.0636035923F, 0.0638408717F, 0.0640785710F, 0.0643166901F, + 0.0645552288F, 0.0647941870F, 0.0650335645F, 0.0652733610F, + 0.0655135765F, 0.0657542108F, 0.0659952636F, 0.0662367348F, + 0.0664786242F, 0.0667209316F, 0.0669636570F, 0.0672068000F, + 0.0674503605F, 0.0676943384F, 0.0679387334F, 0.0681835454F, + 0.0684287742F, 0.0686744196F, 0.0689204814F, 0.0691669595F, + 0.0694138536F, 0.0696611637F, 0.0699088894F, 0.0701570307F, + 0.0704055873F, 0.0706545590F, 0.0709039458F, 0.0711537473F, + 0.0714039634F, 0.0716545939F, 0.0719056387F, 0.0721570975F, + 0.0724089702F, 0.0726612565F, 0.0729139563F, 0.0731670694F, + 0.0734205956F, 0.0736745347F, 0.0739288866F, 0.0741836510F, + 0.0744388277F, 0.0746944166F, 0.0749504175F, 0.0752068301F, + 0.0754636543F, 0.0757208899F, 0.0759785367F, 0.0762365946F, + 0.0764950632F, 0.0767539424F, 0.0770132320F, 0.0772729319F, + 0.0775330418F, 0.0777935616F, 0.0780544909F, 0.0783158298F, + 0.0785775778F, 0.0788397349F, 0.0791023009F, 0.0793652755F, + 0.0796286585F, 0.0798924498F, 0.0801566492F, 0.0804212564F, + 0.0806862712F, 0.0809516935F, 0.0812175231F, 0.0814837597F, + 0.0817504031F, 0.0820174532F, 0.0822849097F, 0.0825527724F, + 0.0828210412F, 0.0830897158F, 0.0833587960F, 0.0836282816F, + 0.0838981724F, 0.0841684682F, 0.0844391688F, 0.0847102740F, + 0.0849817835F, 0.0852536973F, 0.0855260150F, 0.0857987364F, + 0.0860718614F, 0.0863453897F, 0.0866193211F, 0.0868936554F, + 0.0871683924F, 0.0874435319F, 0.0877190737F, 0.0879950175F, + 0.0882713632F, 0.0885481105F, 0.0888252592F, 0.0891028091F, + 0.0893807600F, 0.0896591117F, 0.0899378639F, 0.0902170165F, + 0.0904965692F, 0.0907765218F, 0.0910568740F, 0.0913376258F, + 0.0916187767F, 0.0919003268F, 0.0921822756F, 0.0924646230F, + 0.0927473687F, 0.0930305126F, 0.0933140545F, 0.0935979940F, + 0.0938823310F, 0.0941670653F, 0.0944521966F, 0.0947377247F, + 0.0950236494F, 0.0953099704F, 0.0955966876F, 0.0958838007F, + 0.0961713094F, 0.0964592136F, 0.0967475131F, 0.0970362075F, + 0.0973252967F, 0.0976147805F, 0.0979046585F, 0.0981949307F, + 0.0984855967F, 0.0987766563F, 0.0990681093F, 0.0993599555F, + 0.0996521945F, 0.0999448263F, 0.1002378506F, 0.1005312671F, + 0.1008250755F, 0.1011192757F, 0.1014138675F, 0.1017088505F, + 0.1020042246F, 0.1022999895F, 0.1025961450F, 0.1028926909F, + 0.1031896268F, 0.1034869526F, 0.1037846680F, 0.1040827729F, + 0.1043812668F, 0.1046801497F, 0.1049794213F, 0.1052790813F, + 0.1055791294F, 0.1058795656F, 0.1061803894F, 0.1064816006F, + 0.1067831991F, 0.1070851846F, 0.1073875568F, 0.1076903155F, + 0.1079934604F, 0.1082969913F, 0.1086009079F, 0.1089052101F, + 0.1092098975F, 0.1095149699F, 0.1098204270F, 0.1101262687F, + 0.1104324946F, 0.1107391045F, 0.1110460982F, 0.1113534754F, + 0.1116612359F, 0.1119693793F, 0.1122779055F, 0.1125868142F, + 0.1128961052F, 0.1132057781F, 0.1135158328F, 0.1138262690F, + 0.1141370863F, 0.1144482847F, 0.1147598638F, 0.1150718233F, + 0.1153841631F, 0.1156968828F, 0.1160099822F, 0.1163234610F, + 0.1166373190F, 0.1169515559F, 0.1172661714F, 0.1175811654F, + 0.1178965374F, 0.1182122874F, 0.1185284149F, 0.1188449198F, + 0.1191618018F, 0.1194790606F, 0.1197966960F, 0.1201147076F, + 0.1204330953F, 0.1207518587F, 0.1210709976F, 0.1213905118F, + 0.1217104009F, 0.1220306647F, 0.1223513029F, 0.1226723153F, + 0.1229937016F, 0.1233154615F, 0.1236375948F, 0.1239601011F, + 0.1242829803F, 0.1246062319F, 0.1249298559F, 0.1252538518F, + 0.1255782195F, 0.1259029586F, 0.1262280689F, 0.1265535501F, + 0.1268794019F, 0.1272056241F, 0.1275322163F, 0.1278591784F, + 0.1281865099F, 0.1285142108F, 0.1288422805F, 0.1291707190F, + 0.1294995259F, 0.1298287009F, 0.1301582437F, 0.1304881542F, + 0.1308184319F, 0.1311490766F, 0.1314800881F, 0.1318114660F, + 0.1321432100F, 0.1324753200F, 0.1328077955F, 0.1331406364F, + 0.1334738422F, 0.1338074129F, 0.1341413479F, 0.1344756472F, + 0.1348103103F, 0.1351453370F, 0.1354807270F, 0.1358164801F, + 0.1361525959F, 0.1364890741F, 0.1368259145F, 0.1371631167F, + 0.1375006805F, 0.1378386056F, 0.1381768917F, 0.1385155384F, + 0.1388545456F, 0.1391939129F, 0.1395336400F, 0.1398737266F, + 0.1402141724F, 0.1405549772F, 0.1408961406F, 0.1412376623F, + 0.1415795421F, 0.1419217797F, 0.1422643746F, 0.1426073268F, + 0.1429506358F, 0.1432943013F, 0.1436383231F, 0.1439827008F, + 0.1443274342F, 0.1446725229F, 0.1450179667F, 0.1453637652F, + 0.1457099181F, 0.1460564252F, 0.1464032861F, 0.1467505006F, + 0.1470980682F, 0.1474459888F, 0.1477942620F, 0.1481428875F, + 0.1484918651F, 0.1488411942F, 0.1491908748F, 0.1495409065F, + 0.1498912889F, 0.1502420218F, 0.1505931048F, 0.1509445376F, + 0.1512963200F, 0.1516484516F, 0.1520009321F, 0.1523537612F, + 0.1527069385F, 0.1530604638F, 0.1534143368F, 0.1537685571F, + 0.1541231244F, 0.1544780384F, 0.1548332987F, 0.1551889052F, + 0.1555448574F, 0.1559011550F, 0.1562577978F, 0.1566147853F, + 0.1569721173F, 0.1573297935F, 0.1576878135F, 0.1580461771F, + 0.1584048838F, 0.1587639334F, 0.1591233255F, 0.1594830599F, + 0.1598431361F, 0.1602035540F, 0.1605643131F, 0.1609254131F, + 0.1612868537F, 0.1616486346F, 0.1620107555F, 0.1623732160F, + 0.1627360158F, 0.1630991545F, 0.1634626319F, 0.1638264476F, + 0.1641906013F, 0.1645550926F, 0.1649199212F, 0.1652850869F, + 0.1656505892F, 0.1660164278F, 0.1663826024F, 0.1667491127F, + 0.1671159583F, 0.1674831388F, 0.1678506541F, 0.1682185036F, + 0.1685866872F, 0.1689552044F, 0.1693240549F, 0.1696932384F, + 0.1700627545F, 0.1704326029F, 0.1708027833F, 0.1711732952F, + 0.1715441385F, 0.1719153127F, 0.1722868175F, 0.1726586526F, + 0.1730308176F, 0.1734033121F, 0.1737761359F, 0.1741492886F, + 0.1745227698F, 0.1748965792F, 0.1752707164F, 0.1756451812F, + 0.1760199731F, 0.1763950918F, 0.1767705370F, 0.1771463083F, + 0.1775224054F, 0.1778988279F, 0.1782755754F, 0.1786526477F, + 0.1790300444F, 0.1794077651F, 0.1797858094F, 0.1801641771F, + 0.1805428677F, 0.1809218810F, 0.1813012165F, 0.1816808739F, + 0.1820608528F, 0.1824411530F, 0.1828217739F, 0.1832027154F, + 0.1835839770F, 0.1839655584F, 0.1843474592F, 0.1847296790F, + 0.1851122175F, 0.1854950744F, 0.1858782492F, 0.1862617417F, + 0.1866455514F, 0.1870296780F, 0.1874141211F, 0.1877988804F, + 0.1881839555F, 0.1885693461F, 0.1889550517F, 0.1893410721F, + 0.1897274068F, 0.1901140555F, 0.1905010178F, 0.1908882933F, + 0.1912758818F, 0.1916637828F, 0.1920519959F, 0.1924405208F, + 0.1928293571F, 0.1932185044F, 0.1936079625F, 0.1939977308F, + 0.1943878091F, 0.1947781969F, 0.1951688939F, 0.1955598998F, + 0.1959512141F, 0.1963428364F, 0.1967347665F, 0.1971270038F, + 0.1975195482F, 0.1979123990F, 0.1983055561F, 0.1986990190F, + 0.1990927873F, 0.1994868607F, 0.1998812388F, 0.2002759212F, + 0.2006709075F, 0.2010661974F, 0.2014617904F, 0.2018576862F, + 0.2022538844F, 0.2026503847F, 0.2030471865F, 0.2034442897F, + 0.2038416937F, 0.2042393982F, 0.2046374028F, 0.2050357071F, + 0.2054343107F, 0.2058332133F, 0.2062324145F, 0.2066319138F, + 0.2070317110F, 0.2074318055F, 0.2078321970F, 0.2082328852F, + 0.2086338696F, 0.2090351498F, 0.2094367255F, 0.2098385962F, + 0.2102407617F, 0.2106432213F, 0.2110459749F, 0.2114490220F, + 0.2118523621F, 0.2122559950F, 0.2126599202F, 0.2130641373F, + 0.2134686459F, 0.2138734456F, 0.2142785361F, 0.2146839168F, + 0.2150895875F, 0.2154955478F, 0.2159017972F, 0.2163083353F, + 0.2167151617F, 0.2171222761F, 0.2175296780F, 0.2179373670F, + 0.2183453428F, 0.2187536049F, 0.2191621529F, 0.2195709864F, + 0.2199801051F, 0.2203895085F, 0.2207991961F, 0.2212091677F, + 0.2216194228F, 0.2220299610F, 0.2224407818F, 0.2228518850F, + 0.2232632699F, 0.2236749364F, 0.2240868839F, 0.2244991121F, + 0.2249116204F, 0.2253244086F, 0.2257374763F, 0.2261508229F, + 0.2265644481F, 0.2269783514F, 0.2273925326F, 0.2278069911F, + 0.2282217265F, 0.2286367384F, 0.2290520265F, 0.2294675902F, + 0.2298834292F, 0.2302995431F, 0.2307159314F, 0.2311325937F, + 0.2315495297F, 0.2319667388F, 0.2323842207F, 0.2328019749F, + 0.2332200011F, 0.2336382988F, 0.2340568675F, 0.2344757070F, + 0.2348948166F, 0.2353141961F, 0.2357338450F, 0.2361537629F, + 0.2365739493F, 0.2369944038F, 0.2374151261F, 0.2378361156F, + 0.2382573720F, 0.2386788948F, 0.2391006836F, 0.2395227380F, + 0.2399450575F, 0.2403676417F, 0.2407904902F, 0.2412136026F, + 0.2416369783F, 0.2420606171F, 0.2424845185F, 0.2429086820F, + 0.2433331072F, 0.2437577936F, 0.2441827409F, 0.2446079486F, + 0.2450334163F, 0.2454591435F, 0.2458851298F, 0.2463113747F, + 0.2467378779F, 0.2471646389F, 0.2475916573F, 0.2480189325F, + 0.2484464643F, 0.2488742521F, 0.2493022955F, 0.2497305940F, + 0.2501591473F, 0.2505879549F, 0.2510170163F, 0.2514463311F, + 0.2518758989F, 0.2523057193F, 0.2527357916F, 0.2531661157F, + 0.2535966909F, 0.2540275169F, 0.2544585931F, 0.2548899193F, + 0.2553214948F, 0.2557533193F, 0.2561853924F, 0.2566177135F, + 0.2570502822F, 0.2574830981F, 0.2579161608F, 0.2583494697F, + 0.2587830245F, 0.2592168246F, 0.2596508697F, 0.2600851593F, + 0.2605196929F, 0.2609544701F, 0.2613894904F, 0.2618247534F, + 0.2622602586F, 0.2626960055F, 0.2631319938F, 0.2635682230F, + 0.2640046925F, 0.2644414021F, 0.2648783511F, 0.2653155391F, + 0.2657529657F, 0.2661906305F, 0.2666285329F, 0.2670666725F, + 0.2675050489F, 0.2679436616F, 0.2683825101F, 0.2688215940F, + 0.2692609127F, 0.2697004660F, 0.2701402532F, 0.2705802739F, + 0.2710205278F, 0.2714610142F, 0.2719017327F, 0.2723426830F, + 0.2727838644F, 0.2732252766F, 0.2736669191F, 0.2741087914F, + 0.2745508930F, 0.2749932235F, 0.2754357824F, 0.2758785693F, + 0.2763215837F, 0.2767648251F, 0.2772082930F, 0.2776519870F, + 0.2780959066F, 0.2785400513F, 0.2789844207F, 0.2794290143F, + 0.2798738316F, 0.2803188722F, 0.2807641355F, 0.2812096211F, + 0.2816553286F, 0.2821012574F, 0.2825474071F, 0.2829937773F, + 0.2834403673F, 0.2838871768F, 0.2843342053F, 0.2847814523F, + 0.2852289174F, 0.2856765999F, 0.2861244996F, 0.2865726159F, + 0.2870209482F, 0.2874694962F, 0.2879182594F, 0.2883672372F, + 0.2888164293F, 0.2892658350F, 0.2897154540F, 0.2901652858F, + 0.2906153298F, 0.2910655856F, 0.2915160527F, 0.2919667306F, + 0.2924176189F, 0.2928687171F, 0.2933200246F, 0.2937715409F, + 0.2942232657F, 0.2946751984F, 0.2951273386F, 0.2955796856F, + 0.2960322391F, 0.2964849986F, 0.2969379636F, 0.2973911335F, + 0.2978445080F, 0.2982980864F, 0.2987518684F, 0.2992058534F, + 0.2996600409F, 0.3001144305F, 0.3005690217F, 0.3010238139F, + 0.3014788067F, 0.3019339995F, 0.3023893920F, 0.3028449835F, + 0.3033007736F, 0.3037567618F, 0.3042129477F, 0.3046693306F, + 0.3051259102F, 0.3055826859F, 0.3060396572F, 0.3064968236F, + 0.3069541847F, 0.3074117399F, 0.3078694887F, 0.3083274307F, + 0.3087855653F, 0.3092438920F, 0.3097024104F, 0.3101611199F, + 0.3106200200F, 0.3110791103F, 0.3115383902F, 0.3119978592F, + 0.3124575169F, 0.3129173627F, 0.3133773961F, 0.3138376166F, + 0.3142980238F, 0.3147586170F, 0.3152193959F, 0.3156803598F, + 0.3161415084F, 0.3166028410F, 0.3170643573F, 0.3175260566F, + 0.3179879384F, 0.3184500023F, 0.3189122478F, 0.3193746743F, + 0.3198372814F, 0.3203000685F, 0.3207630351F, 0.3212261807F, + 0.3216895048F, 0.3221530069F, 0.3226166865F, 0.3230805430F, + 0.3235445760F, 0.3240087849F, 0.3244731693F, 0.3249377285F, + 0.3254024622F, 0.3258673698F, 0.3263324507F, 0.3267977045F, + 0.3272631306F, 0.3277287286F, 0.3281944978F, 0.3286604379F, + 0.3291265482F, 0.3295928284F, 0.3300592777F, 0.3305258958F, + 0.3309926821F, 0.3314596361F, 0.3319267573F, 0.3323940451F, + 0.3328614990F, 0.3333291186F, 0.3337969033F, 0.3342648525F, + 0.3347329658F, 0.3352012427F, 0.3356696825F, 0.3361382849F, + 0.3366070492F, 0.3370759749F, 0.3375450616F, 0.3380143087F, + 0.3384837156F, 0.3389532819F, 0.3394230071F, 0.3398928905F, + 0.3403629317F, 0.3408331302F, 0.3413034854F, 0.3417739967F, + 0.3422446638F, 0.3427154860F, 0.3431864628F, 0.3436575938F, + 0.3441288782F, 0.3446003158F, 0.3450719058F, 0.3455436478F, + 0.3460155412F, 0.3464875856F, 0.3469597804F, 0.3474321250F, + 0.3479046189F, 0.3483772617F, 0.3488500527F, 0.3493229914F, + 0.3497960774F, 0.3502693100F, 0.3507426887F, 0.3512162131F, + 0.3516898825F, 0.3521636965F, 0.3526376545F, 0.3531117559F, + 0.3535860003F, 0.3540603870F, 0.3545349157F, 0.3550095856F, + 0.3554843964F, 0.3559593474F, 0.3564344381F, 0.3569096680F, + 0.3573850366F, 0.3578605432F, 0.3583361875F, 0.3588119687F, + 0.3592878865F, 0.3597639402F, 0.3602401293F, 0.3607164533F, + 0.3611929117F, 0.3616695038F, 0.3621462292F, 0.3626230873F, + 0.3631000776F, 0.3635771995F, 0.3640544525F, 0.3645318360F, + 0.3650093496F, 0.3654869926F, 0.3659647645F, 0.3664426648F, + 0.3669206930F, 0.3673988484F, 0.3678771306F, 0.3683555390F, + 0.3688340731F, 0.3693127322F, 0.3697915160F, 0.3702704237F, + 0.3707494549F, 0.3712286091F, 0.3717078857F, 0.3721872840F, + 0.3726668037F, 0.3731464441F, 0.3736262047F, 0.3741060850F, + 0.3745860843F, 0.3750662023F, 0.3755464382F, 0.3760267915F, + 0.3765072618F, 0.3769878484F, 0.3774685509F, 0.3779493686F, + 0.3784303010F, 0.3789113475F, 0.3793925076F, 0.3798737809F, + 0.3803551666F, 0.3808366642F, 0.3813182733F, 0.3817999932F, + 0.3822818234F, 0.3827637633F, 0.3832458124F, 0.3837279702F, + 0.3842102360F, 0.3846926093F, 0.3851750897F, 0.3856576764F, + 0.3861403690F, 0.3866231670F, 0.3871060696F, 0.3875890765F, + 0.3880721870F, 0.3885554007F, 0.3890387168F, 0.3895221349F, + 0.3900056544F, 0.3904892748F, 0.3909729955F, 0.3914568160F, + 0.3919407356F, 0.3924247539F, 0.3929088702F, 0.3933930841F, + 0.3938773949F, 0.3943618021F, 0.3948463052F, 0.3953309035F, + 0.3958155966F, 0.3963003838F, 0.3967852646F, 0.3972702385F, + 0.3977553048F, 0.3982404631F, 0.3987257127F, 0.3992110531F, + 0.3996964838F, 0.4001820041F, 0.4006676136F, 0.4011533116F, + 0.4016390976F, 0.4021249710F, 0.4026109313F, 0.4030969779F, + 0.4035831102F, 0.4040693277F, 0.4045556299F, 0.4050420160F, + 0.4055284857F, 0.4060150383F, 0.4065016732F, 0.4069883899F, + 0.4074751879F, 0.4079620665F, 0.4084490252F, 0.4089360635F, + 0.4094231807F, 0.4099103763F, 0.4103976498F, 0.4108850005F, + 0.4113724280F, 0.4118599315F, 0.4123475107F, 0.4128351648F, + 0.4133228934F, 0.4138106959F, 0.4142985716F, 0.4147865201F, + 0.4152745408F, 0.4157626330F, 0.4162507963F, 0.4167390301F, + 0.4172273337F, 0.4177157067F, 0.4182041484F, 0.4186926583F, + 0.4191812359F, 0.4196698805F, 0.4201585915F, 0.4206473685F, + 0.4211362108F, 0.4216251179F, 0.4221140892F, 0.4226031241F, + 0.4230922221F, 0.4235813826F, 0.4240706050F, 0.4245598887F, + 0.4250492332F, 0.4255386379F, 0.4260281022F, 0.4265176256F, + 0.4270072075F, 0.4274968473F, 0.4279865445F, 0.4284762984F, + 0.4289661086F, 0.4294559743F, 0.4299458951F, 0.4304358704F, + 0.4309258996F, 0.4314159822F, 0.4319061175F, 0.4323963050F, + 0.4328865441F, 0.4333768342F, 0.4338671749F, 0.4343575654F, + 0.4348480052F, 0.4353384938F, 0.4358290306F, 0.4363196149F, + 0.4368102463F, 0.4373009241F, 0.4377916478F, 0.4382824168F, + 0.4387732305F, 0.4392640884F, 0.4397549899F, 0.4402459343F, + 0.4407369212F, 0.4412279499F, 0.4417190198F, 0.4422101305F, + 0.4427012813F, 0.4431924717F, 0.4436837010F, 0.4441749686F, + 0.4446662742F, 0.4451576169F, 0.4456489963F, 0.4461404118F, + 0.4466318628F, 0.4471233487F, 0.4476148690F, 0.4481064230F, + 0.4485980103F, 0.4490896302F, 0.4495812821F, 0.4500729654F, + 0.4505646797F, 0.4510564243F, 0.4515481986F, 0.4520400021F, + 0.4525318341F, 0.4530236942F, 0.4535155816F, 0.4540074959F, + 0.4544994365F, 0.4549914028F, 0.4554833941F, 0.4559754100F, + 0.4564674499F, 0.4569595131F, 0.4574515991F, 0.4579437074F, + 0.4584358372F, 0.4589279881F, 0.4594201595F, 0.4599123508F, + 0.4604045615F, 0.4608967908F, 0.4613890383F, 0.4618813034F, + 0.4623735855F, 0.4628658841F, 0.4633581984F, 0.4638505281F, + 0.4643428724F, 0.4648352308F, 0.4653276028F, 0.4658199877F, + 0.4663123849F, 0.4668047940F, 0.4672972143F, 0.4677896451F, + 0.4682820861F, 0.4687745365F, 0.4692669958F, 0.4697594634F, + 0.4702519387F, 0.4707444211F, 0.4712369102F, 0.4717294052F, + 0.4722219056F, 0.4727144109F, 0.4732069204F, 0.4736994336F, + 0.4741919498F, 0.4746844686F, 0.4751769893F, 0.4756695113F, + 0.4761620341F, 0.4766545571F, 0.4771470797F, 0.4776396013F, + 0.4781321213F, 0.4786246392F, 0.4791171544F, 0.4796096663F, + 0.4801021744F, 0.4805946779F, 0.4810871765F, 0.4815796694F, + 0.4820721561F, 0.4825646360F, 0.4830571086F, 0.4835495732F, + 0.4840420293F, 0.4845344763F, 0.4850269136F, 0.4855193407F, + 0.4860117569F, 0.4865041617F, 0.4869965545F, 0.4874889347F, + 0.4879813018F, 0.4884736551F, 0.4889659941F, 0.4894583182F, + 0.4899506268F, 0.4904429193F, 0.4909351952F, 0.4914274538F, + 0.4919196947F, 0.4924119172F, 0.4929041207F, 0.4933963046F, + 0.4938884685F, 0.4943806116F, 0.4948727335F, 0.4953648335F, + 0.4958569110F, 0.4963489656F, 0.4968409965F, 0.4973330032F, + 0.4978249852F, 0.4983169419F, 0.4988088726F, 0.4993007768F, + 0.4997926539F, 0.5002845034F, 0.5007763247F, 0.5012681171F, + 0.5017598801F, 0.5022516132F, 0.5027433157F, 0.5032349871F, + 0.5037266268F, 0.5042182341F, 0.5047098086F, 0.5052013497F, + 0.5056928567F, 0.5061843292F, 0.5066757664F, 0.5071671679F, + 0.5076585330F, 0.5081498613F, 0.5086411520F, 0.5091324047F, + 0.5096236187F, 0.5101147934F, 0.5106059284F, 0.5110970230F, + 0.5115880766F, 0.5120790887F, 0.5125700587F, 0.5130609860F, + 0.5135518700F, 0.5140427102F, 0.5145335059F, 0.5150242566F, + 0.5155149618F, 0.5160056208F, 0.5164962331F, 0.5169867980F, + 0.5174773151F, 0.5179677837F, 0.5184582033F, 0.5189485733F, + 0.5194388931F, 0.5199291621F, 0.5204193798F, 0.5209095455F, + 0.5213996588F, 0.5218897190F, 0.5223797256F, 0.5228696779F, + 0.5233595755F, 0.5238494177F, 0.5243392039F, 0.5248289337F, + 0.5253186063F, 0.5258082213F, 0.5262977781F, 0.5267872760F, + 0.5272767146F, 0.5277660932F, 0.5282554112F, 0.5287446682F, + 0.5292338635F, 0.5297229965F, 0.5302120667F, 0.5307010736F, + 0.5311900164F, 0.5316788947F, 0.5321677079F, 0.5326564554F, + 0.5331451366F, 0.5336337511F, 0.5341222981F, 0.5346107771F, + 0.5350991876F, 0.5355875290F, 0.5360758007F, 0.5365640021F, + 0.5370521327F, 0.5375401920F, 0.5380281792F, 0.5385160939F, + 0.5390039355F, 0.5394917034F, 0.5399793971F, 0.5404670159F, + 0.5409545594F, 0.5414420269F, 0.5419294179F, 0.5424167318F, + 0.5429039680F, 0.5433911261F, 0.5438782053F, 0.5443652051F, + 0.5448521250F, 0.5453389644F, 0.5458257228F, 0.5463123995F, + 0.5467989940F, 0.5472855057F, 0.5477719341F, 0.5482582786F, + 0.5487445387F, 0.5492307137F, 0.5497168031F, 0.5502028063F, + 0.5506887228F, 0.5511745520F, 0.5516602934F, 0.5521459463F, + 0.5526315103F, 0.5531169847F, 0.5536023690F, 0.5540876626F, + 0.5545728649F, 0.5550579755F, 0.5555429937F, 0.5560279189F, + 0.5565127507F, 0.5569974884F, 0.5574821315F, 0.5579666794F, + 0.5584511316F, 0.5589354875F, 0.5594197465F, 0.5599039080F, + 0.5603879716F, 0.5608719367F, 0.5613558026F, 0.5618395689F, + 0.5623232350F, 0.5628068002F, 0.5632902642F, 0.5637736262F, + 0.5642568858F, 0.5647400423F, 0.5652230953F, 0.5657060442F, + 0.5661888883F, 0.5666716272F, 0.5671542603F, 0.5676367870F, + 0.5681192069F, 0.5686015192F, 0.5690837235F, 0.5695658192F, + 0.5700478058F, 0.5705296827F, 0.5710114494F, 0.5714931052F, + 0.5719746497F, 0.5724560822F, 0.5729374023F, 0.5734186094F, + 0.5738997029F, 0.5743806823F, 0.5748615470F, 0.5753422965F, + 0.5758229301F, 0.5763034475F, 0.5767838480F, 0.5772641310F, + 0.5777442960F, 0.5782243426F, 0.5787042700F, 0.5791840778F, + 0.5796637654F, 0.5801433322F, 0.5806227778F, 0.5811021016F, + 0.5815813029F, 0.5820603814F, 0.5825393363F, 0.5830181673F, + 0.5834968737F, 0.5839754549F, 0.5844539105F, 0.5849322399F, + 0.5854104425F, 0.5858885179F, 0.5863664653F, 0.5868442844F, + 0.5873219746F, 0.5877995353F, 0.5882769660F, 0.5887542661F, + 0.5892314351F, 0.5897084724F, 0.5901853776F, 0.5906621500F, + 0.5911387892F, 0.5916152945F, 0.5920916655F, 0.5925679016F, + 0.5930440022F, 0.5935199669F, 0.5939957950F, 0.5944714861F, + 0.5949470396F, 0.5954224550F, 0.5958977317F, 0.5963728692F, + 0.5968478669F, 0.5973227244F, 0.5977974411F, 0.5982720163F, + 0.5987464497F, 0.5992207407F, 0.5996948887F, 0.6001688932F, + 0.6006427537F, 0.6011164696F, 0.6015900405F, 0.6020634657F, + 0.6025367447F, 0.6030098770F, 0.6034828621F, 0.6039556995F, + 0.6044283885F, 0.6049009288F, 0.6053733196F, 0.6058455606F, + 0.6063176512F, 0.6067895909F, 0.6072613790F, 0.6077330152F, + 0.6082044989F, 0.6086758295F, 0.6091470065F, 0.6096180294F, + 0.6100888977F, 0.6105596108F, 0.6110301682F, 0.6115005694F, + 0.6119708139F, 0.6124409011F, 0.6129108305F, 0.6133806017F, + 0.6138502139F, 0.6143196669F, 0.6147889599F, 0.6152580926F, + 0.6157270643F, 0.6161958746F, 0.6166645230F, 0.6171330088F, + 0.6176013317F, 0.6180694910F, 0.6185374863F, 0.6190053171F, + 0.6194729827F, 0.6199404828F, 0.6204078167F, 0.6208749841F, + 0.6213419842F, 0.6218088168F, 0.6222754811F, 0.6227419768F, + 0.6232083032F, 0.6236744600F, 0.6241404465F, 0.6246062622F, + 0.6250719067F, 0.6255373795F, 0.6260026799F, 0.6264678076F, + 0.6269327619F, 0.6273975425F, 0.6278621487F, 0.6283265800F, + 0.6287908361F, 0.6292549163F, 0.6297188201F, 0.6301825471F, + 0.6306460966F, 0.6311094683F, 0.6315726617F, 0.6320356761F, + 0.6324985111F, 0.6329611662F, 0.6334236410F, 0.6338859348F, + 0.6343480472F, 0.6348099777F, 0.6352717257F, 0.6357332909F, + 0.6361946726F, 0.6366558704F, 0.6371168837F, 0.6375777122F, + 0.6380383552F, 0.6384988123F, 0.6389590830F, 0.6394191668F, + 0.6398790631F, 0.6403387716F, 0.6407982916F, 0.6412576228F, + 0.6417167645F, 0.6421757163F, 0.6426344778F, 0.6430930483F, + 0.6435514275F, 0.6440096149F, 0.6444676098F, 0.6449254119F, + 0.6453830207F, 0.6458404356F, 0.6462976562F, 0.6467546820F, + 0.6472115125F, 0.6476681472F, 0.6481245856F, 0.6485808273F, + 0.6490368717F, 0.6494927183F, 0.6499483667F, 0.6504038164F, + 0.6508590670F, 0.6513141178F, 0.6517689684F, 0.6522236185F, + 0.6526780673F, 0.6531323146F, 0.6535863598F, 0.6540402024F, + 0.6544938419F, 0.6549472779F, 0.6554005099F, 0.6558535373F, + 0.6563063598F, 0.6567589769F, 0.6572113880F, 0.6576635927F, + 0.6581155906F, 0.6585673810F, 0.6590189637F, 0.6594703380F, + 0.6599215035F, 0.6603724598F, 0.6608232064F, 0.6612737427F, + 0.6617240684F, 0.6621741829F, 0.6626240859F, 0.6630737767F, + 0.6635232550F, 0.6639725202F, 0.6644215720F, 0.6648704098F, + 0.6653190332F, 0.6657674417F, 0.6662156348F, 0.6666636121F, + 0.6671113731F, 0.6675589174F, 0.6680062445F, 0.6684533538F, + 0.6689002450F, 0.6693469177F, 0.6697933712F, 0.6702396052F, + 0.6706856193F, 0.6711314129F, 0.6715769855F, 0.6720223369F, + 0.6724674664F, 0.6729123736F, 0.6733570581F, 0.6738015194F, + 0.6742457570F, 0.6746897706F, 0.6751335596F, 0.6755771236F, + 0.6760204621F, 0.6764635747F, 0.6769064609F, 0.6773491204F, + 0.6777915525F, 0.6782337570F, 0.6786757332F, 0.6791174809F, + 0.6795589995F, 0.6800002886F, 0.6804413477F, 0.6808821765F, + 0.6813227743F, 0.6817631409F, 0.6822032758F, 0.6826431785F, + 0.6830828485F, 0.6835222855F, 0.6839614890F, 0.6844004585F, + 0.6848391936F, 0.6852776939F, 0.6857159589F, 0.6861539883F, + 0.6865917815F, 0.6870293381F, 0.6874666576F, 0.6879037398F, + 0.6883405840F, 0.6887771899F, 0.6892135571F, 0.6896496850F, + 0.6900855733F, 0.6905212216F, 0.6909566294F, 0.6913917963F, + 0.6918267218F, 0.6922614055F, 0.6926958471F, 0.6931300459F, + 0.6935640018F, 0.6939977141F, 0.6944311825F, 0.6948644066F, + 0.6952973859F, 0.6957301200F, 0.6961626085F, 0.6965948510F, + 0.6970268470F, 0.6974585961F, 0.6978900980F, 0.6983213521F, + 0.6987523580F, 0.6991831154F, 0.6996136238F, 0.7000438828F, + 0.7004738921F, 0.7009036510F, 0.7013331594F, 0.7017624166F, + 0.7021914224F, 0.7026201763F, 0.7030486779F, 0.7034769268F, + 0.7039049226F, 0.7043326648F, 0.7047601531F, 0.7051873870F, + 0.7056143662F, 0.7060410902F, 0.7064675586F, 0.7068937711F, + 0.7073197271F, 0.7077454264F, 0.7081708684F, 0.7085960529F, + 0.7090209793F, 0.7094456474F, 0.7098700566F, 0.7102942066F, + 0.7107180970F, 0.7111417274F, 0.7115650974F, 0.7119882066F, + 0.7124110545F, 0.7128336409F, 0.7132559653F, 0.7136780272F, + 0.7140998264F, 0.7145213624F, 0.7149426348F, 0.7153636433F, + 0.7157843874F, 0.7162048668F, 0.7166250810F, 0.7170450296F, + 0.7174647124F, 0.7178841289F, 0.7183032786F, 0.7187221613F, + 0.7191407765F, 0.7195591239F, 0.7199772030F, 0.7203950135F, + 0.7208125550F, 0.7212298271F, 0.7216468294F, 0.7220635616F, + 0.7224800233F, 0.7228962140F, 0.7233121335F, 0.7237277813F, + 0.7241431571F, 0.7245582604F, 0.7249730910F, 0.7253876484F, + 0.7258019322F, 0.7262159422F, 0.7266296778F, 0.7270431388F, + 0.7274563247F, 0.7278692353F, 0.7282818700F, 0.7286942287F, + 0.7291063108F, 0.7295181160F, 0.7299296440F, 0.7303408944F, + 0.7307518669F, 0.7311625609F, 0.7315729763F, 0.7319831126F, + 0.7323929695F, 0.7328025466F, 0.7332118435F, 0.7336208600F, + 0.7340295955F, 0.7344380499F, 0.7348462226F, 0.7352541134F, + 0.7356617220F, 0.7360690478F, 0.7364760907F, 0.7368828502F, + 0.7372893259F, 0.7376955176F, 0.7381014249F, 0.7385070475F, + 0.7389123849F, 0.7393174368F, 0.7397222029F, 0.7401266829F, + 0.7405308763F, 0.7409347829F, 0.7413384023F, 0.7417417341F, + 0.7421447780F, 0.7425475338F, 0.7429500009F, 0.7433521791F, + 0.7437540681F, 0.7441556674F, 0.7445569769F, 0.7449579960F, + 0.7453587245F, 0.7457591621F, 0.7461593084F, 0.7465591631F, + 0.7469587259F, 0.7473579963F, 0.7477569741F, 0.7481556590F, + 0.7485540506F, 0.7489521486F, 0.7493499526F, 0.7497474623F, + 0.7501446775F, 0.7505415977F, 0.7509382227F, 0.7513345521F, + 0.7517305856F, 0.7521263229F, 0.7525217636F, 0.7529169074F, + 0.7533117541F, 0.7537063032F, 0.7541005545F, 0.7544945076F, + 0.7548881623F, 0.7552815182F, 0.7556745749F, 0.7560673323F, + 0.7564597899F, 0.7568519474F, 0.7572438046F, 0.7576353611F, + 0.7580266166F, 0.7584175708F, 0.7588082235F, 0.7591985741F, + 0.7595886226F, 0.7599783685F, 0.7603678116F, 0.7607569515F, + 0.7611457879F, 0.7615343206F, 0.7619225493F, 0.7623104735F, + 0.7626980931F, 0.7630854078F, 0.7634724171F, 0.7638591209F, + 0.7642455188F, 0.7646316106F, 0.7650173959F, 0.7654028744F, + 0.7657880459F, 0.7661729100F, 0.7665574664F, 0.7669417150F, + 0.7673256553F, 0.7677092871F, 0.7680926100F, 0.7684756239F, + 0.7688583284F, 0.7692407232F, 0.7696228080F, 0.7700045826F, + 0.7703860467F, 0.7707671999F, 0.7711480420F, 0.7715285728F, + 0.7719087918F, 0.7722886989F, 0.7726682938F, 0.7730475762F, + 0.7734265458F, 0.7738052023F, 0.7741835454F, 0.7745615750F, + 0.7749392906F, 0.7753166921F, 0.7756937791F, 0.7760705514F, + 0.7764470087F, 0.7768231508F, 0.7771989773F, 0.7775744880F, + 0.7779496827F, 0.7783245610F, 0.7786991227F, 0.7790733676F, + 0.7794472953F, 0.7798209056F, 0.7801941982F, 0.7805671729F, + 0.7809398294F, 0.7813121675F, 0.7816841869F, 0.7820558873F, + 0.7824272684F, 0.7827983301F, 0.7831690720F, 0.7835394940F, + 0.7839095957F, 0.7842793768F, 0.7846488373F, 0.7850179767F, + 0.7853867948F, 0.7857552914F, 0.7861234663F, 0.7864913191F, + 0.7868588497F, 0.7872260578F, 0.7875929431F, 0.7879595055F, + 0.7883257445F, 0.7886916601F, 0.7890572520F, 0.7894225198F, + 0.7897874635F, 0.7901520827F, 0.7905163772F, 0.7908803468F, + 0.7912439912F, 0.7916073102F, 0.7919703035F, 0.7923329710F, + 0.7926953124F, 0.7930573274F, 0.7934190158F, 0.7937803774F, + 0.7941414120F, 0.7945021193F, 0.7948624991F, 0.7952225511F, + 0.7955822752F, 0.7959416711F, 0.7963007387F, 0.7966594775F, + 0.7970178875F, 0.7973759685F, 0.7977337201F, 0.7980911422F, + 0.7984482346F, 0.7988049970F, 0.7991614292F, 0.7995175310F, + 0.7998733022F, 0.8002287426F, 0.8005838519F, 0.8009386299F, + 0.8012930765F, 0.8016471914F, 0.8020009744F, 0.8023544253F, + 0.8027075438F, 0.8030603298F, 0.8034127831F, 0.8037649035F, + 0.8041166906F, 0.8044681445F, 0.8048192647F, 0.8051700512F, + 0.8055205038F, 0.8058706222F, 0.8062204062F, 0.8065698556F, + 0.8069189702F, 0.8072677499F, 0.8076161944F, 0.8079643036F, + 0.8083120772F, 0.8086595151F, 0.8090066170F, 0.8093533827F, + 0.8096998122F, 0.8100459051F, 0.8103916613F, 0.8107370806F, + 0.8110821628F, 0.8114269077F, 0.8117713151F, 0.8121153849F, + 0.8124591169F, 0.8128025108F, 0.8131455666F, 0.8134882839F, + 0.8138306627F, 0.8141727027F, 0.8145144038F, 0.8148557658F, + 0.8151967886F, 0.8155374718F, 0.8158778154F, 0.8162178192F, + 0.8165574830F, 0.8168968067F, 0.8172357900F, 0.8175744328F, + 0.8179127349F, 0.8182506962F, 0.8185883164F, 0.8189255955F, + 0.8192625332F, 0.8195991295F, 0.8199353840F, 0.8202712967F, + 0.8206068673F, 0.8209420958F, 0.8212769820F, 0.8216115256F, + 0.8219457266F, 0.8222795848F, 0.8226131000F, 0.8229462721F, + 0.8232791009F, 0.8236115863F, 0.8239437280F, 0.8242755260F, + 0.8246069801F, 0.8249380901F, 0.8252688559F, 0.8255992774F, + 0.8259293544F, 0.8262590867F, 0.8265884741F, 0.8269175167F, + 0.8272462141F, 0.8275745663F, 0.8279025732F, 0.8282302344F, + 0.8285575501F, 0.8288845199F, 0.8292111437F, 0.8295374215F, + 0.8298633530F, 0.8301889382F, 0.8305141768F, 0.8308390688F, + 0.8311636141F, 0.8314878124F, 0.8318116637F, 0.8321351678F, + 0.8324583246F, 0.8327811340F, 0.8331035957F, 0.8334257098F, + 0.8337474761F, 0.8340688944F, 0.8343899647F, 0.8347106867F, + 0.8350310605F, 0.8353510857F, 0.8356707624F, 0.8359900904F, + 0.8363090696F, 0.8366276999F, 0.8369459811F, 0.8372639131F, + 0.8375814958F, 0.8378987292F, 0.8382156130F, 0.8385321472F, + 0.8388483316F, 0.8391641662F, 0.8394796508F, 0.8397947853F, + 0.8401095697F, 0.8404240037F, 0.8407380873F, 0.8410518204F, + 0.8413652029F, 0.8416782347F, 0.8419909156F, 0.8423032456F, + 0.8426152245F, 0.8429268523F, 0.8432381289F, 0.8435490541F, + 0.8438596279F, 0.8441698502F, 0.8444797208F, 0.8447892396F, + 0.8450984067F, 0.8454072218F, 0.8457156849F, 0.8460237959F, + 0.8463315547F, 0.8466389612F, 0.8469460154F, 0.8472527170F, + 0.8475590661F, 0.8478650625F, 0.8481707063F, 0.8484759971F, + 0.8487809351F, 0.8490855201F, 0.8493897521F, 0.8496936308F, + 0.8499971564F, 0.8503003286F, 0.8506031474F, 0.8509056128F, + 0.8512077246F, 0.8515094828F, 0.8518108872F, 0.8521119379F, + 0.8524126348F, 0.8527129777F, 0.8530129666F, 0.8533126015F, + 0.8536118822F, 0.8539108087F, 0.8542093809F, 0.8545075988F, + 0.8548054623F, 0.8551029712F, 0.8554001257F, 0.8556969255F, + 0.8559933707F, 0.8562894611F, 0.8565851968F, 0.8568805775F, + 0.8571756034F, 0.8574702743F, 0.8577645902F, 0.8580585509F, + 0.8583521566F, 0.8586454070F, 0.8589383021F, 0.8592308420F, + 0.8595230265F, 0.8598148556F, 0.8601063292F, 0.8603974473F, + 0.8606882098F, 0.8609786167F, 0.8612686680F, 0.8615583636F, + 0.8618477034F, 0.8621366874F, 0.8624253156F, 0.8627135878F, + 0.8630015042F, 0.8632890646F, 0.8635762690F, 0.8638631173F, + 0.8641496096F, 0.8644357457F, 0.8647215257F, 0.8650069495F, + 0.8652920171F, 0.8655767283F, 0.8658610833F, 0.8661450820F, + 0.8664287243F, 0.8667120102F, 0.8669949397F, 0.8672775127F, + 0.8675597293F, 0.8678415894F, 0.8681230929F, 0.8684042398F, + 0.8686850302F, 0.8689654640F, 0.8692455412F, 0.8695252617F, + 0.8698046255F, 0.8700836327F, 0.8703622831F, 0.8706405768F, + 0.8709185138F, 0.8711960940F, 0.8714733174F, 0.8717501840F, + 0.8720266939F, 0.8723028469F, 0.8725786430F, 0.8728540824F, + 0.8731291648F, 0.8734038905F, 0.8736782592F, 0.8739522711F, + 0.8742259261F, 0.8744992242F, 0.8747721653F, 0.8750447496F, + 0.8753169770F, 0.8755888475F, 0.8758603611F, 0.8761315177F, + 0.8764023175F, 0.8766727603F, 0.8769428462F, 0.8772125752F, + 0.8774819474F, 0.8777509626F, 0.8780196209F, 0.8782879224F, + 0.8785558669F, 0.8788234546F, 0.8790906854F, 0.8793575594F, + 0.8796240765F, 0.8798902368F, 0.8801560403F, 0.8804214870F, + 0.8806865768F, 0.8809513099F, 0.8812156863F, 0.8814797059F, + 0.8817433687F, 0.8820066749F, 0.8822696243F, 0.8825322171F, + 0.8827944532F, 0.8830563327F, 0.8833178556F, 0.8835790219F, + 0.8838398316F, 0.8841002848F, 0.8843603815F, 0.8846201217F, + 0.8848795054F, 0.8851385327F, 0.8853972036F, 0.8856555182F, + 0.8859134764F, 0.8861710783F, 0.8864283239F, 0.8866852133F, + 0.8869417464F, 0.8871979234F, 0.8874537443F, 0.8877092090F, + 0.8879643177F, 0.8882190704F, 0.8884734671F, 0.8887275078F, + 0.8889811927F, 0.8892345216F, 0.8894874948F, 0.8897401122F, + 0.8899923738F, 0.8902442798F, 0.8904958301F, 0.8907470248F, + 0.8909978640F, 0.8912483477F, 0.8914984759F, 0.8917482487F, + 0.8919976662F, 0.8922467284F, 0.8924954353F, 0.8927437871F, + 0.8929917837F, 0.8932394252F, 0.8934867118F, 0.8937336433F, + 0.8939802199F, 0.8942264417F, 0.8944723087F, 0.8947178210F, + 0.8949629785F, 0.8952077815F, 0.8954522299F, 0.8956963239F, + 0.8959400634F, 0.8961834486F, 0.8964264795F, 0.8966691561F, + 0.8969114786F, 0.8971534470F, 0.8973950614F, 0.8976363219F, + 0.8978772284F, 0.8981177812F, 0.8983579802F, 0.8985978256F, + 0.8988373174F, 0.8990764556F, 0.8993152405F, 0.8995536720F, + 0.8997917502F, 0.9000294751F, 0.9002668470F, 0.9005038658F, + 0.9007405317F, 0.9009768446F, 0.9012128048F, 0.9014484123F, + 0.9016836671F, 0.9019185693F, 0.9021531191F, 0.9023873165F, + 0.9026211616F, 0.9028546546F, 0.9030877954F, 0.9033205841F, + 0.9035530210F, 0.9037851059F, 0.9040168392F, 0.9042482207F, + 0.9044792507F, 0.9047099293F, 0.9049402564F, 0.9051702323F, + 0.9053998569F, 0.9056291305F, 0.9058580531F, 0.9060866248F, + 0.9063148457F, 0.9065427159F, 0.9067702355F, 0.9069974046F, + 0.9072242233F, 0.9074506917F, 0.9076768100F, 0.9079025782F, + 0.9081279964F, 0.9083530647F, 0.9085777833F, 0.9088021523F, + 0.9090261717F, 0.9092498417F, 0.9094731623F, 0.9096961338F, + 0.9099187561F, 0.9101410295F, 0.9103629540F, 0.9105845297F, + 0.9108057568F, 0.9110266354F, 0.9112471656F, 0.9114673475F, + 0.9116871812F, 0.9119066668F, 0.9121258046F, 0.9123445945F, + 0.9125630367F, 0.9127811314F, 0.9129988786F, 0.9132162785F, + 0.9134333312F, 0.9136500368F, 0.9138663954F, 0.9140824073F, + 0.9142980724F, 0.9145133910F, 0.9147283632F, 0.9149429890F, + 0.9151572687F, 0.9153712023F, 0.9155847900F, 0.9157980319F, + 0.9160109282F, 0.9162234790F, 0.9164356844F, 0.9166475445F, + 0.9168590595F, 0.9170702296F, 0.9172810548F, 0.9174915354F, + 0.9177016714F, 0.9179114629F, 0.9181209102F, 0.9183300134F, + 0.9185387726F, 0.9187471879F, 0.9189552595F, 0.9191629876F, + 0.9193703723F, 0.9195774136F, 0.9197841119F, 0.9199904672F, + 0.9201964797F, 0.9204021495F, 0.9206074767F, 0.9208124616F, + 0.9210171043F, 0.9212214049F, 0.9214253636F, 0.9216289805F, + 0.9218322558F, 0.9220351896F, 0.9222377821F, 0.9224400335F, + 0.9226419439F, 0.9228435134F, 0.9230447423F, 0.9232456307F, + 0.9234461787F, 0.9236463865F, 0.9238462543F, 0.9240457822F, + 0.9242449704F, 0.9244438190F, 0.9246423282F, 0.9248404983F, + 0.9250383293F, 0.9252358214F, 0.9254329747F, 0.9256297896F, + 0.9258262660F, 0.9260224042F, 0.9262182044F, 0.9264136667F, + 0.9266087913F, 0.9268035783F, 0.9269980280F, 0.9271921405F, + 0.9273859160F, 0.9275793546F, 0.9277724566F, 0.9279652221F, + 0.9281576513F, 0.9283497443F, 0.9285415014F, 0.9287329227F, + 0.9289240084F, 0.9291147586F, 0.9293051737F, 0.9294952536F, + 0.9296849987F, 0.9298744091F, 0.9300634850F, 0.9302522266F, + 0.9304406340F, 0.9306287074F, 0.9308164471F, 0.9310038532F, + 0.9311909259F, 0.9313776654F, 0.9315640719F, 0.9317501455F, + 0.9319358865F, 0.9321212951F, 0.9323063713F, 0.9324911155F, + 0.9326755279F, 0.9328596085F, 0.9330433577F, 0.9332267756F, + 0.9334098623F, 0.9335926182F, 0.9337750434F, 0.9339571380F, + 0.9341389023F, 0.9343203366F, 0.9345014409F, 0.9346822155F, + 0.9348626606F, 0.9350427763F, 0.9352225630F, 0.9354020207F, + 0.9355811498F, 0.9357599503F, 0.9359384226F, 0.9361165667F, + 0.9362943830F, 0.9364718716F, 0.9366490327F, 0.9368258666F, + 0.9370023733F, 0.9371785533F, 0.9373544066F, 0.9375299335F, + 0.9377051341F, 0.9378800087F, 0.9380545576F, 0.9382287809F, + 0.9384026787F, 0.9385762515F, 0.9387494993F, 0.9389224223F, + 0.9390950209F, 0.9392672951F, 0.9394392453F, 0.9396108716F, + 0.9397821743F, 0.9399531536F, 0.9401238096F, 0.9402941427F, + 0.9404641530F, 0.9406338407F, 0.9408032061F, 0.9409722495F, + 0.9411409709F, 0.9413093707F, 0.9414774491F, 0.9416452062F, + 0.9418126424F, 0.9419797579F, 0.9421465528F, 0.9423130274F, + 0.9424791819F, 0.9426450166F, 0.9428105317F, 0.9429757274F, + 0.9431406039F, 0.9433051616F, 0.9434694005F, 0.9436333209F, + 0.9437969232F, 0.9439602074F, 0.9441231739F, 0.9442858229F, + 0.9444481545F, 0.9446101691F, 0.9447718669F, 0.9449332481F, + 0.9450943129F, 0.9452550617F, 0.9454154945F, 0.9455756118F, + 0.9457354136F, 0.9458949003F, 0.9460540721F, 0.9462129292F, + 0.9463714719F, 0.9465297003F, 0.9466876149F, 0.9468452157F, + 0.9470025031F, 0.9471594772F, 0.9473161384F, 0.9474724869F, + 0.9476285229F, 0.9477842466F, 0.9479396584F, 0.9480947585F, + 0.9482495470F, 0.9484040243F, 0.9485581906F, 0.9487120462F, + 0.9488655913F, 0.9490188262F, 0.9491717511F, 0.9493243662F, + 0.9494766718F, 0.9496286683F, 0.9497803557F, 0.9499317345F, + 0.9500828047F, 0.9502335668F, 0.9503840209F, 0.9505341673F, + 0.9506840062F, 0.9508335380F, 0.9509827629F, 0.9511316810F, + 0.9512802928F, 0.9514285984F, 0.9515765982F, 0.9517242923F, + 0.9518716810F, 0.9520187646F, 0.9521655434F, 0.9523120176F, + 0.9524581875F, 0.9526040534F, 0.9527496154F, 0.9528948739F, + 0.9530398292F, 0.9531844814F, 0.9533288310F, 0.9534728780F, + 0.9536166229F, 0.9537600659F, 0.9539032071F, 0.9540460470F, + 0.9541885858F, 0.9543308237F, 0.9544727611F, 0.9546143981F, + 0.9547557351F, 0.9548967723F, 0.9550375100F, 0.9551779485F, + 0.9553180881F, 0.9554579290F, 0.9555974714F, 0.9557367158F, + 0.9558756623F, 0.9560143112F, 0.9561526628F, 0.9562907174F, + 0.9564284752F, 0.9565659366F, 0.9567031017F, 0.9568399710F, + 0.9569765446F, 0.9571128229F, 0.9572488061F, 0.9573844944F, + 0.9575198883F, 0.9576549879F, 0.9577897936F, 0.9579243056F, + 0.9580585242F, 0.9581924497F, 0.9583260824F, 0.9584594226F, + 0.9585924705F, 0.9587252264F, 0.9588576906F, 0.9589898634F, + 0.9591217452F, 0.9592533360F, 0.9593846364F, 0.9595156465F, + 0.9596463666F, 0.9597767971F, 0.9599069382F, 0.9600367901F, + 0.9601663533F, 0.9602956279F, 0.9604246143F, 0.9605533128F, + 0.9606817236F, 0.9608098471F, 0.9609376835F, 0.9610652332F, + 0.9611924963F, 0.9613194733F, 0.9614461644F, 0.9615725699F, + 0.9616986901F, 0.9618245253F, 0.9619500757F, 0.9620753418F, + 0.9622003238F, 0.9623250219F, 0.9624494365F, 0.9625735679F, + 0.9626974163F, 0.9628209821F, 0.9629442656F, 0.9630672671F, + 0.9631899868F, 0.9633124251F, 0.9634345822F, 0.9635564585F, + 0.9636780543F, 0.9637993699F, 0.9639204056F, 0.9640411616F, + 0.9641616383F, 0.9642818359F, 0.9644017549F, 0.9645213955F, + 0.9646407579F, 0.9647598426F, 0.9648786497F, 0.9649971797F, + 0.9651154328F, 0.9652334092F, 0.9653511095F, 0.9654685337F, + 0.9655856823F, 0.9657025556F, 0.9658191538F, 0.9659354773F, + 0.9660515263F, 0.9661673013F, 0.9662828024F, 0.9663980300F, + 0.9665129845F, 0.9666276660F, 0.9667420750F, 0.9668562118F, + 0.9669700766F, 0.9670836698F, 0.9671969917F, 0.9673100425F, + 0.9674228227F, 0.9675353325F, 0.9676475722F, 0.9677595422F, + 0.9678712428F, 0.9679826742F, 0.9680938368F, 0.9682047309F, + 0.9683153569F, 0.9684257150F, 0.9685358056F, 0.9686456289F, + 0.9687551853F, 0.9688644752F, 0.9689734987F, 0.9690822564F, + 0.9691907483F, 0.9692989750F, 0.9694069367F, 0.9695146337F, + 0.9696220663F, 0.9697292349F, 0.9698361398F, 0.9699427813F, + 0.9700491597F, 0.9701552754F, 0.9702611286F, 0.9703667197F, + 0.9704720490F, 0.9705771169F, 0.9706819236F, 0.9707864695F, + 0.9708907549F, 0.9709947802F, 0.9710985456F, 0.9712020514F, + 0.9713052981F, 0.9714082859F, 0.9715110151F, 0.9716134862F, + 0.9717156993F, 0.9718176549F, 0.9719193532F, 0.9720207946F, + 0.9721219794F, 0.9722229080F, 0.9723235806F, 0.9724239976F, + 0.9725241593F, 0.9726240661F, 0.9727237183F, 0.9728231161F, + 0.9729222601F, 0.9730211503F, 0.9731197873F, 0.9732181713F, + 0.9733163027F, 0.9734141817F, 0.9735118088F, 0.9736091842F, + 0.9737063083F, 0.9738031814F, 0.9738998039F, 0.9739961760F, + 0.9740922981F, 0.9741881706F, 0.9742837938F, 0.9743791680F, + 0.9744742935F, 0.9745691707F, 0.9746637999F, 0.9747581814F, + 0.9748523157F, 0.9749462029F, 0.9750398435F, 0.9751332378F, + 0.9752263861F, 0.9753192887F, 0.9754119461F, 0.9755043585F, + 0.9755965262F, 0.9756884496F, 0.9757801291F, 0.9758715650F, + 0.9759627575F, 0.9760537071F, 0.9761444141F, 0.9762348789F, + 0.9763251016F, 0.9764150828F, 0.9765048228F, 0.9765943218F, + 0.9766835802F, 0.9767725984F, 0.9768613767F, 0.9769499154F, + 0.9770382149F, 0.9771262755F, 0.9772140976F, 0.9773016815F, + 0.9773890275F, 0.9774761360F, 0.9775630073F, 0.9776496418F, + 0.9777360398F, 0.9778222016F, 0.9779081277F, 0.9779938182F, + 0.9780792736F, 0.9781644943F, 0.9782494805F, 0.9783342326F, + 0.9784187509F, 0.9785030359F, 0.9785870877F, 0.9786709069F, + 0.9787544936F, 0.9788378484F, 0.9789209714F, 0.9790038631F, + 0.9790865238F, 0.9791689538F, 0.9792511535F, 0.9793331232F, + 0.9794148633F, 0.9794963742F, 0.9795776561F, 0.9796587094F, + 0.9797395345F, 0.9798201316F, 0.9799005013F, 0.9799806437F, + 0.9800605593F, 0.9801402483F, 0.9802197112F, 0.9802989483F, + 0.9803779600F, 0.9804567465F, 0.9805353082F, 0.9806136455F, + 0.9806917587F, 0.9807696482F, 0.9808473143F, 0.9809247574F, + 0.9810019778F, 0.9810789759F, 0.9811557519F, 0.9812323064F, + 0.9813086395F, 0.9813847517F, 0.9814606433F, 0.9815363147F, + 0.9816117662F, 0.9816869981F, 0.9817620108F, 0.9818368047F, + 0.9819113801F, 0.9819857374F, 0.9820598769F, 0.9821337989F, + 0.9822075038F, 0.9822809920F, 0.9823542638F, 0.9824273195F, + 0.9825001596F, 0.9825727843F, 0.9826451940F, 0.9827173891F, + 0.9827893700F, 0.9828611368F, 0.9829326901F, 0.9830040302F, + 0.9830751574F, 0.9831460720F, 0.9832167745F, 0.9832872652F, + 0.9833575444F, 0.9834276124F, 0.9834974697F, 0.9835671166F, + 0.9836365535F, 0.9837057806F, 0.9837747983F, 0.9838436071F, + 0.9839122072F, 0.9839805990F, 0.9840487829F, 0.9841167591F, + 0.9841845282F, 0.9842520903F, 0.9843194459F, 0.9843865953F, + 0.9844535389F, 0.9845202771F, 0.9845868101F, 0.9846531383F, + 0.9847192622F, 0.9847851820F, 0.9848508980F, 0.9849164108F, + 0.9849817205F, 0.9850468276F, 0.9851117324F, 0.9851764352F, + 0.9852409365F, 0.9853052366F, 0.9853693358F, 0.9854332344F, + 0.9854969330F, 0.9855604317F, 0.9856237309F, 0.9856868310F, + 0.9857497325F, 0.9858124355F, 0.9858749404F, 0.9859372477F, + 0.9859993577F, 0.9860612707F, 0.9861229871F, 0.9861845072F, + 0.9862458315F, 0.9863069601F, 0.9863678936F, 0.9864286322F, + 0.9864891764F, 0.9865495264F, 0.9866096826F, 0.9866696454F, + 0.9867294152F, 0.9867889922F, 0.9868483769F, 0.9869075695F, + 0.9869665706F, 0.9870253803F, 0.9870839991F, 0.9871424273F, + 0.9872006653F, 0.9872587135F, 0.9873165721F, 0.9873742415F, + 0.9874317222F, 0.9874890144F, 0.9875461185F, 0.9876030348F, + 0.9876597638F, 0.9877163057F, 0.9877726610F, 0.9878288300F, + 0.9878848130F, 0.9879406104F, 0.9879962225F, 0.9880516497F, + 0.9881068924F, 0.9881619509F, 0.9882168256F, 0.9882715168F, + 0.9883260249F, 0.9883803502F, 0.9884344931F, 0.9884884539F, + 0.9885422331F, 0.9885958309F, 0.9886492477F, 0.9887024838F, + 0.9887555397F, 0.9888084157F, 0.9888611120F, 0.9889136292F, + 0.9889659675F, 0.9890181273F, 0.9890701089F, 0.9891219128F, + 0.9891735392F, 0.9892249885F, 0.9892762610F, 0.9893273572F, + 0.9893782774F, 0.9894290219F, 0.9894795911F, 0.9895299853F, + 0.9895802049F, 0.9896302502F, 0.9896801217F, 0.9897298196F, + 0.9897793443F, 0.9898286961F, 0.9898778755F, 0.9899268828F, + 0.9899757183F, 0.9900243823F, 0.9900728753F, 0.9901211976F, + 0.9901693495F, 0.9902173314F, 0.9902651436F, 0.9903127865F, + 0.9903602605F, 0.9904075659F, 0.9904547031F, 0.9905016723F, + 0.9905484740F, 0.9905951086F, 0.9906415763F, 0.9906878775F, + 0.9907340126F, 0.9907799819F, 0.9908257858F, 0.9908714247F, + 0.9909168988F, 0.9909622086F, 0.9910073543F, 0.9910523364F, + 0.9910971552F, 0.9911418110F, 0.9911863042F, 0.9912306351F, + 0.9912748042F, 0.9913188117F, 0.9913626580F, 0.9914063435F, + 0.9914498684F, 0.9914932333F, 0.9915364383F, 0.9915794839F, + 0.9916223703F, 0.9916650981F, 0.9917076674F, 0.9917500787F, + 0.9917923323F, 0.9918344286F, 0.9918763679F, 0.9919181505F, + 0.9919597769F, 0.9920012473F, 0.9920425621F, 0.9920837217F, + 0.9921247263F, 0.9921655765F, 0.9922062724F, 0.9922468145F, + 0.9922872030F, 0.9923274385F, 0.9923675211F, 0.9924074513F, + 0.9924472294F, 0.9924868557F, 0.9925263306F, 0.9925656544F, + 0.9926048275F, 0.9926438503F, 0.9926827230F, 0.9927214461F, + 0.9927600199F, 0.9927984446F, 0.9928367208F, 0.9928748486F, + 0.9929128285F, 0.9929506608F, 0.9929883459F, 0.9930258841F, + 0.9930632757F, 0.9931005211F, 0.9931376207F, 0.9931745747F, + 0.9932113836F, 0.9932480476F, 0.9932845671F, 0.9933209425F, + 0.9933571742F, 0.9933932623F, 0.9934292074F, 0.9934650097F, + 0.9935006696F, 0.9935361874F, 0.9935715635F, 0.9936067982F, + 0.9936418919F, 0.9936768448F, 0.9937116574F, 0.9937463300F, + 0.9937808629F, 0.9938152565F, 0.9938495111F, 0.9938836271F, + 0.9939176047F, 0.9939514444F, 0.9939851465F, 0.9940187112F, + 0.9940521391F, 0.9940854303F, 0.9941185853F, 0.9941516044F, + 0.9941844879F, 0.9942172361F, 0.9942498495F, 0.9942823283F, + 0.9943146729F, 0.9943468836F, 0.9943789608F, 0.9944109047F, + 0.9944427158F, 0.9944743944F, 0.9945059408F, 0.9945373553F, + 0.9945686384F, 0.9945997902F, 0.9946308112F, 0.9946617017F, + 0.9946924621F, 0.9947230926F, 0.9947535937F, 0.9947839656F, + 0.9948142086F, 0.9948443232F, 0.9948743097F, 0.9949041683F, + 0.9949338995F, 0.9949635035F, 0.9949929807F, 0.9950223315F, + 0.9950515561F, 0.9950806549F, 0.9951096282F, 0.9951384764F, + 0.9951671998F, 0.9951957987F, 0.9952242735F, 0.9952526245F, + 0.9952808520F, 0.9953089564F, 0.9953369380F, 0.9953647971F, + 0.9953925340F, 0.9954201491F, 0.9954476428F, 0.9954750153F, + 0.9955022670F, 0.9955293981F, 0.9955564092F, 0.9955833003F, + 0.9956100720F, 0.9956367245F, 0.9956632582F, 0.9956896733F, + 0.9957159703F, 0.9957421494F, 0.9957682110F, 0.9957941553F, + 0.9958199828F, 0.9958456937F, 0.9958712884F, 0.9958967672F, + 0.9959221305F, 0.9959473784F, 0.9959725115F, 0.9959975300F, + 0.9960224342F, 0.9960472244F, 0.9960719011F, 0.9960964644F, + 0.9961209148F, 0.9961452525F, 0.9961694779F, 0.9961935913F, + 0.9962175930F, 0.9962414834F, 0.9962652627F, 0.9962889313F, + 0.9963124895F, 0.9963359377F, 0.9963592761F, 0.9963825051F, + 0.9964056250F, 0.9964286361F, 0.9964515387F, 0.9964743332F, + 0.9964970198F, 0.9965195990F, 0.9965420709F, 0.9965644360F, + 0.9965866946F, 0.9966088469F, 0.9966308932F, 0.9966528340F, + 0.9966746695F, 0.9966964001F, 0.9967180260F, 0.9967395475F, + 0.9967609651F, 0.9967822789F, 0.9968034894F, 0.9968245968F, + 0.9968456014F, 0.9968665036F, 0.9968873037F, 0.9969080019F, + 0.9969285987F, 0.9969490942F, 0.9969694889F, 0.9969897830F, + 0.9970099769F, 0.9970300708F, 0.9970500651F, 0.9970699601F, + 0.9970897561F, 0.9971094533F, 0.9971290522F, 0.9971485531F, + 0.9971679561F, 0.9971872617F, 0.9972064702F, 0.9972255818F, + 0.9972445968F, 0.9972635157F, 0.9972823386F, 0.9973010659F, + 0.9973196980F, 0.9973382350F, 0.9973566773F, 0.9973750253F, + 0.9973932791F, 0.9974114392F, 0.9974295059F, 0.9974474793F, + 0.9974653599F, 0.9974831480F, 0.9975008438F, 0.9975184476F, + 0.9975359598F, 0.9975533806F, 0.9975707104F, 0.9975879495F, + 0.9976050981F, 0.9976221566F, 0.9976391252F, 0.9976560043F, + 0.9976727941F, 0.9976894950F, 0.9977061073F, 0.9977226312F, + 0.9977390671F, 0.9977554152F, 0.9977716759F, 0.9977878495F, + 0.9978039361F, 0.9978199363F, 0.9978358501F, 0.9978516780F, + 0.9978674202F, 0.9978830771F, 0.9978986488F, 0.9979141358F, + 0.9979295383F, 0.9979448566F, 0.9979600909F, 0.9979752417F, + 0.9979903091F, 0.9980052936F, 0.9980201952F, 0.9980350145F, + 0.9980497515F, 0.9980644067F, 0.9980789804F, 0.9980934727F, + 0.9981078841F, 0.9981222147F, 0.9981364649F, 0.9981506350F, + 0.9981647253F, 0.9981787360F, 0.9981926674F, 0.9982065199F, + 0.9982202936F, 0.9982339890F, 0.9982476062F, 0.9982611456F, + 0.9982746074F, 0.9982879920F, 0.9983012996F, 0.9983145304F, + 0.9983276849F, 0.9983407632F, 0.9983537657F, 0.9983666926F, + 0.9983795442F, 0.9983923208F, 0.9984050226F, 0.9984176501F, + 0.9984302033F, 0.9984426827F, 0.9984550884F, 0.9984674208F, + 0.9984796802F, 0.9984918667F, 0.9985039808F, 0.9985160227F, + 0.9985279926F, 0.9985398909F, 0.9985517177F, 0.9985634734F, + 0.9985751583F, 0.9985867727F, 0.9985983167F, 0.9986097907F, + 0.9986211949F, 0.9986325297F, 0.9986437953F, 0.9986549919F, + 0.9986661199F, 0.9986771795F, 0.9986881710F, 0.9986990946F, + 0.9987099507F, 0.9987207394F, 0.9987314611F, 0.9987421161F, + 0.9987527045F, 0.9987632267F, 0.9987736829F, 0.9987840734F, + 0.9987943985F, 0.9988046584F, 0.9988148534F, 0.9988249838F, + 0.9988350498F, 0.9988450516F, 0.9988549897F, 0.9988648641F, + 0.9988746753F, 0.9988844233F, 0.9988941086F, 0.9989037313F, + 0.9989132918F, 0.9989227902F, 0.9989322269F, 0.9989416021F, + 0.9989509160F, 0.9989601690F, 0.9989693613F, 0.9989784931F, + 0.9989875647F, 0.9989965763F, 0.9990055283F, 0.9990144208F, + 0.9990232541F, 0.9990320286F, 0.9990407443F, 0.9990494016F, + 0.9990580008F, 0.9990665421F, 0.9990750257F, 0.9990834519F, + 0.9990918209F, 0.9991001331F, 0.9991083886F, 0.9991165877F, + 0.9991247307F, 0.9991328177F, 0.9991408491F, 0.9991488251F, + 0.9991567460F, 0.9991646119F, 0.9991724232F, 0.9991801801F, + 0.9991878828F, 0.9991955316F, 0.9992031267F, 0.9992106684F, + 0.9992181569F, 0.9992255925F, 0.9992329753F, 0.9992403057F, + 0.9992475839F, 0.9992548101F, 0.9992619846F, 0.9992691076F, + 0.9992761793F, 0.9992832001F, 0.9992901701F, 0.9992970895F, + 0.9993039587F, 0.9993107777F, 0.9993175470F, 0.9993242667F, + 0.9993309371F, 0.9993375583F, 0.9993441307F, 0.9993506545F, + 0.9993571298F, 0.9993635570F, 0.9993699362F, 0.9993762678F, + 0.9993825519F, 0.9993887887F, 0.9993949785F, 0.9994011216F, + 0.9994072181F, 0.9994132683F, 0.9994192725F, 0.9994252307F, + 0.9994311434F, 0.9994370107F, 0.9994428327F, 0.9994486099F, + 0.9994543423F, 0.9994600303F, 0.9994656739F, 0.9994712736F, + 0.9994768294F, 0.9994823417F, 0.9994878105F, 0.9994932363F, + 0.9994986191F, 0.9995039592F, 0.9995092568F, 0.9995145122F, + 0.9995197256F, 0.9995248971F, 0.9995300270F, 0.9995351156F, + 0.9995401630F, 0.9995451695F, 0.9995501352F, 0.9995550604F, + 0.9995599454F, 0.9995647903F, 0.9995695953F, 0.9995743607F, + 0.9995790866F, 0.9995837734F, 0.9995884211F, 0.9995930300F, + 0.9995976004F, 0.9996021324F, 0.9996066263F, 0.9996110822F, + 0.9996155004F, 0.9996198810F, 0.9996242244F, 0.9996285306F, + 0.9996327999F, 0.9996370326F, 0.9996412287F, 0.9996453886F, + 0.9996495125F, 0.9996536004F, 0.9996576527F, 0.9996616696F, + 0.9996656512F, 0.9996695977F, 0.9996735094F, 0.9996773865F, + 0.9996812291F, 0.9996850374F, 0.9996888118F, 0.9996925523F, + 0.9996962591F, 0.9996999325F, 0.9997035727F, 0.9997071798F, + 0.9997107541F, 0.9997142957F, 0.9997178049F, 0.9997212818F, + 0.9997247266F, 0.9997281396F, 0.9997315209F, 0.9997348708F, + 0.9997381893F, 0.9997414767F, 0.9997447333F, 0.9997479591F, + 0.9997511544F, 0.9997543194F, 0.9997574542F, 0.9997605591F, + 0.9997636342F, 0.9997666797F, 0.9997696958F, 0.9997726828F, + 0.9997756407F, 0.9997785698F, 0.9997814703F, 0.9997843423F, + 0.9997871860F, 0.9997900016F, 0.9997927894F, 0.9997955494F, + 0.9997982818F, 0.9998009869F, 0.9998036648F, 0.9998063157F, + 0.9998089398F, 0.9998115373F, 0.9998141082F, 0.9998166529F, + 0.9998191715F, 0.9998216642F, 0.9998241311F, 0.9998265724F, + 0.9998289884F, 0.9998313790F, 0.9998337447F, 0.9998360854F, + 0.9998384015F, 0.9998406930F, 0.9998429602F, 0.9998452031F, + 0.9998474221F, 0.9998496171F, 0.9998517885F, 0.9998539364F, + 0.9998560610F, 0.9998581624F, 0.9998602407F, 0.9998622962F, + 0.9998643291F, 0.9998663394F, 0.9998683274F, 0.9998702932F, + 0.9998722370F, 0.9998741589F, 0.9998760591F, 0.9998779378F, + 0.9998797952F, 0.9998816313F, 0.9998834464F, 0.9998852406F, + 0.9998870141F, 0.9998887670F, 0.9998904995F, 0.9998922117F, + 0.9998939039F, 0.9998955761F, 0.9998972285F, 0.9998988613F, + 0.9999004746F, 0.9999020686F, 0.9999036434F, 0.9999051992F, + 0.9999067362F, 0.9999082544F, 0.9999097541F, 0.9999112354F, + 0.9999126984F, 0.9999141433F, 0.9999155703F, 0.9999169794F, + 0.9999183709F, 0.9999197449F, 0.9999211014F, 0.9999224408F, + 0.9999237631F, 0.9999250684F, 0.9999263570F, 0.9999276289F, + 0.9999288843F, 0.9999301233F, 0.9999313461F, 0.9999325529F, + 0.9999337437F, 0.9999349187F, 0.9999360780F, 0.9999372218F, + 0.9999383503F, 0.9999394635F, 0.9999405616F, 0.9999416447F, + 0.9999427129F, 0.9999437665F, 0.9999448055F, 0.9999458301F, + 0.9999468404F, 0.9999478365F, 0.9999488185F, 0.9999497867F, + 0.9999507411F, 0.9999516819F, 0.9999526091F, 0.9999535230F, + 0.9999544236F, 0.9999553111F, 0.9999561856F, 0.9999570472F, + 0.9999578960F, 0.9999587323F, 0.9999595560F, 0.9999603674F, + 0.9999611666F, 0.9999619536F, 0.9999627286F, 0.9999634917F, + 0.9999642431F, 0.9999649828F, 0.9999657110F, 0.9999664278F, + 0.9999671334F, 0.9999678278F, 0.9999685111F, 0.9999691835F, + 0.9999698451F, 0.9999704960F, 0.9999711364F, 0.9999717662F, + 0.9999723858F, 0.9999729950F, 0.9999735942F, 0.9999741834F, + 0.9999747626F, 0.9999753321F, 0.9999758919F, 0.9999764421F, + 0.9999769828F, 0.9999775143F, 0.9999780364F, 0.9999785495F, + 0.9999790535F, 0.9999795485F, 0.9999800348F, 0.9999805124F, + 0.9999809813F, 0.9999814417F, 0.9999818938F, 0.9999823375F, + 0.9999827731F, 0.9999832005F, 0.9999836200F, 0.9999840316F, + 0.9999844353F, 0.9999848314F, 0.9999852199F, 0.9999856008F, + 0.9999859744F, 0.9999863407F, 0.9999866997F, 0.9999870516F, + 0.9999873965F, 0.9999877345F, 0.9999880656F, 0.9999883900F, + 0.9999887078F, 0.9999890190F, 0.9999893237F, 0.9999896220F, + 0.9999899140F, 0.9999901999F, 0.9999904796F, 0.9999907533F, + 0.9999910211F, 0.9999912830F, 0.9999915391F, 0.9999917896F, + 0.9999920345F, 0.9999922738F, 0.9999925077F, 0.9999927363F, + 0.9999929596F, 0.9999931777F, 0.9999933907F, 0.9999935987F, + 0.9999938018F, 0.9999940000F, 0.9999941934F, 0.9999943820F, + 0.9999945661F, 0.9999947456F, 0.9999949206F, 0.9999950912F, + 0.9999952575F, 0.9999954195F, 0.9999955773F, 0.9999957311F, + 0.9999958807F, 0.9999960265F, 0.9999961683F, 0.9999963063F, + 0.9999964405F, 0.9999965710F, 0.9999966979F, 0.9999968213F, + 0.9999969412F, 0.9999970576F, 0.9999971707F, 0.9999972805F, + 0.9999973871F, 0.9999974905F, 0.9999975909F, 0.9999976881F, + 0.9999977824F, 0.9999978738F, 0.9999979624F, 0.9999980481F, + 0.9999981311F, 0.9999982115F, 0.9999982892F, 0.9999983644F, + 0.9999984370F, 0.9999985072F, 0.9999985750F, 0.9999986405F, + 0.9999987037F, 0.9999987647F, 0.9999988235F, 0.9999988802F, + 0.9999989348F, 0.9999989873F, 0.9999990379F, 0.9999990866F, + 0.9999991334F, 0.9999991784F, 0.9999992217F, 0.9999992632F, + 0.9999993030F, 0.9999993411F, 0.9999993777F, 0.9999994128F, + 0.9999994463F, 0.9999994784F, 0.9999995091F, 0.9999995384F, + 0.9999995663F, 0.9999995930F, 0.9999996184F, 0.9999996426F, + 0.9999996657F, 0.9999996876F, 0.9999997084F, 0.9999997282F, + 0.9999997469F, 0.9999997647F, 0.9999997815F, 0.9999997973F, + 0.9999998123F, 0.9999998265F, 0.9999998398F, 0.9999998524F, + 0.9999998642F, 0.9999998753F, 0.9999998857F, 0.9999998954F, + 0.9999999045F, 0.9999999130F, 0.9999999209F, 0.9999999282F, + 0.9999999351F, 0.9999999414F, 0.9999999472F, 0.9999999526F, + 0.9999999576F, 0.9999999622F, 0.9999999664F, 0.9999999702F, + 0.9999999737F, 0.9999999769F, 0.9999999798F, 0.9999999824F, + 0.9999999847F, 0.9999999868F, 0.9999999887F, 0.9999999904F, + 0.9999999919F, 0.9999999932F, 0.9999999943F, 0.9999999953F, + 0.9999999961F, 0.9999999969F, 0.9999999975F, 0.9999999980F, + 0.9999999985F, 0.9999999988F, 0.9999999991F, 0.9999999993F, + 0.9999999995F, 0.9999999997F, 0.9999999998F, 0.9999999999F, + 0.9999999999F, 1.0000000000F, 1.0000000000F, 1.0000000000F, + 1.0000000000F, 1.0000000000F, 1.0000000000F, 1.0000000000F, +}; + +const float ff_vorbis_floor1_inverse_db_table[256]={ + 1.0649863e-07F, 1.1341951e-07F, 1.2079015e-07F, 1.2863978e-07F, + 1.3699951e-07F, 1.4590251e-07F, 1.5538408e-07F, 1.6548181e-07F, + 1.7623575e-07F, 1.8768855e-07F, 1.9988561e-07F, 2.128753e-07F, + 2.2670913e-07F, 2.4144197e-07F, 2.5713223e-07F, 2.7384213e-07F, + 2.9163793e-07F, 3.1059021e-07F, 3.3077411e-07F, 3.5226968e-07F, + 3.7516214e-07F, 3.9954229e-07F, 4.2550680e-07F, 4.5315863e-07F, + 4.8260743e-07F, 5.1396998e-07F, 5.4737065e-07F, 5.8294187e-07F, + 6.2082472e-07F, 6.6116941e-07F, 7.0413592e-07F, 7.4989464e-07F, + 7.9862701e-07F, 8.5052630e-07F, 9.0579828e-07F, 9.6466216e-07F, + 1.0273513e-06F, 1.0941144e-06F, 1.1652161e-06F, 1.2409384e-06F, + 1.3215816e-06F, 1.4074654e-06F, 1.4989305e-06F, 1.5963394e-06F, + 1.7000785e-06F, 1.8105592e-06F, 1.9282195e-06F, 2.0535261e-06F, + 2.1869758e-06F, 2.3290978e-06F, 2.4804557e-06F, 2.6416497e-06F, + 2.8133190e-06F, 2.9961443e-06F, 3.1908506e-06F, 3.3982101e-06F, + 3.6190449e-06F, 3.8542308e-06F, 4.1047004e-06F, 4.3714470e-06F, + 4.6555282e-06F, 4.9580707e-06F, 5.2802740e-06F, 5.6234160e-06F, + 5.9888572e-06F, 6.3780469e-06F, 6.7925283e-06F, 7.2339451e-06F, + 7.7040476e-06F, 8.2047000e-06F, 8.7378876e-06F, 9.3057248e-06F, + 9.9104632e-06F, 1.0554501e-05F, 1.1240392e-05F, 1.1970856e-05F, + 1.2748789e-05F, 1.3577278e-05F, 1.4459606e-05F, 1.5399272e-05F, + 1.6400004e-05F, 1.7465768e-05F, 1.8600792e-05F, 1.9809576e-05F, + 2.1096914e-05F, 2.2467911e-05F, 2.3928002e-05F, 2.5482978e-05F, + 2.7139006e-05F, 2.8902651e-05F, 3.0780908e-05F, 3.2781225e-05F, + 3.4911534e-05F, 3.7180282e-05F, 3.9596466e-05F, 4.2169667e-05F, + 4.4910090e-05F, 4.7828601e-05F, 5.0936773e-05F, 5.4246931e-05F, + 5.7772202e-05F, 6.1526565e-05F, 6.5524908e-05F, 6.9783085e-05F, + 7.4317983e-05F, 7.9147585e-05F, 8.4291040e-05F, 8.9768747e-05F, + 9.5602426e-05F, 0.00010181521F, 0.00010843174F, 0.00011547824F, + 0.00012298267F, 0.00013097477F, 0.00013948625F, 0.00014855085F, + 0.00015820453F, 0.00016848555F, 0.00017943469F, 0.00019109536F, + 0.00020351382F, 0.00021673929F, 0.00023082423F, 0.00024582449F, + 0.00026179955F, 0.00027881276F, 0.00029693158F, 0.00031622787F, + 0.00033677814F, 0.00035866388F, 0.00038197188F, 0.00040679456F, + 0.00043323036F, 0.00046138411F, 0.00049136745F, 0.00052329927F, + 0.00055730621F, 0.00059352311F, 0.00063209358F, 0.00067317058F, + 0.00071691700F, 0.00076350630F, 0.00081312324F, 0.00086596457F, + 0.00092223983F, 0.00098217216F, 0.0010459992F, 0.0011139742F, + 0.0011863665F, 0.0012634633F, 0.0013455702F, 0.0014330129F, + 0.0015261382F, 0.0016253153F, 0.0017309374F, 0.0018434235F, + 0.0019632195F, 0.0020908006F, 0.0022266726F, 0.0023713743F, + 0.0025254795F, 0.0026895994F, 0.0028643847F, 0.0030505286F, + 0.0032487691F, 0.0034598925F, 0.0036847358F, 0.0039241906F, + 0.0041792066F, 0.0044507950F, 0.0047400328F, 0.0050480668F, + 0.0053761186F, 0.0057254891F, 0.0060975636F, 0.0064938176F, + 0.0069158225F, 0.0073652516F, 0.0078438871F, 0.0083536271F, + 0.0088964928F, 0.009474637F, 0.010090352F, 0.010746080F, + 0.011444421F, 0.012188144F, 0.012980198F, 0.013823725F, + 0.014722068F, 0.015678791F, 0.016697687F, 0.017782797F, + 0.018938423F, 0.020169149F, 0.021479854F, 0.022875735F, + 0.024362330F, 0.025945531F, 0.027631618F, 0.029427276F, + 0.031339626F, 0.033376252F, 0.035545228F, 0.037855157F, + 0.040315199F, 0.042935108F, 0.045725273F, 0.048696758F, + 0.051861348F, 0.055231591F, 0.058820850F, 0.062643361F, + 0.066714279F, 0.071049749F, 0.075666962F, 0.080584227F, + 0.085821044F, 0.091398179F, 0.097337747F, 0.10366330F, + 0.11039993F, 0.11757434F, 0.12521498F, 0.13335215F, + 0.14201813F, 0.15124727F, 0.16107617F, 0.17154380F, + 0.18269168F, 0.19456402F, 0.20720788F, 0.22067342F, + 0.23501402F, 0.25028656F, 0.26655159F, 0.28387361F, + 0.30232132F, 0.32196786F, 0.34289114F, 0.36517414F, + 0.38890521F, 0.41417847F, 0.44109412F, 0.46975890F, + 0.50028648F, 0.53279791F, 0.56742212F, 0.60429640F, + 0.64356699F, 0.68538959F, 0.72993007F, 0.77736504F, + 0.82788260F, 0.88168307F, 0.9389798F, 1.F, +}; + +const float * ff_vorbis_vwin[8] = { vwin64, vwin128, vwin256, vwin512, vwin1024, vwin2048, vwin4096, vwin8192 }; + diff --git a/src/libffmpeg/libavcodec/vp5.c b/src/libffmpeg/libavcodec/vp5.c new file mode 100644 index 000000000..ac953c7aa --- /dev/null +++ b/src/libffmpeg/libavcodec/vp5.c @@ -0,0 +1,290 @@ +/** + * @file vp5.c + * VP5 compatible video decoder + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include <stdlib.h> +#include <string.h> + +#include "avcodec.h" +#include "dsputil.h" +#include "bitstream.h" +#include "mpegvideo.h" + +#include "vp56.h" +#include "vp56data.h" +#include "vp5data.h" + + +static int vp5_parse_header(vp56_context_t *s, uint8_t *buf, int buf_size, + int *golden_frame) +{ + vp56_range_coder_t *c = &s->c; + int rows, cols; + + vp56_init_range_decoder(&s->c, buf, buf_size); + s->frames[VP56_FRAME_CURRENT].key_frame = !vp56_rac_get(c); + vp56_rac_get(c); + vp56_init_dequant(s, vp56_rac_gets(c, 6)); + if (s->frames[VP56_FRAME_CURRENT].key_frame) + { + vp56_rac_gets(c, 8); + if(vp56_rac_gets(c, 5) > 5) + return 0; + vp56_rac_gets(c, 2); + if (vp56_rac_get(c)) { + av_log(s->avctx, AV_LOG_ERROR, "interlacing not supported\n"); + return 0; + } + rows = vp56_rac_gets(c, 8); /* number of stored macroblock rows */ + cols = vp56_rac_gets(c, 8); /* number of stored macroblock cols */ + vp56_rac_gets(c, 8); /* number of displayed macroblock rows */ + vp56_rac_gets(c, 8); /* number of displayed macroblock cols */ + vp56_rac_gets(c, 2); + if (16*cols != s->avctx->coded_width || + 16*rows != s->avctx->coded_height) { + avcodec_set_dimensions(s->avctx, 16*cols, 16*rows); + return 2; + } + } + return 1; +} + +/* Gives very similar result than the vp6 version except in a few cases */ +static int vp5_adjust(int v, int t) +{ + int s2, s1 = v >> 31; + v ^= s1; + v -= s1; + v *= v < 2*t; + v -= t; + s2 = v >> 31; + v ^= s2; + v -= s2; + v = t - v; + v += s1; + v ^= s1; + return v; +} + +static void vp5_parse_vector_adjustment(vp56_context_t *s, vp56_mv_t *vect) +{ + vp56_range_coder_t *c = &s->c; + int comp, di; + + for (comp=0; comp<2; comp++) { + int delta = 0; + if (vp56_rac_get_prob(c, s->vector_model_dct[comp])) { + int sign = vp56_rac_get_prob(c, s->vector_model_sig[comp]); + di = vp56_rac_get_prob(c, s->vector_model_pdi[comp][0]); + di |= vp56_rac_get_prob(c, s->vector_model_pdi[comp][1]) << 1; + delta = vp56_rac_get_tree(c, vp56_pva_tree, + s->vector_model_pdv[comp]); + delta = di | (delta << 2); + delta = (delta ^ -sign) + sign; + } + if (!comp) + vect->x = delta; + else + vect->y = delta; + } +} + +static void vp5_parse_vector_models(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + int comp, node; + + for (comp=0; comp<2; comp++) { + if (vp56_rac_get_prob(c, vp5_vmc_pct[comp][0])) + s->vector_model_dct[comp] = vp56_rac_gets_nn(c, 7); + if (vp56_rac_get_prob(c, vp5_vmc_pct[comp][1])) + s->vector_model_sig[comp] = vp56_rac_gets_nn(c, 7); + if (vp56_rac_get_prob(c, vp5_vmc_pct[comp][2])) + s->vector_model_pdi[comp][0] = vp56_rac_gets_nn(c, 7); + if (vp56_rac_get_prob(c, vp5_vmc_pct[comp][3])) + s->vector_model_pdi[comp][1] = vp56_rac_gets_nn(c, 7); + } + + for (comp=0; comp<2; comp++) + for (node=0; node<7; node++) + if (vp56_rac_get_prob(c, vp5_vmc_pct[comp][4 + node])) + s->vector_model_pdv[comp][node] = vp56_rac_gets_nn(c, 7); +} + +static void vp5_parse_coeff_models(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + uint8_t def_prob[11]; + int node, cg, ctx; + int ct; /* code type */ + int pt; /* plane type (0 for Y, 1 for U or V) */ + + memset(def_prob, 0x80, sizeof(def_prob)); + + for (pt=0; pt<2; pt++) + for (node=0; node<11; node++) + if (vp56_rac_get_prob(c, vp5_dccv_pct[pt][node])) { + def_prob[node] = vp56_rac_gets_nn(c, 7); + s->coeff_model_dccv[pt][node] = def_prob[node]; + } else if (s->frames[VP56_FRAME_CURRENT].key_frame) { + s->coeff_model_dccv[pt][node] = def_prob[node]; + } + + for (ct=0; ct<3; ct++) + for (pt=0; pt<2; pt++) + for (cg=0; cg<6; cg++) + for (node=0; node<11; node++) + if (vp56_rac_get_prob(c, vp5_ract_pct[ct][pt][cg][node])) { + def_prob[node] = vp56_rac_gets_nn(c, 7); + s->coeff_model_ract[pt][ct][cg][node] = def_prob[node]; + } else if (s->frames[VP56_FRAME_CURRENT].key_frame) { + s->coeff_model_ract[pt][ct][cg][node] = def_prob[node]; + } + + /* coeff_model_dcct is a linear combination of coeff_model_dccv */ + for (pt=0; pt<2; pt++) + for (ctx=0; ctx<36; ctx++) + for (node=0; node<5; node++) + s->coeff_model_dcct[pt][ctx][node] = clip(((s->coeff_model_dccv[pt][node] * vp5_dccv_lc[node][ctx][0] + 128) >> 8) + vp5_dccv_lc[node][ctx][1], 1, 254); + + /* coeff_model_acct is a linear combination of coeff_model_ract */ + for (ct=0; ct<3; ct++) + for (pt=0; pt<2; pt++) + for (cg=0; cg<3; cg++) + for (ctx=0; ctx<6; ctx++) + for (node=0; node<5; node++) + s->coeff_model_acct[pt][ct][cg][ctx][node] = clip(((s->coeff_model_ract[pt][ct][cg][node] * vp5_ract_lc[ct][cg][node][ctx][0] + 128) >> 8) + vp5_ract_lc[ct][cg][node][ctx][1], 1, 254); +} + +static void vp5_parse_coeff(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + uint8_t *permute = s->scantable.permutated; + uint8_t *model, *model2; + int coeff, sign, coeff_idx; + int b, i, cg, idx, ctx, ctx_last; + int pt = 0; /* plane type (0 for Y, 1 for U or V) */ + + for (b=0; b<6; b++) { + int ct = 1; /* code type */ + + if (b > 3) pt = 1; + + ctx = 6*s->coeff_ctx[vp56_b6to4[b]][0] + + s->above_blocks[s->above_block_idx[b]].not_null_dc; + model = s->coeff_model_dccv[pt]; + model2 = s->coeff_model_dcct[pt][ctx]; + + for (coeff_idx=0; coeff_idx<64; ) { + if (vp56_rac_get_prob(c, model2[0])) { + if (vp56_rac_get_prob(c, model2[2])) { + if (vp56_rac_get_prob(c, model2[3])) { + s->coeff_ctx[vp56_b6to4[b]][coeff_idx] = 4; + idx = vp56_rac_get_tree(c, vp56_pc_tree, model); + sign = vp56_rac_get(c); + coeff = vp56_coeff_bias[idx]; + for (i=vp56_coeff_bit_length[idx]; i>=0; i--) + coeff += vp56_rac_get_prob(c, vp56_coeff_parse_table[idx][i]) << i; + } else { + if (vp56_rac_get_prob(c, model2[4])) { + coeff = 3 + vp56_rac_get_prob(c, model[5]); + s->coeff_ctx[vp56_b6to4[b]][coeff_idx] = 3; + } else { + coeff = 2; + s->coeff_ctx[vp56_b6to4[b]][coeff_idx] = 2; + } + sign = vp56_rac_get(c); + } + ct = 2; + } else { + ct = 1; + s->coeff_ctx[vp56_b6to4[b]][coeff_idx] = 1; + sign = vp56_rac_get(c); + coeff = 1; + } + coeff = (coeff ^ -sign) + sign; + if (coeff_idx) + coeff *= s->dequant_ac; + s->block_coeff[b][permute[coeff_idx]] = coeff; + } else { + if (ct && !vp56_rac_get_prob(c, model2[1])) + break; + ct = 0; + s->coeff_ctx[vp56_b6to4[b]][coeff_idx] = 0; + } + + cg = vp5_coeff_groups[++coeff_idx]; + ctx = s->coeff_ctx[vp56_b6to4[b]][coeff_idx]; + model = s->coeff_model_ract[pt][ct][cg]; + model2 = cg > 2 ? model : s->coeff_model_acct[pt][ct][cg][ctx]; + } + + ctx_last = FFMIN(s->coeff_ctx_last[vp56_b6to4[b]], 24); + s->coeff_ctx_last[vp56_b6to4[b]] = coeff_idx; + if (coeff_idx < ctx_last) + for (i=coeff_idx; i<=ctx_last; i++) + s->coeff_ctx[vp56_b6to4[b]][i] = 5; + s->above_blocks[s->above_block_idx[b]].not_null_dc = s->coeff_ctx[vp56_b6to4[b]][0]; + } +} + +static void vp5_default_models_init(vp56_context_t *s) +{ + int i; + + for (i=0; i<2; i++) { + s->vector_model_sig[i] = 0x80; + s->vector_model_dct[i] = 0x80; + s->vector_model_pdi[i][0] = 0x55; + s->vector_model_pdi[i][1] = 0x80; + } + memcpy(s->mb_types_stats, vp56_def_mb_types_stats, sizeof(s->mb_types_stats)); + memset(s->vector_model_pdv, 0x80, sizeof(s->vector_model_pdv)); +} + +static int vp5_decode_init(AVCodecContext *avctx) +{ + vp56_context_t *s = avctx->priv_data; + + vp56_init(s, avctx, 1); + s->vp56_coord_div = vp5_coord_div; + s->parse_vector_adjustment = vp5_parse_vector_adjustment; + s->adjust = vp5_adjust; + s->parse_coeff = vp5_parse_coeff; + s->default_models_init = vp5_default_models_init; + s->parse_vector_models = vp5_parse_vector_models; + s->parse_coeff_models = vp5_parse_coeff_models; + s->parse_header = vp5_parse_header; + + return 0; +} + +AVCodec vp5_decoder = { + "vp5", + CODEC_TYPE_VIDEO, + CODEC_ID_VP5, + sizeof(vp56_context_t), + vp5_decode_init, + NULL, + vp56_free, + vp56_decode_frame, +}; diff --git a/src/libffmpeg/libavcodec/vp56.c b/src/libffmpeg/libavcodec/vp56.c new file mode 100644 index 000000000..eb78d02e4 --- /dev/null +++ b/src/libffmpeg/libavcodec/vp56.c @@ -0,0 +1,665 @@ +/** + * @file vp56.c + * VP5 and VP6 compatible video decoder (common features) + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "avcodec.h" + +#include "vp56.h" +#include "vp56data.h" + + +void vp56_init_dequant(vp56_context_t *s, int quantizer) +{ + s->quantizer = quantizer; + s->dequant_dc = vp56_dc_dequant[quantizer] << 2; + s->dequant_ac = vp56_ac_dequant[quantizer] << 2; +} + +static int vp56_get_vectors_predictors(vp56_context_t *s, int row, int col, + vp56_frame_t ref_frame) +{ + int nb_pred = 0; + vp56_mv_t vect[2] = {{0,0}, {0,0}}; + int pos, offset; + vp56_mv_t mvp; + + for (pos=0; pos<12; pos++) { + mvp.x = col + vp56_candidate_predictor_pos[pos][0]; + mvp.y = row + vp56_candidate_predictor_pos[pos][1]; + if (mvp.x < 0 || mvp.x >= s->mb_width || + mvp.y < 0 || mvp.y >= s->mb_height) + continue; + offset = mvp.x + s->mb_width*mvp.y; + + if (vp56_reference_frame[s->macroblocks[offset].type] != ref_frame) + continue; + if ((s->macroblocks[offset].mv.x == vect[0].x && + s->macroblocks[offset].mv.y == vect[0].y) || + (s->macroblocks[offset].mv.x == 0 && + s->macroblocks[offset].mv.y == 0)) + continue; + + vect[nb_pred++] = s->macroblocks[offset].mv; + if (nb_pred > 1) { + nb_pred = -1; + break; + } + s->vector_candidate_pos = pos; + } + + s->vector_candidate[0] = vect[0]; + s->vector_candidate[1] = vect[1]; + + return nb_pred+1; +} + +static void vp56_parse_mb_type_models(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + int i, ctx, type; + + for (ctx=0; ctx<3; ctx++) { + if (vp56_rac_get_prob(c, 174)) { + int idx = vp56_rac_gets(c, 4); + memcpy(s->mb_types_stats[ctx],vp56_pre_def_mb_type_stats[idx][ctx], + sizeof(s->mb_types_stats[ctx])); + } + if (vp56_rac_get_prob(c, 254)) { + for (type=0; type<10; type++) { + for(i=0; i<2; i++) { + if (vp56_rac_get_prob(c, 205)) { + int delta, sign = vp56_rac_get(c); + + delta = vp56_rac_get_tree(c, vp56_pmbtm_tree, + vp56_mb_type_model_model); + if (!delta) + delta = 4 * vp56_rac_gets(c, 7); + s->mb_types_stats[ctx][type][i] += (delta ^ -sign) + sign; + } + } + } + } + } + + /* compute MB type probability tables based on previous MB type */ + for (ctx=0; ctx<3; ctx++) { + int p[10]; + + for (type=0; type<10; type++) + p[type] = 100 * s->mb_types_stats[ctx][type][1]; + + for (type=0; type<10; type++) { + int p02, p34, p0234, p17, p56, p89, p5689, p156789; + + /* conservative MB type probability */ + s->mb_type_model[ctx][type][0] = 255 - (255 * s->mb_types_stats[ctx][type][0]) / (1 + s->mb_types_stats[ctx][type][0] + s->mb_types_stats[ctx][type][1]); + + p[type] = 0; /* same MB type => weight is null */ + + /* binary tree parsing probabilities */ + p02 = p[0] + p[2]; + p34 = p[3] + p[4]; + p0234 = p02 + p34; + p17 = p[1] + p[7]; + p56 = p[5] + p[6]; + p89 = p[8] + p[9]; + p5689 = p56 + p89; + p156789 = p17 + p5689; + + s->mb_type_model[ctx][type][1] = 1 + 255 * p0234/(1+p0234+p156789); + s->mb_type_model[ctx][type][2] = 1 + 255 * p02 / (1+p0234); + s->mb_type_model[ctx][type][3] = 1 + 255 * p17 / (1+p156789); + s->mb_type_model[ctx][type][4] = 1 + 255 * p[0] / (1+p02); + s->mb_type_model[ctx][type][5] = 1 + 255 * p[3] / (1+p34); + s->mb_type_model[ctx][type][6] = 1 + 255 * p[1] / (1+p17); + s->mb_type_model[ctx][type][7] = 1 + 255 * p56 / (1+p5689); + s->mb_type_model[ctx][type][8] = 1 + 255 * p[5] / (1+p56); + s->mb_type_model[ctx][type][9] = 1 + 255 * p[8] / (1+p89); + + /* restore initial value */ + p[type] = 100 * s->mb_types_stats[ctx][type][1]; + } + } +} + +static vp56_mb_t vp56_parse_mb_type(vp56_context_t *s, + vp56_mb_t prev_type, int ctx) +{ + uint8_t *mb_type_model = s->mb_type_model[ctx][prev_type]; + vp56_range_coder_t *c = &s->c; + + if (vp56_rac_get_prob(c, mb_type_model[0])) + return prev_type; + else + return vp56_rac_get_tree(c, vp56_pmbt_tree, mb_type_model); +} + +static void vp56_decode_4mv(vp56_context_t *s, int row, int col) +{ + vp56_mv_t mv = {0,0}; + int type[4]; + int b; + + /* parse each block type */ + for (b=0; b<4; b++) { + type[b] = vp56_rac_gets(&s->c, 2); + if (type[b]) + type[b]++; /* only returns 0, 2, 3 or 4 (all INTER_PF) */ + } + + /* get vectors */ + for (b=0; b<4; b++) { + switch (type[b]) { + case VP56_MB_INTER_NOVEC_PF: + s->mv[b] = (vp56_mv_t) {0,0}; + break; + case VP56_MB_INTER_DELTA_PF: + s->parse_vector_adjustment(s, &s->mv[b]); + break; + case VP56_MB_INTER_V1_PF: + s->mv[b] = s->vector_candidate[0]; + break; + case VP56_MB_INTER_V2_PF: + s->mv[b] = s->vector_candidate[1]; + break; + } + mv.x += s->mv[b].x; + mv.y += s->mv[b].y; + } + + /* this is the one selected for the whole MB for prediction */ + s->macroblocks[row * s->mb_width + col].mv = s->mv[3]; + + /* chroma vectors are average luma vectors */ + if (s->avctx->codec->id == CODEC_ID_VP5) { + s->mv[4].x = s->mv[5].x = RSHIFT(mv.x,2); + s->mv[4].y = s->mv[5].y = RSHIFT(mv.y,2); + } else { + s->mv[4] = s->mv[5] = (vp56_mv_t) {mv.x/4, mv.y/4}; + } +} + +static vp56_mb_t vp56_decode_mv(vp56_context_t *s, int row, int col) +{ + vp56_mv_t *mv, vect = {0,0}; + int ctx, b; + + ctx = vp56_get_vectors_predictors(s, row, col, VP56_FRAME_PREVIOUS); + s->mb_type = vp56_parse_mb_type(s, s->mb_type, ctx); + s->macroblocks[row * s->mb_width + col].type = s->mb_type; + + switch (s->mb_type) { + case VP56_MB_INTER_V1_PF: + mv = &s->vector_candidate[0]; + break; + + case VP56_MB_INTER_V2_PF: + mv = &s->vector_candidate[1]; + break; + + case VP56_MB_INTER_V1_GF: + vp56_get_vectors_predictors(s, row, col, VP56_FRAME_GOLDEN); + mv = &s->vector_candidate[0]; + break; + + case VP56_MB_INTER_V2_GF: + vp56_get_vectors_predictors(s, row, col, VP56_FRAME_GOLDEN); + mv = &s->vector_candidate[1]; + break; + + case VP56_MB_INTER_DELTA_PF: + s->parse_vector_adjustment(s, &vect); + mv = &vect; + break; + + case VP56_MB_INTER_DELTA_GF: + vp56_get_vectors_predictors(s, row, col, VP56_FRAME_GOLDEN); + s->parse_vector_adjustment(s, &vect); + mv = &vect; + break; + + case VP56_MB_INTER_4V: + vp56_decode_4mv(s, row, col); + return s->mb_type; + + default: + mv = &vect; + break; + } + + s->macroblocks[row*s->mb_width + col].mv = *mv; + + /* same vector for all blocks */ + for (b=0; b<6; b++) + s->mv[b] = *mv; + + return s->mb_type; +} + +static void vp56_add_predictors_dc(vp56_context_t *s, vp56_frame_t ref_frame) +{ + int idx = s->scantable.permutated[0]; + int i; + + for (i=0; i<6; i++) { + vp56_ref_dc_t *ab = &s->above_blocks[s->above_block_idx[i]]; + vp56_ref_dc_t *lb = &s->left_block[vp56_b6to4[i]]; + int count = 0; + int dc = 0; + + if (ref_frame == lb->ref_frame) { + dc += lb->dc_coeff; + count++; + } + if (ref_frame == ab->ref_frame) { + dc += ab->dc_coeff; + count++; + } + if (s->avctx->codec->id == CODEC_ID_VP5) { + if (count < 2 && ref_frame == ab[-1].ref_frame) { + dc += ab[-1].dc_coeff; + count++; + } + if (count < 2 && ref_frame == ab[1].ref_frame) { + dc += ab[1].dc_coeff; + count++; + } + } + if (count == 0) + dc = s->prev_dc[vp56_b6to3[i]][ref_frame]; + else if (count == 2) + dc /= 2; + + s->block_coeff[i][idx] += dc; + s->prev_dc[vp56_b6to3[i]][ref_frame] = s->block_coeff[i][idx]; + ab->dc_coeff = s->block_coeff[i][idx]; + ab->ref_frame = ref_frame; + lb->dc_coeff = s->block_coeff[i][idx]; + lb->ref_frame = ref_frame; + s->block_coeff[i][idx] *= s->dequant_dc; + } +} + +static void vp56_edge_filter(vp56_context_t *s, uint8_t *yuv, + int pix_inc, int line_inc, int t) +{ + int pix2_inc = 2 * pix_inc; + int i, v; + + for (i=0; i<12; i++) { + v = (yuv[-pix2_inc] + 3*(yuv[0]-yuv[-pix_inc]) - yuv[pix_inc] + 4) >>3; + v = s->adjust(v, t); + yuv[-pix_inc] = clip_uint8(yuv[-pix_inc] + v); + yuv[0] = clip_uint8(yuv[0] - v); + yuv += line_inc; + } +} + +static void vp56_deblock_filter(vp56_context_t *s, uint8_t *yuv, + int stride, int dx, int dy) +{ + int t = vp56_filter_threshold[s->quantizer]; + if (dx) vp56_edge_filter(s, yuv + 10-dx , 1, stride, t); + if (dy) vp56_edge_filter(s, yuv + stride*(10-dy), stride, 1, t); +} + +static void vp56_mc(vp56_context_t *s, int b, uint8_t *src, + int stride, int x, int y) +{ + int plane = vp56_b6to3[b]; + uint8_t *dst= s->frames[VP56_FRAME_CURRENT].data[plane]+s->block_offset[b]; + uint8_t *src_block; + int src_offset; + int overlap_offset = 0; + int mask = s->vp56_coord_div[b] - 1; + int deblock_filtering = s->deblock_filtering; + int dx; + int dy; + + if (s->avctx->skip_loop_filter >= AVDISCARD_ALL || + (s->avctx->skip_loop_filter >= AVDISCARD_NONKEY + && !s->frames[VP56_FRAME_CURRENT].key_frame)) + deblock_filtering = 0; + + dx = s->mv[b].x / s->vp56_coord_div[b]; + dy = s->mv[b].y / s->vp56_coord_div[b]; + + if (b >= 4) { + x /= 2; + y /= 2; + } + x += dx - 2; + y += dy - 2; + + if (x<0 || x+12>=s->plane_width[plane] || + y<0 || y+12>=s->plane_height[plane]) { + ff_emulated_edge_mc(s->edge_emu_buffer, + src + s->block_offset[b] + (dy-2)*stride + (dx-2), + stride, 12, 12, x, y, + s->plane_width[plane], + s->plane_height[plane]); + src_block = s->edge_emu_buffer; + src_offset = 2 + 2*stride; + } else if (deblock_filtering) { + /* only need a 12x12 block, but there is no such dsp function, */ + /* so copy a 16x12 block */ + s->dsp.put_pixels_tab[0][0](s->edge_emu_buffer, + src + s->block_offset[b] + (dy-2)*stride + (dx-2), + stride, 12); + src_block = s->edge_emu_buffer; + src_offset = 2 + 2*stride; + } else { + src_block = src; + src_offset = s->block_offset[b] + dy*stride + dx; + } + + if (deblock_filtering) + vp56_deblock_filter(s, src_block, stride, dx&7, dy&7); + + if (s->mv[b].x & mask) + overlap_offset += (s->mv[b].x > 0) ? 1 : -1; + if (s->mv[b].y & mask) + overlap_offset += (s->mv[b].y > 0) ? stride : -stride; + + if (overlap_offset) { + if (s->filter) + s->filter(s, dst, src_block, src_offset, src_offset+overlap_offset, + stride, s->mv[b], mask, s->filter_selection, b<4); + else + s->dsp.put_no_rnd_pixels_l2[1](dst, src_block+src_offset, + src_block+src_offset+overlap_offset, + stride, 8); + } else { + s->dsp.put_pixels_tab[1][0](dst, src_block+src_offset, stride, 8); + } +} + +static void vp56_decode_mb(vp56_context_t *s, int row, int col) +{ + AVFrame *frame_current, *frame_ref; + vp56_mb_t mb_type; + vp56_frame_t ref_frame; + int b, plan, off; + + if (s->frames[VP56_FRAME_CURRENT].key_frame) + mb_type = VP56_MB_INTRA; + else + mb_type = vp56_decode_mv(s, row, col); + ref_frame = vp56_reference_frame[mb_type]; + + memset(s->block_coeff, 0, sizeof(s->block_coeff)); + + s->parse_coeff(s); + + vp56_add_predictors_dc(s, ref_frame); + + frame_current = &s->frames[VP56_FRAME_CURRENT]; + frame_ref = &s->frames[ref_frame]; + + switch (mb_type) { + case VP56_MB_INTRA: + for (b=0; b<6; b++) { + plan = vp56_b6to3[b]; + s->dsp.idct_put(frame_current->data[plan] + s->block_offset[b], + s->stride[plan], s->block_coeff[b]); + } + break; + + case VP56_MB_INTER_NOVEC_PF: + case VP56_MB_INTER_NOVEC_GF: + for (b=0; b<6; b++) { + plan = vp56_b6to3[b]; + off = s->block_offset[b]; + s->dsp.put_pixels_tab[1][0](frame_current->data[plan] + off, + frame_ref->data[plan] + off, + s->stride[plan], 8); + s->dsp.idct_add(frame_current->data[plan] + off, + s->stride[plan], s->block_coeff[b]); + } + break; + + case VP56_MB_INTER_DELTA_PF: + case VP56_MB_INTER_V1_PF: + case VP56_MB_INTER_V2_PF: + case VP56_MB_INTER_DELTA_GF: + case VP56_MB_INTER_4V: + case VP56_MB_INTER_V1_GF: + case VP56_MB_INTER_V2_GF: + for (b=0; b<6; b++) { + int x_off = b==1 || b==3 ? 8 : 0; + int y_off = b==2 || b==3 ? 8 : 0; + plan = vp56_b6to3[b]; + vp56_mc(s, b, frame_ref->data[plan], s->stride[plan], + 16*col+x_off, 16*row+y_off); + s->dsp.idct_add(frame_current->data[plan] + s->block_offset[b], + s->stride[plan], s->block_coeff[b]); + } + break; + } +} + +static int vp56_size_changed(AVCodecContext *avctx, vp56_context_t *s) +{ + int stride = s->frames[VP56_FRAME_CURRENT].linesize[0]; + int i; + + s->plane_width[0] = s->avctx->coded_width; + s->plane_width[1] = s->plane_width[2] = s->avctx->coded_width/2; + s->plane_height[0] = s->avctx->coded_height; + s->plane_height[1] = s->plane_height[2] = s->avctx->coded_height/2; + + for (i=0; i<3; i++) + s->stride[i] = s->flip * s->frames[VP56_FRAME_CURRENT].linesize[i]; + + s->mb_width = (s->avctx->coded_width+15) / 16; + s->mb_height = (s->avctx->coded_height+15) / 16; + + if (s->mb_width > 1000 || s->mb_height > 1000) { + av_log(avctx, AV_LOG_ERROR, "picture too big\n"); + return -1; + } + + s->above_blocks = av_realloc(s->above_blocks, + (4*s->mb_width+6) * sizeof(*s->above_blocks)); + s->macroblocks = av_realloc(s->macroblocks, + s->mb_width*s->mb_height*sizeof(*s->macroblocks)); + av_free(s->edge_emu_buffer_alloc); + s->edge_emu_buffer_alloc = av_malloc(16*stride); + s->edge_emu_buffer = s->edge_emu_buffer_alloc; + if (s->flip < 0) + s->edge_emu_buffer += 15 * stride; + + return 0; +} + +int vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size, + uint8_t *buf, int buf_size) +{ + vp56_context_t *s = avctx->priv_data; + AVFrame *const p = &s->frames[VP56_FRAME_CURRENT]; + AVFrame *picture = data; + int mb_row, mb_col, mb_row_flip, mb_offset = 0; + int block, y, uv, stride_y, stride_uv; + int golden_frame = 0; + int res; + + res = s->parse_header(s, buf, buf_size, &golden_frame); + if (!res) + return -1; + + p->reference = 1; + if (avctx->get_buffer(avctx, p) < 0) { + av_log(avctx, AV_LOG_ERROR, "get_buffer() failed\n"); + return -1; + } + + if (res == 2) + if (vp56_size_changed(avctx, s)) { + avctx->release_buffer(avctx, p); + return -1; + } + + if (p->key_frame) { + p->pict_type = FF_I_TYPE; + s->default_models_init(s); + for (block=0; block<s->mb_height*s->mb_width; block++) + s->macroblocks[block].type = VP56_MB_INTRA; + } else { + p->pict_type = FF_P_TYPE; + vp56_parse_mb_type_models(s); + s->parse_vector_models(s); + s->mb_type = VP56_MB_INTER_NOVEC_PF; + } + + s->parse_coeff_models(s); + + memset(s->prev_dc, 0, sizeof(s->prev_dc)); + s->prev_dc[1][VP56_FRAME_CURRENT] = 128; + s->prev_dc[2][VP56_FRAME_CURRENT] = 128; + + for (block=0; block < 4*s->mb_width+6; block++) { + s->above_blocks[block].ref_frame = -1; + s->above_blocks[block].dc_coeff = 0; + s->above_blocks[block].not_null_dc = 0; + } + s->above_blocks[2*s->mb_width + 2].ref_frame = 0; + s->above_blocks[3*s->mb_width + 4].ref_frame = 0; + + stride_y = p->linesize[0]; + stride_uv = p->linesize[1]; + + if (s->flip < 0) + mb_offset = 7; + + /* main macroblocks loop */ + for (mb_row=0; mb_row<s->mb_height; mb_row++) { + if (s->flip < 0) + mb_row_flip = s->mb_height - mb_row - 1; + else + mb_row_flip = mb_row; + + for (block=0; block<4; block++) { + s->left_block[block].ref_frame = -1; + s->left_block[block].dc_coeff = 0; + s->left_block[block].not_null_dc = 0; + memset(s->coeff_ctx[block], 0, 64*sizeof(s->coeff_ctx[block][0])); + } + memset(s->coeff_ctx_last, 24, sizeof(s->coeff_ctx_last)); + + s->above_block_idx[0] = 1; + s->above_block_idx[1] = 2; + s->above_block_idx[2] = 1; + s->above_block_idx[3] = 2; + s->above_block_idx[4] = 2*s->mb_width + 2 + 1; + s->above_block_idx[5] = 3*s->mb_width + 4 + 1; + + s->block_offset[s->frbi] = (mb_row_flip*16 + mb_offset) * stride_y; + s->block_offset[s->srbi] = s->block_offset[s->frbi] + 8*stride_y; + s->block_offset[1] = s->block_offset[0] + 8; + s->block_offset[3] = s->block_offset[2] + 8; + s->block_offset[4] = (mb_row_flip*8 + mb_offset) * stride_uv; + s->block_offset[5] = s->block_offset[4]; + + for (mb_col=0; mb_col<s->mb_width; mb_col++) { + vp56_decode_mb(s, mb_row, mb_col); + + for (y=0; y<4; y++) { + s->above_block_idx[y] += 2; + s->block_offset[y] += 16; + } + + for (uv=4; uv<6; uv++) { + s->above_block_idx[uv] += 1; + s->block_offset[uv] += 8; + } + } + } + + if (s->frames[VP56_FRAME_PREVIOUS].data[0] + && (s->frames[VP56_FRAME_PREVIOUS].data[0] + != s->frames[VP56_FRAME_GOLDEN].data[0])) { + avctx->release_buffer(avctx, &s->frames[VP56_FRAME_PREVIOUS]); + } + if (p->key_frame || golden_frame) { + if (s->frames[VP56_FRAME_GOLDEN].data[0]) + avctx->release_buffer(avctx, &s->frames[VP56_FRAME_GOLDEN]); + s->frames[VP56_FRAME_GOLDEN] = *p; + } + s->frames[VP56_FRAME_PREVIOUS] = *p; + + *picture = *p; + *data_size = sizeof(AVPicture); + + return buf_size; +} + +void vp56_init(vp56_context_t *s, AVCodecContext *avctx, int flip) +{ + int i; + + s->avctx = avctx; + avctx->pix_fmt = PIX_FMT_YUV420P; + + if (s->avctx->idct_algo == FF_IDCT_AUTO) + s->avctx->idct_algo = FF_IDCT_VP3; + dsputil_init(&s->dsp, s->avctx); + ff_init_scantable(s->dsp.idct_permutation, &s->scantable,ff_zigzag_direct); + + avcodec_set_dimensions(s->avctx, 0, 0); + + for (i=0; i<3; i++) + s->frames[i].data[0] = NULL; + s->edge_emu_buffer_alloc = NULL; + + s->above_blocks = NULL; + s->macroblocks = NULL; + s->quantizer = -1; + s->deblock_filtering = 1; + + s->filter = NULL; + + if (flip) { + s->flip = -1; + s->frbi = 2; + s->srbi = 0; + } else { + s->flip = 1; + s->frbi = 0; + s->srbi = 2; + } +} + +int vp56_free(AVCodecContext *avctx) +{ + vp56_context_t *s = avctx->priv_data; + + av_free(s->above_blocks); + av_free(s->macroblocks); + av_free(s->edge_emu_buffer_alloc); + if (s->frames[VP56_FRAME_GOLDEN].data[0] + && (s->frames[VP56_FRAME_PREVIOUS].data[0] + != s->frames[VP56_FRAME_GOLDEN].data[0])) + avctx->release_buffer(avctx, &s->frames[VP56_FRAME_GOLDEN]); + if (s->frames[VP56_FRAME_PREVIOUS].data[0]) + avctx->release_buffer(avctx, &s->frames[VP56_FRAME_PREVIOUS]); + return 0; +} diff --git a/src/libffmpeg/libavcodec/vp56.h b/src/libffmpeg/libavcodec/vp56.h new file mode 100644 index 000000000..50e201550 --- /dev/null +++ b/src/libffmpeg/libavcodec/vp56.h @@ -0,0 +1,252 @@ +/** + * @file vp56.h + * VP5 and VP6 compatible video decoder (common features) + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef VP56_H +#define VP56_H + +#include "vp56data.h" +#include "dsputil.h" +#include "mpegvideo.h" + + +typedef struct vp56_context vp56_context_t; +typedef struct vp56_mv vp56_mv_t; + +typedef void (*vp56_parse_vector_adjustment_t)(vp56_context_t *s, + vp56_mv_t *vect); +typedef int (*vp56_adjust_t)(int v, int t); +typedef void (*vp56_filter_t)(vp56_context_t *s, uint8_t *dst, uint8_t *src, + int offset1, int offset2, int stride, + vp56_mv_t mv, int mask, int select, int luma); +typedef void (*vp56_parse_coeff_t)(vp56_context_t *s); +typedef void (*vp56_default_models_init_t)(vp56_context_t *s); +typedef void (*vp56_parse_vector_models_t)(vp56_context_t *s); +typedef void (*vp56_parse_coeff_models_t)(vp56_context_t *s); +typedef int (*vp56_parse_header_t)(vp56_context_t *s, uint8_t *buf, + int buf_size, int *golden_frame); + +typedef struct { + int high; + int bits; + const uint8_t *buffer; + unsigned long code_word; +} vp56_range_coder_t; + +typedef struct { + uint8_t not_null_dc; + vp56_frame_t ref_frame; + DCTELEM dc_coeff; +} vp56_ref_dc_t; + +struct vp56_mv { + int x; + int y; +}; + +typedef struct { + uint8_t type; + vp56_mv_t mv; +} vp56_macroblock_t; + +struct vp56_context { + AVCodecContext *avctx; + DSPContext dsp; + ScanTable scantable; + AVFrame frames[3]; + uint8_t *edge_emu_buffer_alloc; + uint8_t *edge_emu_buffer; + vp56_range_coder_t c; + vp56_range_coder_t cc; + vp56_range_coder_t *ccp; + int sub_version; + + /* frame info */ + int plane_width[3]; + int plane_height[3]; + int mb_width; /* number of horizontal MB */ + int mb_height; /* number of vertical MB */ + int block_offset[6]; + + int quantizer; + uint16_t dequant_dc; + uint16_t dequant_ac; + + /* DC predictors management */ + vp56_ref_dc_t *above_blocks; + vp56_ref_dc_t left_block[4]; + int above_block_idx[6]; + DCTELEM prev_dc[3][3]; /* [plan][ref_frame] */ + + /* blocks / macroblock */ + vp56_mb_t mb_type; + vp56_macroblock_t *macroblocks; + DECLARE_ALIGNED_16(DCTELEM, block_coeff[6][64]); + uint8_t coeff_reorder[64]; /* used in vp6 only */ + uint8_t coeff_index_to_pos[64]; /* used in vp6 only */ + + /* motion vectors */ + vp56_mv_t mv[6]; /* vectors for each block in MB */ + vp56_mv_t vector_candidate[2]; + int vector_candidate_pos; + + /* filtering hints */ + int filter_header; /* used in vp6 only */ + int deblock_filtering; + int filter_selection; + int filter_mode; + int max_vector_length; + int sample_variance_threshold; + + /* AC models */ + uint8_t vector_model_sig[2]; /* delta sign */ + uint8_t vector_model_dct[2]; /* delta coding types */ + uint8_t vector_model_pdi[2][2]; /* predefined delta init */ + uint8_t vector_model_pdv[2][7]; /* predefined delta values */ + uint8_t vector_model_fdv[2][8]; /* 8 bit delta value definition */ + uint8_t mb_type_model[3][10][10]; /* model for decoding MB type */ + uint8_t coeff_model_dccv[2][11]; /* DC coeff value */ + uint8_t coeff_model_ract[2][3][6][11]; /* Run/AC coding type and AC coeff value */ + uint8_t coeff_model_acct[2][3][3][6][5];/* vp5 only AC coding type for coding group < 3 */ + uint8_t coeff_model_dcct[2][36][5]; /* DC coeff coding type */ + uint8_t coeff_model_runv[2][14]; /* run value (vp6 only) */ + uint8_t mb_types_stats[3][10][2]; /* contextual, next MB type stats */ + uint8_t coeff_ctx[4][64]; /* used in vp5 only */ + uint8_t coeff_ctx_last[4]; /* used in vp5 only */ + + /* upside-down flipping hints */ + int flip; /* are we flipping ? */ + int frbi; /* first row block index in MB */ + int srbi; /* second row block index in MB */ + int stride[3]; /* stride for each plan */ + + const uint8_t *vp56_coord_div; + vp56_parse_vector_adjustment_t parse_vector_adjustment; + vp56_adjust_t adjust; + vp56_filter_t filter; + vp56_parse_coeff_t parse_coeff; + vp56_default_models_init_t default_models_init; + vp56_parse_vector_models_t parse_vector_models; + vp56_parse_coeff_models_t parse_coeff_models; + vp56_parse_header_t parse_header; +}; + + +void vp56_init(vp56_context_t *s, AVCodecContext *avctx, int flip); +int vp56_free(AVCodecContext *avctx); +void vp56_init_dequant(vp56_context_t *s, int quantizer); +int vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size, + uint8_t *buf, int buf_size); + + +/** + * vp56 specific range coder implementation + */ + +static inline void vp56_init_range_decoder(vp56_range_coder_t *c, + const uint8_t *buf, int buf_size) +{ + c->high = 255; + c->bits = 8; + c->buffer = buf; + c->code_word = *c->buffer++ << 8; + c->code_word |= *c->buffer++; +} + +static inline int vp56_rac_get_prob(vp56_range_coder_t *c, uint8_t prob) +{ + unsigned int low = 1 + (((c->high - 1) * prob) / 256); + unsigned int low_shift = low << 8; + int bit = c->code_word >= low_shift; + + if (bit) { + c->high -= low; + c->code_word -= low_shift; + } else { + c->high = low; + } + + /* normalize */ + while (c->high < 128) { + c->high <<= 1; + c->code_word <<= 1; + if (--c->bits == 0) { + c->bits = 8; + c->code_word |= *c->buffer++; + } + } + return bit; +} + +static inline int vp56_rac_get(vp56_range_coder_t *c) +{ + /* equiprobable */ + int low = (c->high + 1) >> 1; + unsigned int low_shift = low << 8; + int bit = c->code_word >= low_shift; + if (bit) { + c->high = (c->high - low) << 1; + c->code_word -= low_shift; + } else { + c->high = low << 1; + } + + /* normalize */ + c->code_word <<= 1; + if (--c->bits == 0) { + c->bits = 8; + c->code_word |= *c->buffer++; + } + return bit; +} + +static inline int vp56_rac_gets(vp56_range_coder_t *c, int bits) +{ + int value = 0; + + while (bits--) { + value = (value << 1) | vp56_rac_get(c); + } + + return value; +} + +static inline int vp56_rac_gets_nn(vp56_range_coder_t *c, int bits) +{ + int v = vp56_rac_gets(c, 7) << 1; + return v + !v; +} + +static inline int vp56_rac_get_tree(vp56_range_coder_t *c, + const vp56_tree_t *tree, + const uint8_t *probs) +{ + while (tree->val > 0) { + if (vp56_rac_get_prob(c, probs[tree->prob_idx])) + tree += tree->val; + else + tree++; + } + return -tree->val; +} + +#endif /* VP56_H */ diff --git a/src/libffmpeg/libavcodec/vp56data.c b/src/libffmpeg/libavcodec/vp56data.c new file mode 100644 index 000000000..e75c6d1ce --- /dev/null +++ b/src/libffmpeg/libavcodec/vp56data.c @@ -0,0 +1,66 @@ +/** + * @file vp56data.c + * VP5 and VP6 compatible video decoder (common data) + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#include "vp56data.h" + +const uint8_t vp56_b6to3[] = { 0, 0, 0, 0, 1, 2 }; +const uint8_t vp56_b6to4[] = { 0, 0, 1, 1, 2, 3 }; + +const uint8_t vp56_coeff_parse_table[6][11] = { + { 159, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, + { 145, 165, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, + { 140, 148, 173, 0, 0, 0, 0, 0, 0, 0, 0 }, + { 135, 140, 155, 176, 0, 0, 0, 0, 0, 0, 0 }, + { 130, 134, 141, 157, 180, 0, 0, 0, 0, 0, 0 }, + { 129, 130, 133, 140, 153, 177, 196, 230, 243, 254, 254 }, +}; + +const uint8_t vp56_def_mb_types_stats[3][10][2] = { + { { 69, 42 }, { 1, 2 }, { 1, 7 }, { 44, 42 }, { 6, 22 }, + { 1, 3 }, { 0, 2 }, { 1, 5 }, { 0, 1 }, { 0, 0 }, }, + { { 229, 8 }, { 1, 1 }, { 0, 8 }, { 0, 0 }, { 0, 0 }, + { 1, 2 }, { 0, 1 }, { 0, 0 }, { 1, 1 }, { 0, 0 }, }, + { { 122, 35 }, { 1, 1 }, { 1, 6 }, { 46, 34 }, { 0, 0 }, + { 1, 2 }, { 0, 1 }, { 0, 1 }, { 1, 1 }, { 0, 0 }, }, +}; + +const vp56_tree_t vp56_pva_tree[] = { + { 8, 0}, + { 4, 1}, + { 2, 2}, {-0}, {-1}, + { 2, 3}, {-2}, {-3}, + { 4, 4}, + { 2, 5}, {-4}, {-5}, + { 2, 6}, {-6}, {-7}, +}; + +const vp56_tree_t vp56_pc_tree[] = { + { 4, 6}, + { 2, 7}, {-0}, {-1}, + { 4, 8}, + { 2, 9}, {-2}, {-3}, + { 2,10}, {-4}, {-5}, +}; + +const uint8_t vp56_coeff_bias[] = { 5, 7, 11, 19, 35, 67 }; +const uint8_t vp56_coeff_bit_length[] = { 0, 1, 2, 3, 4, 10 }; diff --git a/src/libffmpeg/libavcodec/vp56data.h b/src/libffmpeg/libavcodec/vp56data.h new file mode 100644 index 000000000..dbf92dd68 --- /dev/null +++ b/src/libffmpeg/libavcodec/vp56data.h @@ -0,0 +1,248 @@ +/** + * @file vp56data.h + * VP5 and VP6 compatible video decoder (common data) + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef VP56DATA_H +#define VP56DATA_H + +#include "common.h" + +typedef enum { + VP56_FRAME_CURRENT = 0, + VP56_FRAME_PREVIOUS = 1, + VP56_FRAME_GOLDEN = 2, +} vp56_frame_t; + +typedef enum { + VP56_MB_INTER_NOVEC_PF = 0, /**< Inter MB, no vector, from previous frame */ + VP56_MB_INTRA = 1, /**< Intra MB */ + VP56_MB_INTER_DELTA_PF = 2, /**< Inter MB, above/left vector + delta, from previous frame */ + VP56_MB_INTER_V1_PF = 3, /**< Inter MB, first vector, from previous frame */ + VP56_MB_INTER_V2_PF = 4, /**< Inter MB, second vector, from previous frame */ + VP56_MB_INTER_NOVEC_GF = 5, /**< Inter MB, no vector, from golden frame */ + VP56_MB_INTER_DELTA_GF = 6, /**< Inter MB, above/left vector + delta, from golden frame */ + VP56_MB_INTER_4V = 7, /**< Inter MB, 4 vectors, from previous frame */ + VP56_MB_INTER_V1_GF = 8, /**< Inter MB, first vector, from golden frame */ + VP56_MB_INTER_V2_GF = 9, /**< Inter MB, second vector, from golden frame */ +} vp56_mb_t; + +typedef struct { + int8_t val; + int8_t prob_idx; +} vp56_tree_t; + +extern const uint8_t vp56_b6to3[]; +extern const uint8_t vp56_b6to4[]; +extern const uint8_t vp56_coeff_parse_table[6][11]; +extern const uint8_t vp56_def_mb_types_stats[3][10][2]; +extern const vp56_tree_t vp56_pva_tree[]; +extern const vp56_tree_t vp56_pc_tree[]; +extern const uint8_t vp56_coeff_bias[]; +extern const uint8_t vp56_coeff_bit_length[]; + +static const vp56_frame_t vp56_reference_frame[] = { + VP56_FRAME_PREVIOUS, /* VP56_MB_INTER_NOVEC_PF */ + VP56_FRAME_CURRENT, /* VP56_MB_INTRA */ + VP56_FRAME_PREVIOUS, /* VP56_MB_INTER_DELTA_PF */ + VP56_FRAME_PREVIOUS, /* VP56_MB_INTER_V1_PF */ + VP56_FRAME_PREVIOUS, /* VP56_MB_INTER_V2_PF */ + VP56_FRAME_GOLDEN, /* VP56_MB_INTER_NOVEC_GF */ + VP56_FRAME_GOLDEN, /* VP56_MB_INTER_DELTA_GF */ + VP56_FRAME_PREVIOUS, /* VP56_MB_INTER_4V */ + VP56_FRAME_GOLDEN, /* VP56_MB_INTER_V1_GF */ + VP56_FRAME_GOLDEN, /* VP56_MB_INTER_V2_GF */ +}; + +static const uint8_t vp56_ac_dequant[64] = { + 94, 92, 90, 88, 86, 82, 78, 74, + 70, 66, 62, 58, 54, 53, 52, 51, + 50, 49, 48, 47, 46, 45, 44, 43, + 42, 40, 39, 37, 36, 35, 34, 33, + 32, 31, 30, 29, 28, 27, 26, 25, + 24, 23, 22, 21, 20, 19, 18, 17, + 16, 15, 14, 13, 12, 11, 10, 9, + 8, 7, 6, 5, 4, 3, 2, 1, +}; + +static const uint8_t vp56_dc_dequant[64] = { + 47, 47, 47, 47, 45, 43, 43, 43, + 43, 43, 42, 41, 41, 40, 40, 40, + 40, 35, 35, 35, 35, 33, 33, 33, + 33, 32, 32, 32, 27, 27, 26, 26, + 25, 25, 24, 24, 23, 23, 19, 19, + 19, 19, 18, 18, 17, 16, 16, 16, + 16, 16, 15, 11, 11, 11, 10, 10, + 9, 8, 7, 5, 3, 3, 2, 2, +}; + +static const uint8_t vp56_pre_def_mb_type_stats[16][3][10][2] = { + { { { 9, 15 }, { 32, 25 }, { 7, 19 }, { 9, 21 }, { 1, 12 }, + { 14, 12 }, { 3, 18 }, { 14, 23 }, { 3, 10 }, { 0, 4 }, }, + { { 41, 22 }, { 1, 0 }, { 1, 31 }, { 0, 0 }, { 0, 0 }, + { 0, 1 }, { 1, 7 }, { 0, 1 }, { 98, 25 }, { 4, 10 }, }, + { { 2, 3 }, { 2, 3 }, { 0, 2 }, { 0, 2 }, { 0, 0 }, + { 11, 4 }, { 1, 4 }, { 0, 2 }, { 3, 2 }, { 0, 4 }, }, }, + { { { 48, 39 }, { 1, 2 }, { 11, 27 }, { 29, 44 }, { 7, 27 }, + { 1, 4 }, { 0, 3 }, { 1, 6 }, { 1, 2 }, { 0, 0 }, }, + { { 123, 37 }, { 6, 4 }, { 1, 27 }, { 0, 0 }, { 0, 0 }, + { 5, 8 }, { 1, 7 }, { 0, 1 }, { 12, 10 }, { 0, 2 }, }, + { { 49, 46 }, { 3, 4 }, { 7, 31 }, { 42, 41 }, { 0, 0 }, + { 2, 6 }, { 1, 7 }, { 1, 4 }, { 2, 4 }, { 0, 1 }, }, }, + { { { 21, 32 }, { 1, 2 }, { 4, 10 }, { 32, 43 }, { 6, 23 }, + { 2, 3 }, { 1, 19 }, { 1, 6 }, { 12, 21 }, { 0, 7 }, }, + { { 26, 14 }, { 14, 12 }, { 0, 24 }, { 0, 0 }, { 0, 0 }, + { 55, 17 }, { 1, 9 }, { 0, 36 }, { 5, 7 }, { 1, 3 }, }, + { { 26, 25 }, { 1, 1 }, { 2, 10 }, { 67, 39 }, { 0, 0 }, + { 1, 1 }, { 0, 14 }, { 0, 2 }, { 31, 26 }, { 1, 6 }, }, }, + { { { 69, 83 }, { 0, 0 }, { 0, 2 }, { 10, 29 }, { 3, 12 }, + { 0, 1 }, { 0, 3 }, { 0, 3 }, { 2, 2 }, { 0, 0 }, }, + { { 209, 5 }, { 0, 0 }, { 0, 27 }, { 0, 0 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 0 }, }, + { { 103, 46 }, { 1, 2 }, { 2, 10 }, { 33, 42 }, { 0, 0 }, + { 1, 4 }, { 0, 3 }, { 0, 1 }, { 1, 3 }, { 0, 0 }, }, }, + { { { 11, 20 }, { 1, 4 }, { 18, 36 }, { 43, 48 }, { 13, 35 }, + { 0, 2 }, { 0, 5 }, { 3, 12 }, { 1, 2 }, { 0, 0 }, }, + { { 2, 5 }, { 4, 5 }, { 0, 121 }, { 0, 0 }, { 0, 0 }, + { 0, 3 }, { 2, 4 }, { 1, 4 }, { 2, 2 }, { 0, 1 }, }, + { { 14, 31 }, { 9, 13 }, { 14, 54 }, { 22, 29 }, { 0, 0 }, + { 2, 6 }, { 4, 18 }, { 6, 13 }, { 1, 5 }, { 0, 1 }, }, }, + { { { 70, 44 }, { 0, 1 }, { 2, 10 }, { 37, 46 }, { 8, 26 }, + { 0, 2 }, { 0, 2 }, { 0, 2 }, { 0, 1 }, { 0, 0 }, }, + { { 175, 5 }, { 0, 1 }, { 0, 48 }, { 0, 0 }, { 0, 0 }, + { 0, 2 }, { 0, 1 }, { 0, 2 }, { 0, 1 }, { 0, 0 }, }, + { { 85, 39 }, { 0, 0 }, { 1, 9 }, { 69, 40 }, { 0, 0 }, + { 0, 1 }, { 0, 3 }, { 0, 1 }, { 2, 3 }, { 0, 0 }, }, }, + { { { 8, 15 }, { 0, 1 }, { 8, 21 }, { 74, 53 }, { 22, 42 }, + { 0, 1 }, { 0, 2 }, { 0, 3 }, { 1, 2 }, { 0, 0 }, }, + { { 83, 5 }, { 2, 3 }, { 0, 102 }, { 0, 0 }, { 0, 0 }, + { 1, 3 }, { 0, 2 }, { 0, 1 }, { 0, 0 }, { 0, 0 }, }, + { { 31, 28 }, { 0, 0 }, { 3, 14 }, { 130, 34 }, { 0, 0 }, + { 0, 1 }, { 0, 3 }, { 0, 1 }, { 3, 3 }, { 0, 1 }, }, }, + { { { 141, 42 }, { 0, 0 }, { 1, 4 }, { 11, 24 }, { 1, 11 }, + { 0, 1 }, { 0, 1 }, { 0, 2 }, { 0, 0 }, { 0, 0 }, }, + { { 233, 6 }, { 0, 0 }, { 0, 8 }, { 0, 0 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 1 }, { 0, 0 }, }, + { { 171, 25 }, { 0, 0 }, { 1, 5 }, { 25, 21 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 0 }, { 0, 0 }, }, }, + { { { 8, 19 }, { 4, 10 }, { 24, 45 }, { 21, 37 }, { 9, 29 }, + { 0, 3 }, { 1, 7 }, { 11, 25 }, { 0, 2 }, { 0, 1 }, }, + { { 34, 16 }, { 112, 21 }, { 1, 28 }, { 0, 0 }, { 0, 0 }, + { 6, 8 }, { 1, 7 }, { 0, 3 }, { 2, 5 }, { 0, 2 }, }, + { { 17, 21 }, { 68, 29 }, { 6, 15 }, { 13, 22 }, { 0, 0 }, + { 6, 12 }, { 3, 14 }, { 4, 10 }, { 1, 7 }, { 0, 3 }, }, }, + { { { 46, 42 }, { 0, 1 }, { 2, 10 }, { 54, 51 }, { 10, 30 }, + { 0, 2 }, { 0, 2 }, { 0, 1 }, { 0, 1 }, { 0, 0 }, }, + { { 159, 35 }, { 2, 2 }, { 0, 25 }, { 0, 0 }, { 0, 0 }, + { 3, 6 }, { 0, 5 }, { 0, 1 }, { 4, 4 }, { 0, 1 }, }, + { { 51, 39 }, { 0, 1 }, { 2, 12 }, { 91, 44 }, { 0, 0 }, + { 0, 2 }, { 0, 3 }, { 0, 1 }, { 2, 3 }, { 0, 1 }, }, }, + { { { 28, 32 }, { 0, 0 }, { 3, 10 }, { 75, 51 }, { 14, 33 }, + { 0, 1 }, { 0, 2 }, { 0, 1 }, { 1, 2 }, { 0, 0 }, }, + { { 75, 39 }, { 5, 7 }, { 2, 48 }, { 0, 0 }, { 0, 0 }, + { 3, 11 }, { 2, 16 }, { 1, 4 }, { 7, 10 }, { 0, 2 }, }, + { { 81, 25 }, { 0, 0 }, { 2, 9 }, { 106, 26 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 1, 1 }, { 0, 0 }, }, }, + { { { 100, 46 }, { 0, 1 }, { 3, 9 }, { 21, 37 }, { 5, 20 }, + { 0, 1 }, { 0, 2 }, { 1, 2 }, { 0, 1 }, { 0, 0 }, }, + { { 212, 21 }, { 0, 1 }, { 0, 9 }, { 0, 0 }, { 0, 0 }, + { 1, 2 }, { 0, 2 }, { 0, 0 }, { 2, 2 }, { 0, 0 }, }, + { { 140, 37 }, { 0, 1 }, { 1, 8 }, { 24, 33 }, { 0, 0 }, + { 1, 2 }, { 0, 2 }, { 0, 1 }, { 1, 2 }, { 0, 0 }, }, }, + { { { 27, 29 }, { 0, 1 }, { 9, 25 }, { 53, 51 }, { 12, 34 }, + { 0, 1 }, { 0, 3 }, { 1, 5 }, { 0, 2 }, { 0, 0 }, }, + { { 4, 2 }, { 0, 0 }, { 0, 172 }, { 0, 0 }, { 0, 0 }, + { 0, 1 }, { 0, 2 }, { 0, 0 }, { 2, 0 }, { 0, 0 }, }, + { { 14, 23 }, { 1, 3 }, { 11, 53 }, { 90, 31 }, { 0, 0 }, + { 0, 3 }, { 1, 5 }, { 2, 6 }, { 1, 2 }, { 0, 0 }, }, }, + { { { 80, 38 }, { 0, 0 }, { 1, 4 }, { 69, 33 }, { 5, 16 }, + { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 1 }, { 0, 0 }, }, + { { 187, 22 }, { 1, 1 }, { 0, 17 }, { 0, 0 }, { 0, 0 }, + { 3, 6 }, { 0, 4 }, { 0, 1 }, { 4, 4 }, { 0, 1 }, }, + { { 123, 29 }, { 0, 0 }, { 1, 7 }, { 57, 30 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 0, 1 }, { 0, 0 }, }, }, + { { { 16, 20 }, { 0, 0 }, { 2, 8 }, { 104, 49 }, { 15, 33 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 1, 1 }, { 0, 0 }, }, + { { 133, 6 }, { 1, 2 }, { 1, 70 }, { 0, 0 }, { 0, 0 }, + { 0, 2 }, { 0, 4 }, { 0, 3 }, { 1, 1 }, { 0, 0 }, }, + { { 13, 14 }, { 0, 0 }, { 4, 20 }, { 175, 20 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 1, 1 }, { 0, 0 }, }, }, + { { { 194, 16 }, { 0, 0 }, { 1, 1 }, { 1, 9 }, { 1, 3 }, + { 0, 0 }, { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 0 }, }, + { { 251, 1 }, { 0, 0 }, { 0, 2 }, { 0, 0 }, { 0, 0 }, + { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0, 0 }, }, + { { 202, 23 }, { 0, 0 }, { 1, 3 }, { 2, 9 }, { 0, 0 }, + { 0, 1 }, { 0, 1 }, { 0, 1 }, { 0, 0 }, { 0, 0 }, }, }, +}; + +static const uint8_t vp56_filter_threshold[] = { + 14, 14, 13, 13, 12, 12, 10, 10, + 10, 10, 8, 8, 8, 8, 8, 8, + 8, 8, 8, 8, 8, 8, 8, 8, + 8, 8, 8, 8, 8, 8, 8, 8, + 8, 8, 8, 8, 7, 7, 7, 7, + 7, 7, 6, 6, 6, 6, 6, 6, + 5, 5, 5, 5, 4, 4, 4, 4, + 4, 4, 4, 3, 3, 3, 3, 2, +}; + +static const uint8_t vp56_mb_type_model_model[] = { + 171, 83, 199, 140, 125, 104, +}; + +static const vp56_tree_t vp56_pmbtm_tree[] = { + { 4, 0}, + { 2, 1}, {-8}, {-4}, + { 8, 2}, + { 6, 3}, + { 4, 4}, + { 2, 5}, {-24}, {-20}, {-16}, {-12}, {-0}, +}; + +static const vp56_tree_t vp56_pmbt_tree[] = { + { 8, 1}, + { 4, 2}, + { 2, 4}, {-VP56_MB_INTER_NOVEC_PF}, {-VP56_MB_INTER_DELTA_PF}, + { 2, 5}, {-VP56_MB_INTER_V1_PF}, {-VP56_MB_INTER_V2_PF}, + { 4, 3}, + { 2, 6}, {-VP56_MB_INTRA}, {-VP56_MB_INTER_4V}, + { 4, 7}, + { 2, 8}, {-VP56_MB_INTER_NOVEC_GF}, {-VP56_MB_INTER_DELTA_GF}, + { 2, 9}, {-VP56_MB_INTER_V1_GF}, {-VP56_MB_INTER_V2_GF}, +}; + +/* relative pos of surrounding blocks, from closest to farthest */ +static const int8_t vp56_candidate_predictor_pos[12][2] = { + { 0, -1 }, + { -1, 0 }, + { -1, -1 }, + { 1, -1 }, + { 0, -2 }, + { -2, 0 }, + { -2, -1 }, + { -1, -2 }, + { 1, -2 }, + { 2, -1 }, + { -2, -2 }, + { 2, -2 }, +}; + +#endif /* VP56DATA */ diff --git a/src/libffmpeg/libavcodec/vp5data.h b/src/libffmpeg/libavcodec/vp5data.h new file mode 100644 index 000000000..effc17c2c --- /dev/null +++ b/src/libffmpeg/libavcodec/vp5data.h @@ -0,0 +1,173 @@ +/** + * @file vp5data.h + * VP5 compatible video decoder + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef VP5DATA_H +#define VP5DATA_H + +static const uint8_t vp5_coeff_groups[] = { + -1, 0, 1, 1, 2, 1, 1, 2, + 2, 1, 1, 2, 2, 2, 1, 2, + 2, 2, 2, 2, 1, 1, 2, 2, + 3, 3, 4, 3, 4, 4, 4, 3, + 3, 3, 3, 3, 4, 3, 3, 3, + 4, 4, 4, 4, 4, 3, 3, 4, + 4, 4, 3, 4, 4, 4, 4, 4, + 4, 4, 5, 5, 5, 5, 5, 5, +}; + +static const uint8_t vp5_vmc_pct[2][11] = { + { 243, 220, 251, 253, 237, 232, 241, 245, 247, 251, 253 }, + { 235, 211, 246, 249, 234, 231, 248, 249, 252, 252, 254 }, +}; + +static const uint8_t vp5_dccv_pct[2][11] = { + { 146, 197, 181, 207, 232, 243, 238, 251, 244, 250, 249 }, + { 179, 219, 214, 240, 250, 254, 244, 254, 254, 254, 254 }, +}; + +static const uint8_t vp5_ract_pct[3][2][6][11] = { + { { { 227, 246, 230, 247, 244, 254, 254, 254, 254, 254, 254 }, + { 202, 254, 209, 231, 231, 249, 249, 253, 254, 254, 254 }, + { 206, 254, 225, 242, 241, 251, 253, 254, 254, 254, 254 }, + { 235, 254, 241, 253, 252, 254, 254, 254, 254, 254, 254 }, + { 234, 254, 248, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } }, + { { 240, 254, 248, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 238, 254, 240, 253, 254, 254, 254, 254, 254, 254, 254 }, + { 244, 254, 251, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } } }, + { { { 206, 203, 227, 239, 247, 254, 253, 254, 254, 254, 254 }, + { 207, 199, 220, 236, 243, 252, 252, 254, 254, 254, 254 }, + { 212, 219, 230, 243, 244, 253, 252, 254, 254, 254, 254 }, + { 236, 237, 247, 252, 253, 254, 254, 254, 254, 254, 254 }, + { 240, 240, 248, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } }, + { { 230, 233, 249, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 238, 238, 250, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 248, 251, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } } }, + { { { 225, 239, 227, 231, 244, 253, 243, 254, 254, 253, 254 }, + { 232, 234, 224, 228, 242, 249, 242, 252, 251, 251, 254 }, + { 235, 249, 238, 240, 251, 254, 249, 254, 253, 253, 254 }, + { 249, 253, 251, 250, 254, 254, 254, 254, 254, 254, 254 }, + { 251, 250, 249, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } }, + { { 243, 244, 250, 250, 254, 254, 254, 254, 254, 254, 254 }, + { 249, 248, 250, 253, 254, 254, 254, 254, 254, 254, 254 }, + { 253, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 }, + { 254, 254, 254, 254, 254, 254, 254, 254, 254, 254, 254 } } }, +}; + +static const int16_t vp5_dccv_lc[5][36][2] = { + { {154, 61}, {141, 54}, { 90, 45}, { 54, 34}, { 54, 13}, {128, 109}, + {136, 54}, {148, 45}, { 92, 41}, { 54, 33}, { 51, 15}, { 87, 113}, + { 87, 44}, { 97, 40}, { 67, 36}, { 46, 29}, { 41, 15}, { 64, 80}, + { 59, 33}, { 61, 31}, { 51, 28}, { 44, 22}, { 33, 12}, { 49, 63}, + { 69, 12}, { 59, 16}, { 46, 14}, { 31, 13}, { 26, 6}, { 92, 26}, + {128, 108}, { 77, 119}, { 54, 84}, { 26, 71}, { 87, 19}, { 95, 155} }, + { {154, 4}, {182, 0}, {159, -8}, {128, -5}, {143, -5}, {187, 55}, + {182, 0}, {228, -3}, {187, -7}, {174, -9}, {189, -11}, {169, 79}, + {161, -9}, {192, -8}, {187, -9}, {169, -10}, {136, -9}, {184, 40}, + {164, -11}, {179, -10}, {174, -10}, {161, -10}, {115, -7}, {197, 20}, + {195, -11}, {195, -11}, {146, -10}, {110, -6}, { 95, -4}, {195, 39}, + {182, 55}, {172, 77}, {177, 37}, {169, 29}, {172, 52}, { 92, 162} }, + { {174, 80}, {164, 80}, { 95, 80}, { 46, 66}, { 56, 24}, { 36, 193}, + {164, 80}, {166, 77}, {105, 76}, { 49, 68}, { 46, 31}, { 49, 186}, + { 97, 78}, {110, 74}, { 72, 72}, { 44, 60}, { 33, 30}, { 69, 131}, + { 61, 61}, { 69, 63}, { 51, 57}, { 31, 48}, { 26, 27}, { 64, 89}, + { 67, 23}, { 51, 32}, { 36, 33}, { 26, 28}, { 20, 12}, { 44, 68}, + { 26, 197}, { 41, 189}, { 61, 129}, { 28, 103}, { 49, 52}, {-12, 245} }, + { {102, 141}, { 79, 166}, { 72, 162}, { 97, 125}, {179, 4}, {307, 0}, + { 72, 168}, { 69, 175}, { 84, 160}, {105, 127}, {148, 34}, {310, 0}, + { 84, 151}, { 82, 161}, { 87, 153}, { 87, 135}, {115, 51}, {317, 0}, + { 97, 125}, {102, 131}, {105, 125}, { 87, 122}, { 84, 64}, { 54, 184}, + {166, 18}, {146, 43}, {125, 51}, { 90, 64}, { 95, 7}, { 38, 154}, + {294, 0}, { 13, 225}, { 10, 225}, { 67, 168}, { 0, 167}, {161, 94} }, + { {172, 76}, {172, 75}, {136, 80}, { 64, 98}, { 74, 67}, {315, 0}, + {169, 76}, {207, 56}, {164, 66}, { 97, 80}, { 67, 72}, {328, 0}, + {136, 80}, {187, 53}, {154, 62}, { 72, 85}, { -2, 105}, {305, 0}, + { 74, 91}, {128, 64}, {113, 64}, { 61, 77}, { 41, 75}, {259, 0}, + { 46, 84}, { 51, 81}, { 28, 89}, { 31, 78}, { 23, 77}, {202, 0}, + {323, 0}, {323, 0}, {300, 0}, {236, 0}, {195, 0}, {328, 0} }, +}; + +static const int16_t vp5_ract_lc[3][3][5][6][2] = { + { { { {276, 0}, {238, 0}, {195, 0}, {156, 0}, {113, 0}, {274, 0} }, + { { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1} }, + { {192, 59}, {182, 50}, {141, 48}, {110, 40}, { 92, 19}, {125,128} }, + { {169, 87}, {169, 83}, {184, 62}, {220, 16}, {184, 0}, {264, 0} }, + { {212, 40}, {212, 36}, {169, 49}, {174, 27}, { 8,120}, {182, 71} } }, + { { {259, 10}, {197, 19}, {143, 22}, {123, 16}, {110, 8}, {133, 88} }, + { { 0, 1}, {256, 0}, { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1} }, + { {207, 46}, {187, 50}, { 97, 83}, { 23,100}, { 41, 56}, { 56,188} }, + { {166, 90}, {146,108}, {161, 88}, {136, 95}, {174, 0}, {266, 0} }, + { {264, 7}, {243, 18}, {184, 43}, {-14,154}, { 20,112}, { 20,199} } }, + { { {230, 26}, {197, 22}, {159, 20}, {146, 12}, {136, 4}, { 54,162} }, + { { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1}, { 0, 1} }, + { {192, 59}, {156, 72}, { 84,101}, { 49,101}, { 79, 47}, { 79,167} }, + { {138,115}, {136,116}, {166, 80}, {238, 0}, {195, 0}, {261, 0} }, + { {225, 33}, {205, 42}, {159, 61}, { 79, 96}, { 92, 66}, { 28,195} } }, + }, { + { { {200, 37}, {197, 18}, {159, 13}, {143, 7}, {102, 5}, {123,126} }, + { {197, 3}, {220, -9}, {210,-12}, {187, -6}, {151, -2}, {174, 80} }, + { {200, 53}, {187, 47}, {159, 40}, {118, 38}, {100, 18}, {141,111} }, + { {179, 78}, {166, 86}, {197, 50}, {207, 27}, {187, 0}, {115,139} }, + { {218, 34}, {220, 29}, {174, 46}, {128, 61}, { 54, 89}, {187, 65} } }, + { { {238, 14}, {197, 18}, {125, 26}, { 90, 25}, { 82, 13}, {161, 86} }, + { {189, 1}, {205, -2}, {156, -4}, {143, -4}, {146, -4}, {172, 72} }, + { {230, 31}, {192, 45}, {102, 76}, { 38, 85}, { 56, 41}, { 64,173} }, + { {166, 91}, {141,111}, {128,116}, {118,109}, {177, 0}, { 23,222} }, + { {253, 14}, {236, 21}, {174, 49}, { 33,118}, { 44, 93}, { 23,187} } }, + { { {218, 28}, {179, 28}, {118, 35}, { 95, 30}, { 72, 24}, {128,108} }, + { {187, 1}, {174, -1}, {125, -1}, {110, -1}, {108, -1}, {202, 52} }, + { {197, 53}, {146, 75}, { 46,118}, { 33,103}, { 64, 50}, {118,126} }, + { {138,114}, {128,122}, {161, 86}, {243, -6}, {195, 0}, { 38,210} }, + { {215, 39}, {179, 58}, { 97,101}, { 95, 85}, { 87, 70}, { 69,152} } }, + }, { + { { {236, 24}, {205, 18}, {172, 12}, {154, 6}, {125, 1}, {169, 75} }, + { {187, 4}, {230, -2}, {228, -4}, {236, -4}, {241, -2}, {192, 66} }, + { {200, 46}, {187, 42}, {159, 34}, {136, 25}, {105, 10}, {179, 62} }, + { {207, 55}, {192, 63}, {192, 54}, {195, 36}, {177, 1}, {143, 98} }, + { {225, 27}, {207, 34}, {200, 30}, {131, 57}, { 97, 60}, {197, 45} } }, + { { {271, 8}, {218, 13}, {133, 19}, { 90, 19}, { 72, 7}, {182, 51} }, + { {179, 1}, {225, -1}, {154, -2}, {110, -1}, { 92, 0}, {195, 41} }, + { {241, 26}, {189, 40}, { 82, 64}, { 33, 60}, { 67, 17}, {120, 94} }, + { {192, 68}, {151, 94}, {146, 90}, {143, 72}, {161, 0}, {113,128} }, + { {256, 12}, {218, 29}, {166, 48}, { 44, 99}, { 31, 87}, {148, 78} } }, + { { {238, 20}, {184, 22}, {113, 27}, { 90, 22}, { 74, 9}, {192, 37} }, + { {184, 0}, {215, -1}, {141, -1}, { 97, 0}, { 49, 0}, {264, 13} }, + { {182, 51}, {138, 61}, { 95, 63}, { 54, 59}, { 64, 25}, {200, 45} }, + { {179, 75}, {156, 87}, {174, 65}, {177, 44}, {174, 0}, {164, 85} }, + { {195, 45}, {148, 65}, {105, 79}, { 95, 72}, { 87, 60}, {169, 63} } }, + } +}; + +static const uint8_t vp5_coord_div[] = { 2, 2, 2, 2, 4, 4 }; + +#endif /* VP5DATA_H */ diff --git a/src/libffmpeg/libavcodec/vp6.c b/src/libffmpeg/libavcodec/vp6.c new file mode 100644 index 000000000..df4ebf87d --- /dev/null +++ b/src/libffmpeg/libavcodec/vp6.c @@ -0,0 +1,556 @@ +/** + * @file vp6.c + * VP6 compatible video decoder + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + * + * + * The VP6F decoder accept an optional 1 byte extradata. It is composed of: + * - upper 4bits: difference between encoded width and visible width + * - lower 4bits: difference between encoded height and visible height + */ + +#include <stdlib.h> + +#include "avcodec.h" +#include "dsputil.h" +#include "bitstream.h" +#include "mpegvideo.h" + +#include "vp56.h" +#include "vp56data.h" +#include "vp6data.h" + + +static int vp6_parse_header(vp56_context_t *s, uint8_t *buf, int buf_size, + int *golden_frame) +{ + vp56_range_coder_t *c = &s->c; + int parse_filter_info = 0; + int coeff_offset = 0; + int vrt_shift = 0; + int sub_version; + int rows, cols; + int res = 1; + int separated_coeff = buf[0] & 1; + + s->frames[VP56_FRAME_CURRENT].key_frame = !(buf[0] & 0x80); + vp56_init_dequant(s, (buf[0] >> 1) & 0x3F); + + if (s->frames[VP56_FRAME_CURRENT].key_frame) { + sub_version = buf[1] >> 3; + if (sub_version > 8) + return 0; + s->filter_header = buf[1] & 0x06; + if (buf[1] & 1) { + av_log(s->avctx, AV_LOG_ERROR, "interlacing not supported\n"); + return 0; + } + if (separated_coeff || !s->filter_header) { + coeff_offset = AV_RB16(buf+2) - 2; + buf += 2; + buf_size -= 2; + } + + rows = buf[2]; /* number of stored macroblock rows */ + cols = buf[3]; /* number of stored macroblock cols */ + /* buf[4] is number of displayed macroblock rows */ + /* buf[5] is number of displayed macroblock cols */ + + if (16*cols != s->avctx->coded_width || + 16*rows != s->avctx->coded_height) { + avcodec_set_dimensions(s->avctx, 16*cols, 16*rows); + if (s->avctx->extradata_size == 1) { + s->avctx->width -= s->avctx->extradata[0] >> 4; + s->avctx->height -= s->avctx->extradata[0] & 0x0F; + } + res = 2; + } + + vp56_init_range_decoder(c, buf+6, buf_size-6); + vp56_rac_gets(c, 2); + + parse_filter_info = s->filter_header; + if (sub_version < 8) + vrt_shift = 5; + s->sub_version = sub_version; + } else { + if (!s->sub_version) + return 0; + + if (separated_coeff || !s->filter_header) { + coeff_offset = AV_RB16(buf+1) - 2; + buf += 2; + buf_size -= 2; + } + vp56_init_range_decoder(c, buf+1, buf_size-1); + + *golden_frame = vp56_rac_get(c); + if (s->filter_header) { + s->deblock_filtering = vp56_rac_get(c); + if (s->deblock_filtering) + vp56_rac_get(c); + if (s->sub_version > 7) + parse_filter_info = vp56_rac_get(c); + } + } + + if (parse_filter_info) { + if (vp56_rac_get(c)) { + s->filter_mode = 2; + s->sample_variance_threshold = vp56_rac_gets(c, 5) << vrt_shift; + s->max_vector_length = 2 << vp56_rac_gets(c, 3); + } else if (vp56_rac_get(c)) { + s->filter_mode = 1; + } else { + s->filter_mode = 0; + } + if (s->sub_version > 7) + s->filter_selection = vp56_rac_gets(c, 4); + else + s->filter_selection = 16; + } + + vp56_rac_get(c); + + if (coeff_offset) { + vp56_init_range_decoder(&s->cc, buf+coeff_offset, + buf_size-coeff_offset); + s->ccp = &s->cc; + } else { + s->ccp = &s->c; + } + + return res; +} + +static void vp6_coeff_order_table_init(vp56_context_t *s) +{ + int i, pos, idx = 1; + + s->coeff_index_to_pos[0] = 0; + for (i=0; i<16; i++) + for (pos=1; pos<64; pos++) + if (s->coeff_reorder[pos] == i) + s->coeff_index_to_pos[idx++] = pos; +} + +static void vp6_default_models_init(vp56_context_t *s) +{ + s->vector_model_dct[0] = 0xA2; + s->vector_model_dct[1] = 0xA4; + s->vector_model_sig[0] = 0x80; + s->vector_model_sig[1] = 0x80; + + memcpy(s->mb_types_stats, vp56_def_mb_types_stats, sizeof(s->mb_types_stats)); + memcpy(s->vector_model_fdv, vp6_def_fdv_vector_model, sizeof(s->vector_model_fdv)); + memcpy(s->vector_model_pdv, vp6_def_pdv_vector_model, sizeof(s->vector_model_pdv)); + memcpy(s->coeff_model_runv, vp6_def_runv_coeff_model, sizeof(s->coeff_model_runv)); + memcpy(s->coeff_reorder, vp6_def_coeff_reorder, sizeof(s->coeff_reorder)); + + vp6_coeff_order_table_init(s); +} + +static void vp6_parse_vector_models(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + int comp, node; + + for (comp=0; comp<2; comp++) { + if (vp56_rac_get_prob(c, vp6_sig_dct_pct[comp][0])) + s->vector_model_dct[comp] = vp56_rac_gets_nn(c, 7); + if (vp56_rac_get_prob(c, vp6_sig_dct_pct[comp][1])) + s->vector_model_sig[comp] = vp56_rac_gets_nn(c, 7); + } + + for (comp=0; comp<2; comp++) + for (node=0; node<7; node++) + if (vp56_rac_get_prob(c, vp6_pdv_pct[comp][node])) + s->vector_model_pdv[comp][node] = vp56_rac_gets_nn(c, 7); + + for (comp=0; comp<2; comp++) + for (node=0; node<8; node++) + if (vp56_rac_get_prob(c, vp6_fdv_pct[comp][node])) + s->vector_model_fdv[comp][node] = vp56_rac_gets_nn(c, 7); +} + +static void vp6_parse_coeff_models(vp56_context_t *s) +{ + vp56_range_coder_t *c = &s->c; + int def_prob[11]; + int node, cg, ctx, pos; + int ct; /* code type */ + int pt; /* plane type (0 for Y, 1 for U or V) */ + + memset(def_prob, 0x80, sizeof(def_prob)); + + for (pt=0; pt<2; pt++) + for (node=0; node<11; node++) + if (vp56_rac_get_prob(c, vp6_dccv_pct[pt][node])) { + def_prob[node] = vp56_rac_gets_nn(c, 7); + s->coeff_model_dccv[pt][node] = def_prob[node]; + } else if (s->frames[VP56_FRAME_CURRENT].key_frame) { + s->coeff_model_dccv[pt][node] = def_prob[node]; + } + + if (vp56_rac_get(c)) { + for (pos=1; pos<64; pos++) + if (vp56_rac_get_prob(c, vp6_coeff_reorder_pct[pos])) + s->coeff_reorder[pos] = vp56_rac_gets(c, 4); + vp6_coeff_order_table_init(s); + } + + for (cg=0; cg<2; cg++) + for (node=0; node<14; node++) + if (vp56_rac_get_prob(c, vp6_runv_pct[cg][node])) + s->coeff_model_runv[cg][node] = vp56_rac_gets_nn(c, 7); + + for (ct=0; ct<3; ct++) + for (pt=0; pt<2; pt++) + for (cg=0; cg<6; cg++) + for (node=0; node<11; node++) + if (vp56_rac_get_prob(c, vp6_ract_pct[ct][pt][cg][node])) { + def_prob[node] = vp56_rac_gets_nn(c, 7); + s->coeff_model_ract[pt][ct][cg][node] = def_prob[node]; + } else if (s->frames[VP56_FRAME_CURRENT].key_frame) { + s->coeff_model_ract[pt][ct][cg][node] = def_prob[node]; + } + + /* coeff_model_dcct is a linear combination of coeff_model_dccv */ + for (pt=0; pt<2; pt++) + for (ctx=0; ctx<3; ctx++) + for (node=0; node<5; node++) + s->coeff_model_dcct[pt][ctx][node] = clip(((s->coeff_model_dccv[pt][node] * vp6_dccv_lc[ctx][node][0] + 128) >> 8) + vp6_dccv_lc[ctx][node][1], 1, 255); +} + +static void vp6_parse_vector_adjustment(vp56_context_t *s, vp56_mv_t *vect) +{ + vp56_range_coder_t *c = &s->c; + int comp; + + *vect = (vp56_mv_t) {0,0}; + if (s->vector_candidate_pos < 2) + *vect = s->vector_candidate[0]; + + for (comp=0; comp<2; comp++) { + int i, delta = 0; + + if (vp56_rac_get_prob(c, s->vector_model_dct[comp])) { + static const uint8_t prob_order[] = {0, 1, 2, 7, 6, 5, 4}; + for (i=0; i<sizeof(prob_order); i++) { + int j = prob_order[i]; + delta |= vp56_rac_get_prob(c, s->vector_model_fdv[comp][j])<<j; + } + if (delta & 0xF0) + delta |= vp56_rac_get_prob(c, s->vector_model_fdv[comp][3])<<3; + else + delta |= 8; + } else { + delta = vp56_rac_get_tree(c, vp56_pva_tree, + s->vector_model_pdv[comp]); + } + + if (delta && vp56_rac_get_prob(c, s->vector_model_sig[comp])) + delta = -delta; + + if (!comp) + vect->x += delta; + else + vect->y += delta; + } +} + +static void vp6_parse_coeff(vp56_context_t *s) +{ + vp56_range_coder_t *c = s->ccp; + uint8_t *permute = s->scantable.permutated; + uint8_t *model, *model2, *model3; + int coeff, sign, coeff_idx; + int b, i, cg, idx, ctx; + int pt = 0; /* plane type (0 for Y, 1 for U or V) */ + + for (b=0; b<6; b++) { + int ct = 1; /* code type */ + int run = 1; + + if (b > 3) pt = 1; + + ctx = s->left_block[vp56_b6to4[b]].not_null_dc + + s->above_blocks[s->above_block_idx[b]].not_null_dc; + model = s->coeff_model_dccv[pt]; + model2 = s->coeff_model_dcct[pt][ctx]; + + for (coeff_idx=0; coeff_idx<64; ) { + if ((coeff_idx>1 && ct==0) || vp56_rac_get_prob(c, model2[0])) { + /* parse a coeff */ + if (coeff_idx == 0) { + s->left_block[vp56_b6to4[b]].not_null_dc = 1; + s->above_blocks[s->above_block_idx[b]].not_null_dc = 1; + } + + if (vp56_rac_get_prob(c, model2[2])) { + if (vp56_rac_get_prob(c, model2[3])) { + idx = vp56_rac_get_tree(c, vp56_pc_tree, model); + coeff = vp56_coeff_bias[idx]; + for (i=vp56_coeff_bit_length[idx]; i>=0; i--) + coeff += vp56_rac_get_prob(c, vp56_coeff_parse_table[idx][i]) << i; + } else { + if (vp56_rac_get_prob(c, model2[4])) + coeff = 3 + vp56_rac_get_prob(c, model[5]); + else + coeff = 2; + } + ct = 2; + } else { + ct = 1; + coeff = 1; + } + sign = vp56_rac_get(c); + coeff = (coeff ^ -sign) + sign; + if (coeff_idx) + coeff *= s->dequant_ac; + idx = s->coeff_index_to_pos[coeff_idx]; + s->block_coeff[b][permute[idx]] = coeff; + run = 1; + } else { + /* parse a run */ + ct = 0; + if (coeff_idx == 0) { + s->left_block[vp56_b6to4[b]].not_null_dc = 0; + s->above_blocks[s->above_block_idx[b]].not_null_dc = 0; + } else { + if (!vp56_rac_get_prob(c, model2[1])) + break; + + model3 = s->coeff_model_runv[coeff_idx >= 6]; + run = vp56_rac_get_tree(c, vp6_pcr_tree, model3); + if (!run) + for (run=9, i=0; i<6; i++) + run += vp56_rac_get_prob(c, model3[i+8]) << i; + } + } + + cg = vp6_coeff_groups[coeff_idx+=run]; + model = model2 = s->coeff_model_ract[pt][ct][cg]; + } + } +} + +static int vp6_adjust(int v, int t) +{ + int V = v, s = v >> 31; + V ^= s; + V -= s; + if (V-t-1 >= (unsigned)(t-1)) + return v; + V = 2*t - V; + V += s; + V ^= s; + return V; +} + +static int vp6_block_variance(uint8_t *src, int stride) +{ + int sum = 0, square_sum = 0; + int y, x; + + for (y=0; y<8; y+=2) { + for (x=0; x<8; x+=2) { + sum += src[x]; + square_sum += src[x]*src[x]; + } + src += 2*stride; + } + return (16*square_sum - sum*sum) >> 8; +} + +static void vp6_filter_hv2(vp56_context_t *s, uint8_t *dst, uint8_t *src, + int stride, int delta, int16_t weight) +{ + s->dsp.put_pixels_tab[1][0](dst, src, stride, 8); + s->dsp.biweight_h264_pixels_tab[3](dst, src+delta, stride, 2, + 8-weight, weight, 0); +} + +static void vp6_filter_hv4(uint8_t *dst, uint8_t *src, int stride, + int delta, const int16_t *weights) +{ + int x, y; + + for (y=0; y<8; y++) { + for (x=0; x<8; x++) { + dst[x] = clip_uint8(( src[x-delta ] * weights[0] + + src[x ] * weights[1] + + src[x+delta ] * weights[2] + + src[x+2*delta] * weights[3] + 64) >> 7); + } + src += stride; + dst += stride; + } +} + +static void vp6_filter_diag2(vp56_context_t *s, uint8_t *dst, uint8_t *src, + int stride, int h_weight, int v_weight) +{ + uint8_t *tmp = s->edge_emu_buffer+16; + int x, xmax; + + s->dsp.put_pixels_tab[1][0](tmp, src, stride, 8); + s->dsp.biweight_h264_pixels_tab[3](tmp, src+1, stride, 2, + 8-h_weight, h_weight, 0); + /* we need a 8x9 block to do vertical filter, so compute one more line */ + for (x=8*stride, xmax=x+8; x<xmax; x++) + tmp[x] = (src[x]*(8-h_weight) + src[x+1]*h_weight + 4) >> 3; + + s->dsp.put_pixels_tab[1][0](dst, tmp, stride, 8); + s->dsp.biweight_h264_pixels_tab[3](dst, tmp+stride, stride, 2, + 8-v_weight, v_weight, 0); +} + +static void vp6_filter_diag4(uint8_t *dst, uint8_t *src, int stride, + const int16_t *h_weights,const int16_t *v_weights) +{ + int x, y; + int tmp[8*11]; + int *t = tmp; + + src -= stride; + + for (y=0; y<11; y++) { + for (x=0; x<8; x++) { + t[x] = clip_uint8(( src[x-1] * h_weights[0] + + src[x ] * h_weights[1] + + src[x+1] * h_weights[2] + + src[x+2] * h_weights[3] + 64) >> 7); + } + src += stride; + t += 8; + } + + t = tmp + 8; + for (y=0; y<8; y++) { + for (x=0; x<8; x++) { + dst[x] = clip_uint8(( t[x-8 ] * v_weights[0] + + t[x ] * v_weights[1] + + t[x+8 ] * v_weights[2] + + t[x+16] * v_weights[3] + 64) >> 7); + } + dst += stride; + t += 8; + } +} + +static void vp6_filter(vp56_context_t *s, uint8_t *dst, uint8_t *src, + int offset1, int offset2, int stride, + vp56_mv_t mv, int mask, int select, int luma) +{ + int filter4 = 0; + int x8 = mv.x & mask; + int y8 = mv.y & mask; + + if (luma) { + x8 *= 2; + y8 *= 2; + filter4 = s->filter_mode; + if (filter4 == 2) { + if (s->max_vector_length && + (FFABS(mv.x) > s->max_vector_length || + FFABS(mv.y) > s->max_vector_length)) { + filter4 = 0; + } else if (s->sample_variance_threshold + && (vp6_block_variance(src+offset1, stride) + < s->sample_variance_threshold)) { + filter4 = 0; + } + } + } + + if ((y8 && (offset2-offset1)*s->flip<0) || (!y8 && offset1 > offset2)) { + offset1 = offset2; + } + + if (filter4) { + if (!y8) { /* left or right combine */ + vp6_filter_hv4(dst, src+offset1, stride, 1, + vp6_block_copy_filter[select][x8]); + } else if (!x8) { /* above or below combine */ + vp6_filter_hv4(dst, src+offset1, stride, stride, + vp6_block_copy_filter[select][y8]); + } else if ((mv.x^mv.y) >> 31) { /* lower-left or upper-right combine */ + vp6_filter_diag4(dst, src+offset1-1, stride, + vp6_block_copy_filter[select][x8], + vp6_block_copy_filter[select][y8]); + } else { /* lower-right or upper-left combine */ + vp6_filter_diag4(dst, src+offset1, stride, + vp6_block_copy_filter[select][x8], + vp6_block_copy_filter[select][y8]); + } + } else { + if (!y8) { /* left or right combine */ + vp6_filter_hv2(s, dst, src+offset1, stride, 1, x8); + } else if (!x8) { /* above or below combine */ + vp6_filter_hv2(s, dst, src+offset1, stride, stride, y8); + } else if ((mv.x^mv.y) >> 31) { /* lower-left or upper-right combine */ + vp6_filter_diag2(s, dst, src+offset1-1, stride, x8, y8); + } else { /* lower-right or upper-left combine */ + vp6_filter_diag2(s, dst, src+offset1, stride, x8, y8); + } + } +} + +static int vp6_decode_init(AVCodecContext *avctx) +{ + vp56_context_t *s = avctx->priv_data; + + vp56_init(s, avctx, avctx->codec->id == CODEC_ID_VP6); + s->vp56_coord_div = vp6_coord_div; + s->parse_vector_adjustment = vp6_parse_vector_adjustment; + s->adjust = vp6_adjust; + s->filter = vp6_filter; + s->parse_coeff = vp6_parse_coeff; + s->default_models_init = vp6_default_models_init; + s->parse_vector_models = vp6_parse_vector_models; + s->parse_coeff_models = vp6_parse_coeff_models; + s->parse_header = vp6_parse_header; + + return 0; +} + +AVCodec vp6_decoder = { + "vp6", + CODEC_TYPE_VIDEO, + CODEC_ID_VP6, + sizeof(vp56_context_t), + vp6_decode_init, + NULL, + vp56_free, + vp56_decode_frame, +}; + +/* flash version, not flipped upside-down */ +AVCodec vp6f_decoder = { + "vp6f", + CODEC_TYPE_VIDEO, + CODEC_ID_VP6F, + sizeof(vp56_context_t), + vp6_decode_init, + NULL, + vp56_free, + vp56_decode_frame, +}; diff --git a/src/libffmpeg/libavcodec/vp6data.h b/src/libffmpeg/libavcodec/vp6data.h new file mode 100644 index 000000000..0545a9d66 --- /dev/null +++ b/src/libffmpeg/libavcodec/vp6data.h @@ -0,0 +1,300 @@ +/** + * @file vp6data.h + * VP6 compatible video decoder + * + * Copyright (C) 2006 Aurelien Jacobs <aurel@gnuage.org> + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef VP6DATA_H +#define VP6DATA_H + +#include "vp56data.h" + +static const uint8_t vp6_def_fdv_vector_model[2][8] = { + { 247, 210, 135, 68, 138, 220, 239, 246 }, + { 244, 184, 201, 44, 173, 221, 239, 253 }, +}; + +static const uint8_t vp6_def_pdv_vector_model[2][7] = { + { 225, 146, 172, 147, 214, 39, 156 }, + { 204, 170, 119, 235, 140, 230, 228 }, +}; + +static const uint8_t vp6_def_coeff_reorder[] = { + 0, 0, 1, 1, 1, 2, 2, 2, + 2, 2, 2, 3, 3, 4, 4, 4, + 5, 5, 5, 5, 6, 6, 7, 7, + 7, 7, 7, 8, 8, 9, 9, 9, + 9, 9, 9, 10, 10, 11, 11, 11, + 11, 11, 11, 12, 12, 12, 12, 12, + 12, 13, 13, 13, 13, 13, 14, 14, + 14, 14, 15, 15, 15, 15, 15, 15, +}; + +static const uint8_t vp6_def_runv_coeff_model[2][14] = { + { 198, 197, 196, 146, 198, 204, 169, 142, 130, 136, 149, 149, 191, 249 }, + { 135, 201, 181, 154, 98, 117, 132, 126, 146, 169, 184, 240, 246, 254 }, +}; + +static const uint8_t vp6_sig_dct_pct[2][2] = { + { 237, 246 }, + { 231, 243 }, +}; + +static const uint8_t vp6_pdv_pct[2][7] = { + { 253, 253, 254, 254, 254, 254, 254 }, + { 245, 253, 254, 254, 254, 254, 254 }, +}; + +static const uint8_t vp6_fdv_pct[2][8] = { + { 254, 254, 254, 254, 254, 250, 250, 252 }, + { 254, 254, 254, 254, 254, 251, 251, 254 }, +}; + +static const uint8_t vp6_dccv_pct[2][11] = { + { 146, 255, 181, 207, 232, 243, 238, 251, 244, 250, 249 }, + { 179, 255, 214, 240, 250, 255, 244, 255, 255, 255, 255 }, +}; + +static const uint8_t vp6_coeff_reorder_pct[] = { + 255, 132, 132, 159, 153, 151, 161, 170, + 164, 162, 136, 110, 103, 114, 129, 118, + 124, 125, 132, 136, 114, 110, 142, 135, + 134, 123, 143, 126, 153, 183, 166, 161, + 171, 180, 179, 164, 203, 218, 225, 217, + 215, 206, 203, 217, 229, 241, 248, 243, + 253, 255, 253, 255, 255, 255, 255, 255, + 255, 255, 255, 255, 255, 255, 255, 255, +}; + +static const uint8_t vp6_runv_pct[2][14] = { + { 219, 246, 238, 249, 232, 239, 249, 255, 248, 253, 239, 244, 241, 248 }, + { 198, 232, 251, 253, 219, 241, 253, 255, 248, 249, 244, 238, 251, 255 }, +}; + +static const uint8_t vp6_ract_pct[3][2][6][11] = { + { { { 227, 246, 230, 247, 244, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 209, 231, 231, 249, 249, 253, 255, 255, 255 }, + { 255, 255, 225, 242, 241, 251, 253, 255, 255, 255, 255 }, + { 255, 255, 241, 253, 252, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 248, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } }, + { { 240, 255, 248, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 240, 253, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } } }, + { { { 206, 203, 227, 239, 247, 255, 253, 255, 255, 255, 255 }, + { 207, 199, 220, 236, 243, 252, 252, 255, 255, 255, 255 }, + { 212, 219, 230, 243, 244, 253, 252, 255, 255, 255, 255 }, + { 236, 237, 247, 252, 253, 255, 255, 255, 255, 255, 255 }, + { 240, 240, 248, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } }, + { { 230, 233, 249, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 238, 238, 250, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 248, 251, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } } }, + { { { 225, 239, 227, 231, 244, 253, 243, 255, 255, 253, 255 }, + { 232, 234, 224, 228, 242, 249, 242, 252, 251, 251, 255 }, + { 235, 249, 238, 240, 251, 255, 249, 255, 253, 253, 255 }, + { 249, 253, 251, 250, 255, 255, 255, 255, 255, 255, 255 }, + { 251, 250, 249, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } }, + { { 243, 244, 250, 250, 255, 255, 255, 255, 255, 255, 255 }, + { 249, 248, 250, 253, 255, 255, 255, 255, 255, 255, 255 }, + { 253, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 }, + { 255, 255, 255, 255, 255, 255, 255, 255, 255, 255, 255 } } } +}; + +static const int vp6_dccv_lc[3][5][2] = { + { { 122, 133 }, { 0, 1 }, { 78, 171 }, { 139, 117 }, { 168, 79 } }, + { { 133, 51 }, { 0, 1 }, { 169, 71 }, { 214, 44 }, { 210, 38 } }, + { { 142, -16 }, { 0, 1 }, { 221, -30 }, { 246, -3 }, { 203, 17 } }, +}; + +static const uint8_t vp6_coeff_groups[] = { + 0, 0, 1, 1, 1, 2, 2, 2, + 2, 2, 2, 3, 3, 3, 3, 3, + 3, 3, 3, 3, 3, 3, 4, 4, + 4, 4, 4, 4, 4, 4, 4, 4, + 4, 4, 4, 4, 4, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, + 5, 5, 5, 5, 5, 5, 5, 5, +}; + +static const int16_t vp6_block_copy_filter[17][8][4] = { + { { 0, 128, 0, 0 }, /* 0 */ + { -3, 122, 9, 0 }, + { -4, 109, 24, -1 }, + { -5, 91, 45, -3 }, + { -4, 68, 68, -4 }, + { -3, 45, 91, -5 }, + { -1, 24, 109, -4 }, + { 0, 9, 122, -3 } }, + { { 0, 128, 0, 0 }, /* 1 */ + { -4, 124, 9, -1 }, + { -5, 110, 25, -2 }, + { -6, 91, 46, -3 }, + { -5, 69, 69, -5 }, + { -3, 46, 91, -6 }, + { -2, 25, 110, -5 }, + { -1, 9, 124, -4 } }, + { { 0, 128, 0, 0 }, /* 2 */ + { -4, 123, 10, -1 }, + { -6, 110, 26, -2 }, + { -7, 92, 47, -4 }, + { -6, 70, 70, -6 }, + { -4, 47, 92, -7 }, + { -2, 26, 110, -6 }, + { -1, 10, 123, -4 } }, + { { 0, 128, 0, 0 }, /* 3 */ + { -5, 124, 10, -1 }, + { -7, 110, 27, -2 }, + { -7, 91, 48, -4 }, + { -6, 70, 70, -6 }, + { -4, 48, 92, -8 }, + { -2, 27, 110, -7 }, + { -1, 10, 124, -5 } }, + { { 0, 128, 0, 0 }, /* 4 */ + { -6, 124, 11, -1 }, + { -8, 111, 28, -3 }, + { -8, 92, 49, -5 }, + { -7, 71, 71, -7 }, + { -5, 49, 92, -8 }, + { -3, 28, 111, -8 }, + { -1, 11, 124, -6 } }, + { { 0, 128, 0, 0 }, /* 5 */ + { -6, 123, 12, -1 }, + { -9, 111, 29, -3 }, + { -9, 93, 50, -6 }, + { -8, 72, 72, -8 }, + { -6, 50, 93, -9 }, + { -3, 29, 111, -9 }, + { -1, 12, 123, -6 } }, + { { 0, 128, 0, 0 }, /* 6 */ + { -7, 124, 12, -1 }, + { -10, 111, 30, -3 }, + { -10, 93, 51, -6 }, + { -9, 73, 73, -9 }, + { -6, 51, 93, -10 }, + { -3, 30, 111, -10 }, + { -1, 12, 124, -7 } }, + { { 0, 128, 0, 0 }, /* 7 */ + { -7, 123, 13, -1 }, + { -11, 112, 31, -4 }, + { -11, 94, 52, -7 }, + { -10, 74, 74, -10 }, + { -7, 52, 94, -11 }, + { -4, 31, 112, -11 }, + { -1, 13, 123, -7 } }, + { { 0, 128, 0, 0 }, /* 8 */ + { -8, 124, 13, -1 }, + { -12, 112, 32, -4 }, + { -12, 94, 53, -7 }, + { -10, 74, 74, -10 }, + { -7, 53, 94, -12 }, + { -4, 32, 112, -12 }, + { -1, 13, 124, -8 } }, + { { 0, 128, 0, 0 }, /* 9 */ + { -9, 124, 14, -1 }, + { -13, 112, 33, -4 }, + { -13, 95, 54, -8 }, + { -11, 75, 75, -11 }, + { -8, 54, 95, -13 }, + { -4, 33, 112, -13 }, + { -1, 14, 124, -9 } }, + { { 0, 128, 0, 0 }, /* 10 */ + { -9, 123, 15, -1 }, + { -14, 113, 34, -5 }, + { -14, 95, 55, -8 }, + { -12, 76, 76, -12 }, + { -8, 55, 95, -14 }, + { -5, 34, 112, -13 }, + { -1, 15, 123, -9 } }, + { { 0, 128, 0, 0 }, /* 11 */ + { -10, 124, 15, -1 }, + { -14, 113, 34, -5 }, + { -15, 96, 56, -9 }, + { -13, 77, 77, -13 }, + { -9, 56, 96, -15 }, + { -5, 34, 113, -14 }, + { -1, 15, 124, -10 } }, + { { 0, 128, 0, 0 }, /* 12 */ + { -10, 123, 16, -1 }, + { -15, 113, 35, -5 }, + { -16, 98, 56, -10 }, + { -14, 78, 78, -14 }, + { -10, 56, 98, -16 }, + { -5, 35, 113, -15 }, + { -1, 16, 123, -10 } }, + { { 0, 128, 0, 0 }, /* 13 */ + { -11, 124, 17, -2 }, + { -16, 113, 36, -5 }, + { -17, 98, 57, -10 }, + { -14, 78, 78, -14 }, + { -10, 57, 98, -17 }, + { -5, 36, 113, -16 }, + { -2, 17, 124, -11 } }, + { { 0, 128, 0, 0 }, /* 14 */ + { -12, 125, 17, -2 }, + { -17, 114, 37, -6 }, + { -18, 99, 58, -11 }, + { -15, 79, 79, -15 }, + { -11, 58, 99, -18 }, + { -6, 37, 114, -17 }, + { -2, 17, 125, -12 } }, + { { 0, 128, 0, 0 }, /* 15 */ + { -12, 124, 18, -2 }, + { -18, 114, 38, -6 }, + { -19, 99, 59, -11 }, + { -16, 80, 80, -16 }, + { -11, 59, 99, -19 }, + { -6, 38, 114, -18 }, + { -2, 18, 124, -12 } }, + { { 0, 128, 0, 0 }, /* 16 */ + { -4, 118, 16, -2 }, + { -7, 106, 34, -5 }, + { -8, 90, 53, -7 }, + { -8, 72, 72, -8 }, + { -7, 53, 90, -8 }, + { -5, 34, 106, -7 }, + { -2, 16, 118, -4 } }, +}; + +static const vp56_tree_t vp6_pcr_tree[] = { + { 8, 0}, + { 4, 1}, + { 2, 2}, {-1}, {-2}, + { 2, 3}, {-3}, {-4}, + { 8, 4}, + { 4, 5}, + { 2, 6}, {-5}, {-6}, + { 2, 7}, {-7}, {-8}, + {-0}, +}; + +static const uint8_t vp6_coord_div[] = { 4, 4, 4, 4, 8, 8 }; + +#endif /* VP6DATA_H */ diff --git a/src/libffmpeg/libavcodec/wavpack.c b/src/libffmpeg/libavcodec/wavpack.c new file mode 100644 index 000000000..b462174da --- /dev/null +++ b/src/libffmpeg/libavcodec/wavpack.c @@ -0,0 +1,556 @@ +/* + * WavPack lossless audio decoder + * Copyright (c) 2006 Konstantin Shishkov + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ +#define ALT_BITSTREAM_READER_LE +#include "avcodec.h" +#include "bitstream.h" + +/** + * @file wavpack.c + * WavPack lossless audio decoder + */ + +#define WV_JOINT 0x0010 + +enum WP_ID_Flags{ + WP_IDF_MASK = 0x1F, + WP_IDF_IGNORE = 0x20, + WP_IDF_ODD = 0x40, + WP_IDF_LONG = 0x80 +}; + +enum WP_ID{ + WP_ID_DUMMY = 0, + WP_ID_ENCINFO, + WP_ID_DECTERMS, + WP_ID_DECWEIGHTS, + WP_ID_DECSAMPLES, + WP_ID_ENTROPY, + WP_ID_HYBRID, + WP_ID_SHAPING, + WP_ID_FLOATINFO, + WP_ID_INT32INFO, + WP_ID_DATA, + WP_ID_CORR, + WP_ID_FLT, + WP_ID_CHANINFO +}; + +#define MAX_TERMS 16 + +typedef struct Decorr { + int delta; + int value; + int weightA; + int weightB; + int samplesA[8]; + int samplesB[8]; +} Decorr; + +typedef struct WavpackContext { + AVCodecContext *avctx; + int stereo; + int joint; + uint32_t CRC; + GetBitContext gb; + int data_size; // in bits + int samples; + int median[6]; + int terms; + Decorr decorr[MAX_TERMS]; + int zero, one, zeroes; +} WavpackContext; + +// exponent table copied from WavPack source +static const uint8_t wp_exp2_table [256] = { + 0x00, 0x01, 0x01, 0x02, 0x03, 0x03, 0x04, 0x05, 0x06, 0x06, 0x07, 0x08, 0x08, 0x09, 0x0a, 0x0b, + 0x0b, 0x0c, 0x0d, 0x0e, 0x0e, 0x0f, 0x10, 0x10, 0x11, 0x12, 0x13, 0x13, 0x14, 0x15, 0x16, 0x16, + 0x17, 0x18, 0x19, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1d, 0x1e, 0x1f, 0x20, 0x20, 0x21, 0x22, 0x23, + 0x24, 0x24, 0x25, 0x26, 0x27, 0x28, 0x28, 0x29, 0x2a, 0x2b, 0x2c, 0x2c, 0x2d, 0x2e, 0x2f, 0x30, + 0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3a, 0x3b, 0x3c, 0x3d, + 0x3e, 0x3f, 0x40, 0x41, 0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x48, 0x49, 0x4a, 0x4b, + 0x4c, 0x4d, 0x4e, 0x4f, 0x50, 0x51, 0x51, 0x52, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, 0x5a, + 0x5b, 0x5c, 0x5d, 0x5e, 0x5e, 0x5f, 0x60, 0x61, 0x62, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, + 0x6a, 0x6b, 0x6c, 0x6d, 0x6e, 0x6f, 0x70, 0x71, 0x72, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, + 0x7a, 0x7b, 0x7c, 0x7d, 0x7e, 0x7f, 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x87, 0x88, 0x89, 0x8a, + 0x8b, 0x8c, 0x8d, 0x8e, 0x8f, 0x90, 0x91, 0x92, 0x93, 0x95, 0x96, 0x97, 0x98, 0x99, 0x9a, 0x9b, + 0x9c, 0x9d, 0x9f, 0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, + 0xaf, 0xb0, 0xb1, 0xb2, 0xb3, 0xb4, 0xb6, 0xb7, 0xb8, 0xb9, 0xba, 0xbc, 0xbd, 0xbe, 0xbf, 0xc0, + 0xc2, 0xc3, 0xc4, 0xc5, 0xc6, 0xc8, 0xc9, 0xca, 0xcb, 0xcd, 0xce, 0xcf, 0xd0, 0xd2, 0xd3, 0xd4, + 0xd6, 0xd7, 0xd8, 0xd9, 0xdb, 0xdc, 0xdd, 0xde, 0xe0, 0xe1, 0xe2, 0xe4, 0xe5, 0xe6, 0xe8, 0xe9, + 0xea, 0xec, 0xed, 0xee, 0xf0, 0xf1, 0xf2, 0xf4, 0xf5, 0xf6, 0xf8, 0xf9, 0xfa, 0xfc, 0xfd, 0xff +}; + +static av_always_inline int wp_exp2(int16_t val) +{ + int res, neg = 0; + + if(val < 0){ + val = -val; + neg = 1; + } + + res = wp_exp2_table[val & 0xFF] | 0x100; + val >>= 8; + res = (val > 9) ? (res << (val - 9)) : (res >> (9 - val)); + return neg ? -res : res; +} + +static inline int get_unary(GetBitContext *gb){ + int r=0; + while(get_bits1(gb) && r<33)r++; + return r; +} + +// macros for manipulating median values +#define GET_MED(n) ((median[n] >> 4) + 1) +#define DEC_MED(n) median[n] -= ((median[n] + (128>>n) - 2) / (128>>n)) * 2 +#define INC_MED(n) median[n] += ((median[n] + (128>>n)) / (128>>n)) * 5 + +// macros for applying weight +#define UPDATE_WEIGHT_CLIP(weight, delta, samples, in) \ + if(samples && in){ \ + if((samples ^ in) < 0){ \ + weight -= delta; \ + if(weight < -1024) weight = -1024; \ + }else{ \ + weight += delta; \ + if(weight > 1024) weight = 1024; \ + } \ + } + + +static av_always_inline int get_tail(GetBitContext *gb, int k) +{ + int p, e, res; + + if(k<1)return 0; + p = av_log2(k); + e = (1 << (p + 1)) - k - 1; + res = p ? get_bits(gb, p) : 0; + if(res >= e){ + res = (res<<1) - e + get_bits1(gb); + } + return res; +} + +static int wv_get_value(WavpackContext *ctx, GetBitContext *gb, int *median, int *last) +{ + int t, t2; + int sign, base, add, ret; + + *last = 0; + + if((ctx->median[0] < 2U) && (ctx->median[3] < 2U) && !ctx->zero && !ctx->one){ + if(ctx->zeroes){ + ctx->zeroes--; + if(ctx->zeroes) + return 0; + }else{ + t = get_unary(gb); + if(t >= 2) t = get_bits(gb, t - 1) | (1 << (t-1)); + ctx->zeroes = t; + if(ctx->zeroes){ + memset(ctx->median, 0, sizeof(ctx->median)); + return 0; + } + } + } + + if(get_bits_count(gb) >= ctx->data_size){ + *last = 1; + return 0; + } + + if(ctx->zero){ + t = 0; + ctx->zero = 0; + }else{ + t = get_unary(gb); + if(get_bits_count(gb) >= ctx->data_size){ + *last = 1; + return 0; + } + if(t == 16) { + t2 = get_unary(gb); + if(t2 < 2) t += t2; + else t += get_bits(gb, t2 - 1) | (1 << (t2 - 1)); + } + + if(ctx->one){ + ctx->one = t&1; + t = (t>>1) + 1; + }else{ + ctx->one = t&1; + t >>= 1; + } + ctx->zero = !ctx->one; + } + + if(!t){ + base = 0; + add = GET_MED(0) - 1; + DEC_MED(0); + }else if(t == 1){ + base = GET_MED(0); + add = GET_MED(1) - 1; + INC_MED(0); + DEC_MED(1); + }else if(t == 2){ + base = GET_MED(0) + GET_MED(1); + add = GET_MED(2) - 1; + INC_MED(0); + INC_MED(1); + DEC_MED(2); + }else{ + base = GET_MED(0) + GET_MED(1) + GET_MED(2) * (t - 2); + add = GET_MED(2) - 1; + INC_MED(0); + INC_MED(1); + INC_MED(2); + } + ret = base + get_tail(gb, add); + sign = get_bits1(gb); + return sign ? ~ret : ret; +} + +static int wv_unpack_stereo(WavpackContext *s, GetBitContext *gb, int16_t *dst) +{ + int i, j, count = 0; + int last, t; + int A, B, L, L2, R, R2; + int pos = 0; + uint32_t crc = 0xFFFFFFFF; + + s->one = s->zero = s->zeroes = 0; + do{ + L = wv_get_value(s, gb, s->median, &last); + if(last) break; + R = wv_get_value(s, gb, s->median + 3, &last); + if(last) break; + for(i = 0; i < s->terms; i++){ + t = s->decorr[i].value; + j = 0; + if(t > 0){ + if(t > 8){ + if(t & 1){ + A = 2 * s->decorr[i].samplesA[0] - s->decorr[i].samplesA[1]; + B = 2 * s->decorr[i].samplesB[0] - s->decorr[i].samplesB[1]; + }else{ + A = (3 * s->decorr[i].samplesA[0] - s->decorr[i].samplesA[1]) >> 1; + B = (3 * s->decorr[i].samplesB[0] - s->decorr[i].samplesB[1]) >> 1; + } + s->decorr[i].samplesA[1] = s->decorr[i].samplesA[0]; + s->decorr[i].samplesB[1] = s->decorr[i].samplesB[0]; + j = 0; + }else{ + A = s->decorr[i].samplesA[pos]; + B = s->decorr[i].samplesB[pos]; + j = (pos + t) & 7; + } + L2 = L + ((s->decorr[i].weightA * A + 512) >> 10); + R2 = R + ((s->decorr[i].weightB * B + 512) >> 10); + if(A && L) s->decorr[i].weightA -= ((((L ^ A) >> 30) & 2) - 1) * s->decorr[i].delta; + if(B && R) s->decorr[i].weightB -= ((((R ^ B) >> 30) & 2) - 1) * s->decorr[i].delta; + s->decorr[i].samplesA[j] = L = L2; + s->decorr[i].samplesB[j] = R = R2; + }else if(t == -1){ + L2 = L + ((s->decorr[i].weightA * s->decorr[i].samplesA[0] + 512) >> 10); + UPDATE_WEIGHT_CLIP(s->decorr[i].weightA, s->decorr[i].delta, s->decorr[i].samplesA[0], L); + L = L2; + R2 = R + ((s->decorr[i].weightB * L2 + 512) >> 10); + UPDATE_WEIGHT_CLIP(s->decorr[i].weightB, s->decorr[i].delta, L2, R); + R = R2; + s->decorr[i].samplesA[0] = R; + }else{ + R2 = R + ((s->decorr[i].weightB * s->decorr[i].samplesB[0] + 512) >> 10); + UPDATE_WEIGHT_CLIP(s->decorr[i].weightB, s->decorr[i].delta, s->decorr[i].samplesB[0], R); + R = R2; + + if(t == -3){ + R2 = s->decorr[i].samplesA[0]; + s->decorr[i].samplesA[0] = R; + } + + L2 = L + ((s->decorr[i].weightA * R2 + 512) >> 10); + UPDATE_WEIGHT_CLIP(s->decorr[i].weightA, s->decorr[i].delta, R2, L); + L = L2; + s->decorr[i].samplesB[0] = L; + } + } + pos = (pos + 1) & 7; + if(s->joint) + L += (R -= (L >> 1)); + crc = (crc * 3 + L) * 3 + R; + *dst++ = L; + *dst++ = R; + + count++; + }while(!last && count < s->samples); + + if(crc != s->CRC){ + av_log(s->avctx, AV_LOG_ERROR, "CRC error\n"); + return -1; + } + return count * 2; +} + +static int wv_unpack_mono(WavpackContext *s, GetBitContext *gb, int16_t *dst) +{ + int i, j, count = 0; + int last, t; + int A, S, T; + int pos = 0; + uint32_t crc = 0xFFFFFFFF; + + s->one = s->zero = s->zeroes = 0; + do{ + T = wv_get_value(s, gb, s->median, &last); + S = 0; + if(last) break; + for(i = 0; i < s->terms; i++){ + t = s->decorr[i].value; + if(t > 8){ + if(t & 1) + A = 2 * s->decorr[i].samplesA[0] - s->decorr[i].samplesA[1]; + else + A = (3 * s->decorr[i].samplesA[0] - s->decorr[i].samplesA[1]) >> 1; + s->decorr[i].samplesA[1] = s->decorr[i].samplesA[0]; + j = 0; + }else{ + A = s->decorr[i].samplesA[pos]; + j = (pos + t) & 7; + } + S = T + ((s->decorr[i].weightA * A + 512) >> 10); + if(A && T) s->decorr[i].weightA -= ((((T ^ A) >> 30) & 2) - 1) * s->decorr[i].delta; + s->decorr[i].samplesA[j] = T = S; + } + pos = (pos + 1) & 7; + crc = crc * 3 + S; + *dst++ = S; + count++; + }while(!last && count < s->samples); + + if(crc != s->CRC){ + av_log(s->avctx, AV_LOG_ERROR, "CRC error\n"); + return -1; + } + return count; +} + +static int wavpack_decode_init(AVCodecContext *avctx) +{ + WavpackContext *s = avctx->priv_data; + + s->avctx = avctx; + s->stereo = (avctx->channels == 2); + + return 0; +} + +static int wavpack_decode_close(AVCodecContext *avctx) +{ +// WavpackContext *s = avctx->priv_data; + + return 0; +} + +static int wavpack_decode_frame(AVCodecContext *avctx, + void *data, int *data_size, + uint8_t *buf, int buf_size) +{ + WavpackContext *s = avctx->priv_data; + int16_t *samples = data; + int samplecount; + int got_terms = 0, got_weights = 0, got_samples = 0, got_entropy = 0, got_bs = 0; + uint8_t* buf_end = buf + buf_size; + int i, j, id, size, ssize, weights, t; + + if (buf_size == 0) return 0; + + memset(s->decorr, 0, MAX_TERMS * sizeof(Decorr)); + + s->samples = AV_RL32(buf); buf += 4; + if(!s->samples) return buf_size; + /* should not happen but who knows */ + if(s->samples * 2 * avctx->channels > AVCODEC_MAX_AUDIO_FRAME_SIZE){ + av_log(avctx, AV_LOG_ERROR, "Packet size is too big to be handled in lavc!\n"); + return -1; + } + s->joint = AV_RL32(buf) & WV_JOINT; buf += 4; + s->CRC = AV_RL32(buf); buf += 4; + // parse metadata blocks + while(buf < buf_end){ + id = *buf++; + size = *buf++; + if(id & WP_IDF_LONG) { + size |= (*buf++) << 8; + size |= (*buf++) << 16; + } + size <<= 1; // size is specified in words + ssize = size; + if(id & WP_IDF_ODD) size--; + if(size < 0){ + av_log(avctx, AV_LOG_ERROR, "Got incorrect block %02X with size %i\n", id, size); + break; + } + if(buf + ssize > buf_end){ + av_log(avctx, AV_LOG_ERROR, "Block size %i is out of bounds\n", size); + break; + } + if(id & WP_IDF_IGNORE){ + buf += ssize; + continue; + } + switch(id & WP_IDF_MASK){ + case WP_ID_DECTERMS: + s->terms = size; + if(s->terms > MAX_TERMS){ + av_log(avctx, AV_LOG_ERROR, "Too many decorrelation terms\n"); + buf += ssize; + continue; + } + for(i = 0; i < s->terms; i++) { + s->decorr[s->terms - i - 1].value = (*buf & 0x1F) - 5; + s->decorr[s->terms - i - 1].delta = *buf >> 5; + buf++; + } + got_terms = 1; + break; + case WP_ID_DECWEIGHTS: + if(!got_terms){ + av_log(avctx, AV_LOG_ERROR, "No decorrelation terms met\n"); + continue; + } + weights = size >> s->stereo; + if(weights > MAX_TERMS || weights > s->terms){ + av_log(avctx, AV_LOG_ERROR, "Too many decorrelation weights\n"); + buf += ssize; + continue; + } + for(i = 0; i < weights; i++) { + t = (int8_t)(*buf++); + s->decorr[s->terms - i - 1].weightA = t << 3; + if(s->decorr[s->terms - i - 1].weightA > 0) + s->decorr[s->terms - i - 1].weightA += (s->decorr[s->terms - i - 1].weightA + 64) >> 7; + if(s->stereo){ + t = (int8_t)(*buf++); + s->decorr[s->terms - i - 1].weightB = t << 3; + if(s->decorr[s->terms - i - 1].weightB > 0) + s->decorr[s->terms - i - 1].weightB += (s->decorr[s->terms - i - 1].weightB + 64) >> 7; + } + } + got_weights = 1; + break; + case WP_ID_DECSAMPLES: + if(!got_terms){ + av_log(avctx, AV_LOG_ERROR, "No decorrelation terms met\n"); + continue; + } + t = 0; + for(i = s->terms - 1; (i >= 0) && (t < size); i--) { + if(s->decorr[i].value > 8){ + s->decorr[i].samplesA[0] = wp_exp2(AV_RL16(buf)); buf += 2; + s->decorr[i].samplesA[1] = wp_exp2(AV_RL16(buf)); buf += 2; + if(s->stereo){ + s->decorr[i].samplesB[0] = wp_exp2(AV_RL16(buf)); buf += 2; + s->decorr[i].samplesB[1] = wp_exp2(AV_RL16(buf)); buf += 2; + t += 4; + } + t += 4; + }else if(s->decorr[i].value < 0){ + s->decorr[i].samplesA[0] = wp_exp2(AV_RL16(buf)); buf += 2; + s->decorr[i].samplesB[0] = wp_exp2(AV_RL16(buf)); buf += 2; + t += 4; + }else{ + for(j = 0; j < s->decorr[i].value; j++){ + s->decorr[i].samplesA[j] = wp_exp2(AV_RL16(buf)); buf += 2; + if(s->stereo){ + s->decorr[i].samplesB[j] = wp_exp2(AV_RL16(buf)); buf += 2; + } + } + t += s->decorr[i].value * 2 * avctx->channels; + } + } + got_samples = 1; + break; + case WP_ID_ENTROPY: + if(size != 6 * avctx->channels){ + av_log(avctx, AV_LOG_ERROR, "Entropy vars size should be %i, got %i", 6 * avctx->channels, size); + buf += ssize; + continue; + } + for(i = 0; i < 3 * avctx->channels; i++){ + s->median[i] = wp_exp2(AV_RL16(buf)); + buf += 2; + } + got_entropy = 1; + break; + case WP_ID_DATA: + init_get_bits(&s->gb, buf, size * 8); + s->data_size = size * 8; + buf += size; + got_bs = 1; + break; + default: + buf += size; + } + if(id & WP_IDF_ODD) buf++; + } + if(!got_terms){ + av_log(avctx, AV_LOG_ERROR, "No block with decorrelation terms\n"); + return -1; + } + if(!got_weights){ + av_log(avctx, AV_LOG_ERROR, "No block with decorrelation weights\n"); + return -1; + } + if(!got_samples){ + av_log(avctx, AV_LOG_ERROR, "No block with decorrelation samples\n"); + return -1; + } + if(!got_entropy){ + av_log(avctx, AV_LOG_ERROR, "No block with entropy info\n"); + return -1; + } + if(!got_bs){ + av_log(avctx, AV_LOG_ERROR, "Packed samples not found\n"); + return -1; + } + + if(s->stereo) + samplecount = wv_unpack_stereo(s, &s->gb, samples); + else + samplecount = wv_unpack_mono(s, &s->gb, samples); + *data_size = samplecount * 2; + + return buf_size; +} + +AVCodec wavpack_decoder = { + "wavpack", + CODEC_TYPE_AUDIO, + CODEC_ID_WAVPACK, + sizeof(WavpackContext), + wavpack_decode_init, + NULL, + wavpack_decode_close, + wavpack_decode_frame, +}; diff --git a/src/libffmpeg/libavutil/.hgignore b/src/libffmpeg/libavutil/.hgignore new file mode 100644 index 000000000..7d926a554 --- /dev/null +++ b/src/libffmpeg/libavutil/.hgignore @@ -0,0 +1,6 @@ +Makefile +Makefile.in +.libs +.deps +*.lo +*.la diff --git a/src/libffmpeg/libavutil/intreadwrite.h b/src/libffmpeg/libavutil/intreadwrite.h new file mode 100644 index 000000000..e6db5ce6f --- /dev/null +++ b/src/libffmpeg/libavutil/intreadwrite.h @@ -0,0 +1,65 @@ +#ifndef INTREADWRITE_H +#define INTREADWRITE_H + +#ifdef __GNUC__ + +struct unaligned_64 { uint64_t l; } __attribute__((packed)); +struct unaligned_32 { uint32_t l; } __attribute__((packed)); +struct unaligned_16 { uint16_t l; } __attribute__((packed)); + +#define LD16(a) (((const struct unaligned_16 *) (a))->l) +#define LD32(a) (((const struct unaligned_32 *) (a))->l) +#define LD64(a) (((const struct unaligned_64 *) (a))->l) + +#define ST16(a, b) (((struct unaligned_16 *) (a))->l) = (b) +#define ST32(a, b) (((struct unaligned_32 *) (a))->l) = (b) + +#else /* __GNUC__ */ + +#define LD16(a) (*((uint16_t*)(a))) +#define LD32(a) (*((uint32_t*)(a))) +#define LD64(a) (*((uint64_t*)(a))) + +#define ST16(a, b) *((uint16_t*)(a)) = (b) +#define ST32(a, b) *((uint32_t*)(a)) = (b) + +#endif /* !__GNUC__ */ + +/* endian macros */ +#define AV_RB8(x) (((uint8_t*)(x))[0]) +#define AV_WB8(p, i, d) { ((uint8_t*)(p))[(i)] = (d); } + +#define AV_RB16(x) ((((uint8_t*)(x))[0] << 8) | ((uint8_t*)(x))[1]) +#define AV_WB16(p, i, d) { \ + ((uint8_t*)(p))[(i)+1] = (d); \ + ((uint8_t*)(p))[(i)] = (d)>>8; } + +#define AV_RB32(x) ((((uint8_t*)(x))[0] << 24) | \ + (((uint8_t*)(x))[1] << 16) | \ + (((uint8_t*)(x))[2] << 8) | \ + ((uint8_t*)(x))[3]) +#define AV_WB32(p, i, d) { \ + ((uint8_t*)(p))[(i)+3] = (d); \ + ((uint8_t*)(p))[(i)+2] = (d)>>8; \ + ((uint8_t*)(p))[(i)+1] = (d)>>16; \ + ((uint8_t*)(p))[(i)] = (d)>>24; } + +#define AV_RL8(x) AV_RB8(x) +#define AV_WL8(p, i, d) AV_WB8(p, i, d) + +#define AV_RL16(x) ((((uint8_t*)(x))[1] << 8) | ((uint8_t*)(x))[0]) +#define AV_WL16(p, i, d) { \ + ((uint8_t*)(p))[(i)] = (d); \ + ((uint8_t*)(p))[(i)+1] = (d)>>8; } + +#define AV_RL32(x) ((((uint8_t*)(x))[3] << 24) | \ + (((uint8_t*)(x))[2] << 16) | \ + (((uint8_t*)(x))[1] << 8) | \ + ((uint8_t*)(x))[0]) +#define AV_WL32(p, i, d) { \ + ((uint8_t*)(p))[(i)] = (d); \ + ((uint8_t*)(p))[(i)+1] = (d)>>8; \ + ((uint8_t*)(p))[(i)+2] = (d)>>16; \ + ((uint8_t*)(p))[(i)+3] = (d)>>24; } + +#endif /* INTREADWRITE_H */ diff --git a/src/libffmpeg/libavutil/mem.c b/src/libffmpeg/libavutil/mem.c new file mode 100644 index 000000000..f43fb5420 --- /dev/null +++ b/src/libffmpeg/libavutil/mem.c @@ -0,0 +1,171 @@ +/* + * default memory allocator for libavutil + * Copyright (c) 2002 Fabrice Bellard. + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/** + * @file mem.c + * default memory allocator for libavutil. + */ + +#include "common.h" + +/* here we can use OS dependant allocation functions */ +#undef malloc +#undef free +#undef realloc + +#ifdef HAVE_MALLOC_H +#include <malloc.h> +#endif + +/* you can redefine av_malloc and av_free in your project to use your + memory allocator. You do not need to suppress this file because the + linker will do it automatically */ + +/** + * Memory allocation of size byte with alignment suitable for all + * memory accesses (including vectors if available on the + * CPU). av_malloc(0) must return a non NULL pointer. + */ +void *av_malloc(unsigned int size) +{ + void *ptr; +#ifdef CONFIG_MEMALIGN_HACK + long diff; +#endif + + /* let's disallow possible ambiguous cases */ + if(size > (INT_MAX-16) ) + return NULL; + +#ifdef CONFIG_MEMALIGN_HACK + ptr = malloc(size+16); + if(!ptr) + return ptr; + diff= ((-(long)ptr - 1)&15) + 1; + ptr += diff; + ((char*)ptr)[-1]= diff; +#elif defined (HAVE_MEMALIGN) + ptr = memalign(16,size); + /* Why 64? + Indeed, we should align it: + on 4 for 386 + on 16 for 486 + on 32 for 586, PPro - k6-III + on 64 for K7 (maybe for P3 too). + Because L1 and L2 caches are aligned on those values. + But I don't want to code such logic here! + */ + /* Why 16? + because some cpus need alignment, for example SSE2 on P4, & most RISC cpus + it will just trigger an exception and the unaligned load will be done in the + exception handler or it will just segfault (SSE2 on P4) + Why not larger? because i didnt see a difference in benchmarks ... + */ + /* benchmarks with p3 + memalign(64)+1 3071,3051,3032 + memalign(64)+2 3051,3032,3041 + memalign(64)+4 2911,2896,2915 + memalign(64)+8 2545,2554,2550 + memalign(64)+16 2543,2572,2563 + memalign(64)+32 2546,2545,2571 + memalign(64)+64 2570,2533,2558 + + btw, malloc seems to do 8 byte alignment by default here + */ +#else + ptr = malloc(size); +#endif + return ptr; +} + +/** + * av_realloc semantics (same as glibc): if ptr is NULL and size > 0, + * identical to malloc(size). If size is zero, it is identical to + * free(ptr) and NULL is returned. + */ +void *av_realloc(void *ptr, unsigned int size) +{ +#ifdef CONFIG_MEMALIGN_HACK + int diff; +#endif + + /* let's disallow possible ambiguous cases */ + if(size > (INT_MAX-16) ) + return NULL; + +#ifdef CONFIG_MEMALIGN_HACK + //FIXME this isn't aligned correctly, though it probably isn't needed + if(!ptr) return av_malloc(size); + diff= ((char*)ptr)[-1]; + return realloc(ptr - diff, size + diff) + diff; +#else + return realloc(ptr, size); +#endif +} + +/** + * Free memory which has been allocated with av_malloc(z)() or av_realloc(). + * NOTE: ptr = NULL is explicetly allowed + * Note2: it is recommended that you use av_freep() instead + */ +void av_free(void *ptr) +{ + /* XXX: this test should not be needed on most libcs */ + if (ptr) +#ifdef CONFIG_MEMALIGN_HACK + free(ptr - ((char*)ptr)[-1]); +#else + free(ptr); +#endif +} + +/** + * Frees memory and sets the pointer to NULL. + * @param arg pointer to the pointer which should be freed + */ +void av_freep(void *arg) +{ + void **ptr= (void**)arg; + av_free(*ptr); + *ptr = NULL; +} + +void *av_mallocz(unsigned int size) +{ + void *ptr; + + ptr = av_malloc(size); + if (ptr) + memset(ptr, 0, size); + return ptr; +} + +char *av_strdup(const char *s) +{ + char *ptr; + int len; + len = strlen(s) + 1; + ptr = av_malloc(len); + if (ptr) + memcpy(ptr, s, len); + return ptr; +} + diff --git a/src/libffmpeg/mpeg_parser.c b/src/libffmpeg/mpeg_parser.c index d9ee602b2..14a08a456 100644 --- a/src/libffmpeg/mpeg_parser.c +++ b/src/libffmpeg/mpeg_parser.c @@ -20,7 +20,7 @@ * Simple MPEG-ES parser/framer by Thibaut Mattern (tmattern@noos.fr) * based on libmpeg2 decoder. * - * $Id: mpeg_parser.c,v 1.4 2006/09/26 01:19:31 dgp85 Exp $ + * $Id: mpeg_parser.c,v 1.7 2007/03/29 18:52:45 dgp85 Exp $ */ #define LOG_MODULE "mpeg_parser" #define LOG_VERBOSE @@ -52,9 +52,17 @@ static const int frame_rate_tab[][2] = { void mpeg_parser_init (mpeg_parser_t *parser) { + parser->chunk_buffer = xine_xmalloc(BUFFER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE); mpeg_parser_reset(parser); } +void mpeg_parser_dispose (mpeg_parser_t *parser) +{ + if ( parser == NULL ) return; + + free(parser->chunk_buffer); +} + void mpeg_parser_reset (mpeg_parser_t *parser) { parser->shift = 0xffffff00; @@ -162,8 +170,7 @@ static int parse_chunk (mpeg_parser_t *parser, int code, uint8_t *buffer, int le case 0xb3: /* sequence_header_code */ { int value; - int width; - int height; + uint16_t width, height; if (parser->is_sequence_needed) { parser->is_sequence_needed = 0; diff --git a/src/libffmpeg/mpeg_parser.h b/src/libffmpeg/mpeg_parser.h index 9cc12f380..859a0fcec 100644 --- a/src/libffmpeg/mpeg_parser.h +++ b/src/libffmpeg/mpeg_parser.h @@ -20,7 +20,7 @@ * Simple MPEG-ES parser/framer by Thibaut Mattern (tmattern@noos.fr) * based on libmpeg2 decoder. * - * $Id: mpeg_parser.h,v 1.2 2004/09/21 19:27:18 tmattern Exp $ + * $Id: mpeg_parser.h,v 1.5 2007/03/29 18:52:45 dgp85 Exp $ */ #ifndef HAVE_MPEG_PARSER_H #define HAVE_MPEG_PARSER_H @@ -37,23 +37,26 @@ #define D_TYPE 4 typedef struct mpeg_parser_s { - uint32_t shift; - int is_sequence_needed; - uint8_t chunk_buffer[BUFFER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE]; + uint8_t *chunk_buffer; uint8_t *chunk_ptr; uint8_t *chunk_start; + uint32_t shift; int buffer_size; uint8_t code; uint8_t picture_coding_type; - int rate_code; + + uint8_t is_sequence_needed:1; + uint8_t is_mpeg1:1; /* public */ + uint8_t has_sequence:1; /* public */ + uint8_t in_slice:1; + + uint8_t rate_code:4; + int aspect_ratio_info; - int in_slice; /* public properties */ - int is_mpeg1; - int has_sequence; - int width; - int height; + uint16_t width; + uint16_t height; int frame_duration; double frame_aspect_ratio; @@ -62,6 +65,9 @@ typedef struct mpeg_parser_s { /* parser initialization */ void mpeg_parser_init (mpeg_parser_t *parser); +/* parser disposal */ +void mpeg_parser_dispose (mpeg_parser_t *parser); + /* read a frame * return a pointer to the first byte of the next frame * or NULL if more bytes are needed diff --git a/src/libffmpeg/video_decoder.c b/src/libffmpeg/video_decoder.c index 2d5733211..e366a945c 100644 --- a/src/libffmpeg/video_decoder.c +++ b/src/libffmpeg/video_decoder.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2001-2004 the xine project + * Copyright (C) 2001-2007 the xine project * * This file is part of xine, a free video player. * @@ -17,7 +17,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA * - * $Id: video_decoder.c,v 1.63.2.1 2006/12/02 01:20:07 dgp85 Exp $ + * $Id: video_decoder.c,v 1.73 2007/03/29 18:41:02 dgp85 Exp $ * * xine video decoder plugin using ffmpeg * @@ -25,6 +25,7 @@ #ifdef HAVE_CONFIG_H #include "config.h" +#include "ffmpeg_config.h" #endif #include <stdlib.h> @@ -74,8 +75,14 @@ struct ff_video_decoder_s { xine_stream_t *stream; int64_t pts; int video_step; - int decoder_ok; - int decoder_init_mode; + + uint8_t decoder_ok:1; + uint8_t decoder_init_mode:1; + uint8_t is_mpeg12:1; + uint8_t pp_available:1; + uint8_t yuv_init:1; + uint8_t is_direct_rendering_disabled:1; + uint8_t cs_convert_init:1; xine_bmiheader bih; unsigned char *buf; @@ -89,15 +96,13 @@ struct ff_video_decoder_s { AVCodecContext *context; AVCodec *codec; - int pp_available; int pp_quality; int pp_flags; pp_context_t *pp_context; pp_mode_t *pp_mode; /* mpeg-es parsing */ - mpeg_parser_t mpeg_parser; - int is_mpeg12; + mpeg_parser_t *mpeg_parser; double aspect_ratio; int aspect_ratio_prio; @@ -105,11 +110,10 @@ struct ff_video_decoder_s { int crop_right, crop_bottom; int output_format; - yuv_planes_t yuv; - int yuv_init; - int cs_convert_init; - int is_direct_rendering_disabled; + xine_list_t *dr1_frames; + + yuv_planes_t yuv; AVPaletteControl palette_control; }; @@ -199,16 +203,25 @@ static int get_buffer(AVCodecContext *context, AVFrame *av_frame){ av_frame->type= FF_BUFFER_TYPE_USER; + xine_list_push_back(this->dr1_frames, av_frame); + return 0; } static void release_buffer(struct AVCodecContext *context, AVFrame *av_frame){ + ff_video_decoder_t *this = (ff_video_decoder_t *)context->opaque; if (av_frame->type == FF_BUFFER_TYPE_USER) { vo_frame_t *img = (vo_frame_t *)av_frame->opaque; + xine_list_iterator_t it; assert(av_frame->opaque); img->free(img); + + it = xine_list_find(this->dr1_frames, av_frame); + assert(it); + if( it != NULL ) + xine_list_remove(this->dr1_frames, it); } else { avcodec_default_release_buffer(context, av_frame); } @@ -233,7 +246,7 @@ static const ff_codec_t ff_video_lookup[] = { {BUF_VIDEO_3IVX, CODEC_ID_MPEG4, "ISO MPEG-4 (3ivx, ffmpeg)"}, {BUF_VIDEO_JPEG, CODEC_ID_MJPEG, "Motion JPEG (ffmpeg)"}, {BUF_VIDEO_MJPEG, CODEC_ID_MJPEG, "Motion JPEG (ffmpeg)"}, - {BUF_VIDEO_MJPEG_B, CODEC_ID_MJPEGB, "Motion JPEG B (ffmpeg"}, + {BUF_VIDEO_MJPEG_B, CODEC_ID_MJPEGB, "Motion JPEG B (ffmpeg)"}, {BUF_VIDEO_I263, CODEC_ID_H263I, "ITU H.263 (ffmpeg)"}, {BUF_VIDEO_H263, CODEC_ID_H263, "H.263 (ffmpeg)"}, {BUF_VIDEO_RV10, CODEC_ID_RV10, "Real Video 1.0 (ffmpeg)"}, @@ -245,6 +258,9 @@ static const ff_codec_t ff_video_lookup[] = { {BUF_VIDEO_DV, CODEC_ID_DVVIDEO, "DV (ffmpeg)"}, {BUF_VIDEO_HUFFYUV, CODEC_ID_HUFFYUV, "HuffYUV (ffmpeg)"}, {BUF_VIDEO_VP31, CODEC_ID_VP3, "On2 VP3.1 (ffmpeg)"}, + {BUF_VIDEO_VP5, CODEC_ID_VP5, "On2 VP5 (ffmpeg)"}, + {BUF_VIDEO_VP6, CODEC_ID_VP6, "On2 VP6 (ffmpeg)"}, + {BUF_VIDEO_VP6F, CODEC_ID_VP6F, "On2 VP6 (ffmpeg)"}, {BUF_VIDEO_4XM, CODEC_ID_4XM, "4X Video (ffmpeg)"}, {BUF_VIDEO_CINEPAK, CODEC_ID_CINEPAK, "Cinepak (ffmpeg)"}, {BUF_VIDEO_MSVC, CODEC_ID_MSVIDEO1, "Microsoft Video 1 (ffmpeg)"}, @@ -372,7 +388,7 @@ static void init_video_codec (ff_video_decoder_t *this, unsigned int codec_type) /* enable direct rendering by default */ this->output_format = XINE_IMGFMT_YV12; #ifdef ENABLE_DIRECT_RENDERING - if( this->codec->capabilities & CODEC_CAP_DR1 ) { + if( this->codec->capabilities & CODEC_CAP_DR1 && this->codec->id != CODEC_ID_H264 ) { this->context->get_buffer = get_buffer; this->context->release_buffer = release_buffer; xprintf(this->stream->xine, XINE_VERBOSITY_LOG, @@ -411,7 +427,7 @@ static void pp_change_quality (ff_video_decoder_t *this) { this->pp_quality = this->class->pp_quality; if(this->pp_available && this->pp_quality) { - if(!this->pp_context) + if(!this->pp_context && this->context) this->pp_context = pp_get_context(this->context->width, this->context->height, this->pp_flags); if(this->pp_mode) @@ -504,7 +520,7 @@ static int ff_handle_mpeg_sequence(ff_video_decoder_t *this, mpeg_parser_t *pars data.pan_scan = 0; xine_event_send(this->stream, &event); } - this->video_step = this->mpeg_parser.frame_duration; + this->video_step = this->mpeg_parser->frame_duration; return 1; } @@ -797,7 +813,7 @@ static void ff_check_bufsize (ff_video_decoder_t *this, int size) { xprintf(this->stream->xine, XINE_VERBOSITY_LOG, _("ffmpeg_video_dec: increasing buffer to %d to avoid overflow.\n"), this->bufsize); - this->buf = realloc(this->buf, this->bufsize); + this->buf = realloc(this->buf, this->bufsize + FF_INPUT_BUFFER_PADDING_SIZE ); } } @@ -807,8 +823,14 @@ static void ff_handle_preview_buffer (ff_video_decoder_t *this, buf_element_t *b lprintf ("preview buffer\n"); codec_type = buf->type & 0xFFFF0000; - if (codec_type == BUF_VIDEO_MPEG) + if (codec_type == BUF_VIDEO_MPEG) { this->is_mpeg12 = 1; + if ( this->mpeg_parser == NULL ) { + this->mpeg_parser = xine_xmalloc(sizeof(mpeg_parser_t)); + mpeg_parser_init(this->mpeg_parser); + this->decoder_init_mode = 0; + } + } if (this->decoder_init_mode && !this->is_mpeg12) { init_video_codec(this, codec_type); @@ -822,7 +844,7 @@ static void ff_handle_header_buffer (ff_video_decoder_t *this, buf_element_t *bu lprintf ("header buffer\n"); /* accumulate data */ - ff_check_bufsize(this, this->size + buf->size + FF_INPUT_BUFFER_PADDING_SIZE); + ff_check_bufsize(this, this->size + buf->size); xine_fast_memcpy (&this->buf[this->size], buf->content, buf->size); this->size += buf->size; @@ -961,7 +983,7 @@ static void ff_handle_mpeg12_buffer (ff_video_decoder_t *this, buf_element_t *bu got_picture = 0; if (!flush) { - current = mpeg_parser_decode_data(&this->mpeg_parser, + current = mpeg_parser_decode_data(this->mpeg_parser, buf->content + offset, buf->content + offset + size, &next_flush); } else { @@ -973,8 +995,8 @@ static void ff_handle_mpeg12_buffer (ff_video_decoder_t *this, buf_element_t *bu return; } - if (this->mpeg_parser.has_sequence) { - ff_handle_mpeg_sequence(this, &this->mpeg_parser); + if (this->mpeg_parser->has_sequence) { + ff_handle_mpeg_sequence(this, this->mpeg_parser); } if (!this->decoder_ok) @@ -983,16 +1005,16 @@ static void ff_handle_mpeg12_buffer (ff_video_decoder_t *this, buf_element_t *bu if (flush) { lprintf("flush lavc buffers\n"); /* hack: ffmpeg outputs the last frame if size=0 */ - this->mpeg_parser.buffer_size = 0; + this->mpeg_parser->buffer_size = 0; } /* skip decoding b frames if too late */ this->context->hurry_up = (this->skipframes > 0); - lprintf("avcodec_decode_video: size=%d\n", this->mpeg_parser.buffer_size); + lprintf("avcodec_decode_video: size=%d\n", this->mpeg_parser->buffer_size); len = avcodec_decode_video (this->context, this->av_frame, - &got_picture, this->mpeg_parser.chunk_buffer, - this->mpeg_parser.buffer_size); + &got_picture, this->mpeg_parser->chunk_buffer, + this->mpeg_parser->buffer_size); lprintf("avcodec_decode_video: decoded_size=%d, got_picture=%d\n", len, got_picture); len = current - buf->content - offset; @@ -1098,7 +1120,7 @@ static void ff_handle_buffer (ff_video_decoder_t *this, buf_element_t *buf) { lprintf("no memcpy needed to accumulate data\n"); } else { /* copy data into our internal buffer */ - ff_check_bufsize(this, this->size + buf->size + FF_INPUT_BUFFER_PADDING_SIZE); + ff_check_bufsize(this, this->size + buf->size); chunk_buf = this->buf; /* ff_check_bufsize might realloc this->buf */ xine_fast_memcpy (&this->buf[this->size], buf->content, buf->size); @@ -1118,7 +1140,10 @@ static void ff_handle_buffer (ff_video_decoder_t *this, buf_element_t *buf) { int codec_type = buf->type & 0xFFFF0000; /* pad input data */ - chunk_buf[this->size] = 0; + /* note: bitstream, alt bitstream reader or something will cause + * severe mpeg4 artifacts if padding is less than 32 bits. + */ + memset(&chunk_buf[this->size], 0, FF_INPUT_BUFFER_PADDING_SIZE); while (this->size > 0) { @@ -1146,7 +1171,7 @@ static void ff_handle_buffer (ff_video_decoder_t *this, buf_element_t *buf) { this->size -= len; if (this->size > 0) { - ff_check_bufsize(this, this->size + FF_INPUT_BUFFER_PADDING_SIZE); + ff_check_bufsize(this, this->size); memmove (this->buf, &chunk_buf[offset], this->size); chunk_buf = this->buf; } @@ -1252,7 +1277,7 @@ static void ff_handle_buffer (ff_video_decoder_t *this, buf_element_t *buf) { img->crop_bottom = this->crop_bottom; this->skipframes = img->draw(img, this->stream); - + if(free_img) img->free(img); } @@ -1342,7 +1367,7 @@ static void ff_reset (video_decoder_t *this_gen) { avcodec_flush_buffers(this->context); if (this->is_mpeg12) - mpeg_parser_reset(&this->mpeg_parser); + mpeg_parser_reset(this->mpeg_parser); } static void ff_discontinuity (video_decoder_t *this_gen) { @@ -1356,12 +1381,23 @@ static void ff_dispose (video_decoder_t *this_gen) { ff_video_decoder_t *this = (ff_video_decoder_t *) this_gen; lprintf ("ff_dispose\n"); - + if (this->decoder_ok) { + xine_list_iterator_t it; + AVFrame *av_frame; + pthread_mutex_lock(&ffmpeg_lock); avcodec_close (this->context); pthread_mutex_unlock(&ffmpeg_lock); - + + /* frame garbage collector here - workaround for buggy ffmpeg codecs that + * don't release their DR1 frames */ + while( (it = xine_list_front(this->dr1_frames)) != NULL ) + { + av_frame = (AVFrame *)xine_list_get_value(this->dr1_frames, it); + release_buffer(this->context, av_frame); + } + this->stream->video_out->close(this->stream->video_out, this->stream); this->decoder_ok = 0; } @@ -1390,6 +1426,10 @@ static void ff_dispose (video_decoder_t *this_gen) { if(this->pp_mode) pp_free_mode(this->pp_mode); + + mpeg_parser_dispose(this->mpeg_parser); + + xine_list_delete(this->dr1_frames); free (this_gen); } @@ -1429,7 +1469,9 @@ static video_decoder_t *ff_video_open_plugin (video_decoder_class_t *class_gen, this->pp_context = NULL; this->pp_mode = NULL; - mpeg_parser_init(&this->mpeg_parser); + this->mpeg_parser = NULL; + + this->dr1_frames = xine_list_new(); return &this->video_decoder; } @@ -1479,74 +1521,224 @@ void *init_video_plugin (xine_t *xine, void *data) { } static uint32_t supported_video_types[] = { - BUF_VIDEO_MSMPEG4_V1, + #ifdef CONFIG_MSMPEG4V1_DECODER + BUF_VIDEO_MSMPEG4_V1, + #endif + #ifdef CONFIG_MSMPEG4V2_DECODER BUF_VIDEO_MSMPEG4_V2, - BUF_VIDEO_MSMPEG4_V3, - BUF_VIDEO_WMV7, + #endif + #ifdef CONFIG_MSMPEG4V3_DECODER + BUF_VIDEO_MSMPEG4_V3, + #endif + #ifdef CONFIG_WMV1_DECODER + BUF_VIDEO_WMV7, + #endif + #ifdef CONFIG_WMV2_DECODER + BUF_VIDEO_WMV8, + #endif + #ifdef CONFIG_WMV3_DECODER + BUF_VIDEO_WMV9, + #endif + #ifdef CONFIG_MPEG4_DECODER BUF_VIDEO_MPEG4, - BUF_VIDEO_XVID, - BUF_VIDEO_DIVX5, + #endif + #ifdef CONFIG_MPEG4_DECODER + BUF_VIDEO_XVID, + #endif + #ifdef CONFIG_MPEG4_DECODER + BUF_VIDEO_DIVX5, + #endif + #ifdef CONFIG_MPEG4_DECODER BUF_VIDEO_3IVX, + #endif + #ifdef CONFIG_MJPEG_DECODER + BUF_VIDEO_JPEG, + #endif + #ifdef CONFIG_MJPEG_DECODER BUF_VIDEO_MJPEG, + #endif + #ifdef CONFIG_MJPEGB_DECODER BUF_VIDEO_MJPEG_B, + #endif + #ifdef CONFIG_H263I_DECODER + BUF_VIDEO_I263, + #endif + #ifdef CONFIG_H263_DECODER BUF_VIDEO_H263, + #endif + #ifdef CONFIG_RV10_DECODER BUF_VIDEO_RV10, + #endif + #ifdef CONFIG_RV20_DECODER BUF_VIDEO_RV20, + #endif + #ifdef CONFIG_INDEO3_DECODER BUF_VIDEO_IV31, + #endif + #ifdef CONFIG_INDEO3_DECODER BUF_VIDEO_IV32, + #endif + #ifdef CONFIG_SVQ1_DECODER BUF_VIDEO_SORENSON_V1, + #endif + #ifdef CONFIG_SVQ3_DECODER BUF_VIDEO_SORENSON_V3, - BUF_VIDEO_JPEG, - BUF_VIDEO_MPEG, + #endif + #ifdef CONFIG_DVVIDEO_DECODER BUF_VIDEO_DV, + #endif + #ifdef CONFIG_HUFFYUV_DECODER BUF_VIDEO_HUFFYUV, + #endif + #ifdef CONFIG_VP3_DECODER BUF_VIDEO_VP31, + #endif + #ifdef CONFIG_VP5_DECODER + BUF_VIDEO_VP5, + #endif + #ifdef CONFIG_VP6_DECODER + BUF_VIDEO_VP6, + BUF_VIDEO_VP6F, + #endif + #ifdef CONFIG_4XM_DECODER BUF_VIDEO_4XM, + #endif + #ifdef CONFIG_CINEPAK_DECODER BUF_VIDEO_CINEPAK, + #endif + #ifdef CONFIG_MSVIDEO1_DECODER BUF_VIDEO_MSVC, + #endif + #ifdef CONFIG_MSRLE_DECODER BUF_VIDEO_MSRLE, + #endif + #ifdef CONFIG_RPZA_DECODER BUF_VIDEO_RPZA, + #endif + #ifdef CONFIG_CYUV_DECODER BUF_VIDEO_CYUV, + #endif + #ifdef CONFIG_ROQ_DECODER BUF_VIDEO_ROQ, + #endif + #ifdef CONFIG_IDCIN_DECODER BUF_VIDEO_IDCIN, + #endif + #ifdef CONFIG_XAN_WC3_DECODER BUF_VIDEO_WC3, + #endif + #ifdef CONFIG_WS_VQA_DECODER BUF_VIDEO_VQA, + #endif + #ifdef CONFIG_INTERPLAY_VIDEO_DECODER BUF_VIDEO_INTERPLAY, + #endif + #ifdef CONFIG_FLIC_DECODER BUF_VIDEO_FLI, + #endif + #ifdef CONFIG_8BPS_DECODER BUF_VIDEO_8BPS, + #endif + #ifdef CONFIG_SMC_DECODER BUF_VIDEO_SMC, - BUF_VIDEO_VMD, + #endif + #ifdef CONFIG_TRUEMOTION1_DECODER BUF_VIDEO_DUCKTM1, + #endif + #ifdef CONFIG_TRUEMOTION2_DECODER BUF_VIDEO_DUCKTM2, + #endif + #ifdef CONFIG_VMDVIDEO_DECODER + BUF_VIDEO_VMD, + #endif + #ifdef CONFIG_ZLIB_DECODER BUF_VIDEO_ZLIB, + #endif + #ifdef CONFIG_MSZH_DECODER BUF_VIDEO_MSZH, + #endif + #ifdef CONFIG_ASV1_DECODER BUF_VIDEO_ASV1, + #endif + #ifdef CONFIG_ASV2_DECODER BUF_VIDEO_ASV2, + #endif + #ifdef CONFIG_VCR1_DECODER BUF_VIDEO_ATIVCR1, + #endif + #ifdef CONFIG_FLV_DECODER BUF_VIDEO_FLV1, + #endif + #ifdef CONFIG_QTRLE_DECODER BUF_VIDEO_QTRLE, + #endif + #ifdef CONFIG_H264_DECODER BUF_VIDEO_H264, + #endif + #ifdef CONFIG_H261_DECODER BUF_VIDEO_H261, + #endif + #ifdef CONFIG_AASC_DECODER BUF_VIDEO_AASC, + #endif + #ifdef CONFIG_LOCO_DECODER BUF_VIDEO_LOCO, + #endif + #ifdef CONFIG_QDRAW_DECODER BUF_VIDEO_QDRW, + #endif + #ifdef CONFIG_QPEG_DECODER BUF_VIDEO_QPEG, + #endif + #ifdef CONFIG_TSCC_DECODER BUF_VIDEO_TSCC, + #endif + #ifdef CONFIG_ULTI_DECODER BUF_VIDEO_ULTI, + #endif + #ifdef CONFIG_WNV1_DECODER BUF_VIDEO_WNV1, + #endif + #ifdef CONFIG_VIXL_DECODER BUF_VIDEO_XL, + #endif + #ifdef CONFIG_INDEO2_DECODER BUF_VIDEO_RT21, + #endif + #ifdef CONFIG_FRAPS_DECODER BUF_VIDEO_FPS1, - BUF_VIDEO_WMV9, + #endif + #ifdef CONFIG_MPEG1VIDEO_DECODER + BUF_VIDEO_MPEG, + #endif + #ifdef CONFIG_CSCD_DECODER BUF_VIDEO_CSCD, + #endif + #ifdef CONFIG_AVS_DECODER + BUF_VIDEO_AVS, + #endif + #ifdef CONFIG_MMVIDEO_DECODER BUF_VIDEO_ALGMM, + #endif + #ifdef CONFIG_ZMBV_DECODER BUF_VIDEO_ZMBV, - BUF_VIDEO_AVS, + #endif + #ifdef CONFIG_SMACKVIDEO_DECODER BUF_VIDEO_SMACKER, + #endif + #ifdef CONFIG_NUV_DECODER BUF_VIDEO_NUV, + #endif + #ifdef CONFIG_KMVC_DECODER BUF_VIDEO_KMVC, + #endif + #ifdef CONFIG_FLASHSV_DECODER BUF_VIDEO_FLASHSV, + #endif + #ifdef CONFIG_CAVS_DECODER BUF_VIDEO_CAVS, + #endif + 0 }; diff --git a/src/libffmpeg/xine_decoder.c b/src/libffmpeg/xine_decoder.c index 10f420997..2eeb9746b 100644 --- a/src/libffmpeg/xine_decoder.c +++ b/src/libffmpeg/xine_decoder.c @@ -17,7 +17,7 @@ * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA * - * $Id: xine_decoder.c,v 1.171 2006/10/11 12:53:12 dgp85 Exp $ + * $Id: xine_decoder.c,v 1.173 2007/01/13 21:19:52 miguelfreitas Exp $ * * xine decoder plugin using ffmpeg * @@ -25,6 +25,7 @@ #ifdef HAVE_CONFIG_H #include "config.h" +#include "ffmpeg_config.h" #endif #include "xine_internal.h" @@ -39,114 +40,273 @@ pthread_once_t once_control = PTHREAD_ONCE_INIT; pthread_mutex_t ffmpeg_lock; #ifndef HAVE_FFMPEG + +#define REGISTER_ENCODER(X,x) \ + if(ENABLE_##X##_ENCODER) register_avcodec(&x##_encoder) +#define REGISTER_DECODER(X,x) \ + if(ENABLE_##X##_DECODER) register_avcodec(&x##_decoder) +#define REGISTER_ENCDEC(X,x) REGISTER_ENCODER(X,x); REGISTER_DECODER(X,x) + +#define REGISTER_PARSER(X,x) \ + if(ENABLE_##X##_PARSER) av_register_codec_parser(&x##_parser) + +/* If you do not call this function, then you can select exactly which + formats you want to support */ + +/** + * simple call to register all the codecs. + */ void avcodec_register_all(void) { static int inited = 0; - + if (inited != 0) - return; + return; inited = 1; - /* decoders */ - register_avcodec(&h263_decoder); - register_avcodec(&mpeg4_decoder); - register_avcodec(&msmpeg4v1_decoder); - register_avcodec(&msmpeg4v2_decoder); - register_avcodec(&msmpeg4v3_decoder); - register_avcodec(&wmv1_decoder); - register_avcodec(&wmv2_decoder); - register_avcodec(&h263i_decoder); - register_avcodec(&rv10_decoder); - register_avcodec(&rv20_decoder); - register_avcodec(&svq1_decoder); - register_avcodec(&svq3_decoder); - register_avcodec(&wmav1_decoder); - register_avcodec(&wmav2_decoder); - register_avcodec(&indeo3_decoder); - register_avcodec(&mpeg1video_decoder); - register_avcodec(&dvvideo_decoder); - register_avcodec(&pcm_s16le_decoder); - register_avcodec(&mjpeg_decoder); - register_avcodec(&mjpegb_decoder); - register_avcodec(&mp2_decoder); - register_avcodec(&mp3_decoder); - register_avcodec(&mace3_decoder); - register_avcodec(&mace6_decoder); - register_avcodec(&huffyuv_decoder); - register_avcodec(&cyuv_decoder); - register_avcodec(&h264_decoder); - register_avcodec(&vp3_decoder); - register_avcodec(&fourxm_decoder); - register_avcodec(&ra_144_decoder); - register_avcodec(&ra_288_decoder); - register_avcodec(&adpcm_ms_decoder); - register_avcodec(&adpcm_ima_qt_decoder); - register_avcodec(&adpcm_ima_wav_decoder); - register_avcodec(&adpcm_ima_dk3_decoder); - register_avcodec(&adpcm_ima_dk4_decoder); - register_avcodec(&adpcm_ima_ws_decoder); - register_avcodec(&adpcm_ima_smjpeg_decoder); - register_avcodec(&adpcm_xa_decoder); - register_avcodec(&adpcm_4xm_decoder); - register_avcodec(&adpcm_ea_decoder); - register_avcodec(&pcm_alaw_decoder); - register_avcodec(&pcm_mulaw_decoder); - register_avcodec(&roq_dpcm_decoder); - register_avcodec(&interplay_dpcm_decoder); - register_avcodec(&cinepak_decoder); - register_avcodec(&msvideo1_decoder); - register_avcodec(&msrle_decoder); - register_avcodec(&rpza_decoder); - register_avcodec(&roq_decoder); - register_avcodec(&idcin_decoder); - register_avcodec(&xan_wc3_decoder); - register_avcodec(&vqa_decoder); - register_avcodec(&interplay_video_decoder); - register_avcodec(&flic_decoder); - register_avcodec(&smc_decoder); - register_avcodec(&eightbps_decoder); - register_avcodec(&vmdvideo_decoder); - register_avcodec(&vmdaudio_decoder); - register_avcodec(&truemotion1_decoder); - register_avcodec(&mszh_decoder); - register_avcodec(&zlib_decoder); - register_avcodec(&xan_dpcm_decoder); - register_avcodec(&asv1_decoder); - register_avcodec(&asv2_decoder); - register_avcodec(&vcr1_decoder); - register_avcodec(&flv_decoder); - register_avcodec(&qtrle_decoder); - register_avcodec(&flac_decoder); - register_avcodec(&aasc_decoder); - register_avcodec(&alac_decoder); - register_avcodec(&h261_decoder); - register_avcodec(&loco_decoder); - register_avcodec(&qdraw_decoder); - register_avcodec(&qpeg_decoder); - register_avcodec(&tscc_decoder); - register_avcodec(&ulti_decoder); - register_avcodec(&wnv1_decoder); - register_avcodec(&xl_decoder); - register_avcodec(&indeo2_decoder); - register_avcodec(&fraps_decoder); - register_avcodec(&shorten_decoder); - register_avcodec(&qdm2_decoder); - register_avcodec(&truemotion2_decoder); - register_avcodec(&wmv3_decoder); - register_avcodec(&cscd_decoder); - register_avcodec(&mmvideo_decoder); - register_avcodec(&zmbv_decoder); - register_avcodec(&avs_decoder); - register_avcodec(&smacker_decoder); - register_avcodec(&smackaud_decoder); - register_avcodec(&nuv_decoder); - register_avcodec(&kmvc_decoder); - register_avcodec(&flashsv_decoder); - register_avcodec(&cavs_decoder); - register_avcodec(&cook_decoder); - register_avcodec(&truespeech_decoder); - register_avcodec(&tta_decoder); + /* video codecs */ + REGISTER_DECODER(AASC, aasc); + REGISTER_ENCDEC (ASV1, asv1); + REGISTER_ENCDEC (ASV2, asv2); + REGISTER_DECODER(AVS, avs); + REGISTER_DECODER(BMP, bmp); + REGISTER_DECODER(CAVS, cavs); + REGISTER_DECODER(CINEPAK, cinepak); + REGISTER_DECODER(CLJR, cljr); + REGISTER_DECODER(CSCD, cscd); + REGISTER_DECODER(CYUV, cyuv); + REGISTER_DECODER(DSICINVIDEO, dsicinvideo); + REGISTER_ENCDEC (DVVIDEO, dvvideo); + REGISTER_DECODER(EIGHTBPS, eightbps); + REGISTER_ENCDEC (FFV1, ffv1); + REGISTER_ENCDEC (FFVHUFF, ffvhuff); + REGISTER_DECODER(FLASHSV, flashsv); + REGISTER_DECODER(FLIC, flic); + REGISTER_ENCDEC (FLV, flv); + REGISTER_DECODER(FOURXM, fourxm); + REGISTER_DECODER(FRAPS, fraps); + REGISTER_ENCDEC (GIF, gif); + REGISTER_ENCDEC (H261, h261); + REGISTER_ENCDEC (H263, h263); + REGISTER_DECODER(H263I, h263i); + REGISTER_ENCODER(H263P, h263p); + REGISTER_DECODER(H264, h264); + REGISTER_ENCDEC (HUFFYUV, huffyuv); + REGISTER_DECODER(IDCIN, idcin); + REGISTER_DECODER(INDEO2, indeo2); + REGISTER_DECODER(INDEO3, indeo3); + REGISTER_DECODER(INTERPLAY_VIDEO, interplay_video); + REGISTER_ENCODER(JPEGLS, jpegls); + REGISTER_DECODER(KMVC, kmvc); + REGISTER_ENCODER(LJPEG, ljpeg); + REGISTER_DECODER(LOCO, loco); + REGISTER_DECODER(MDEC, mdec); + REGISTER_ENCDEC (MJPEG, mjpeg); + REGISTER_DECODER(MJPEGB, mjpegb); + REGISTER_DECODER(MMVIDEO, mmvideo); +#ifdef HAVE_XVMC + REGISTER_DECODER(MPEG_XVMC, mpeg_xvmc); +#endif + REGISTER_ENCDEC (MPEG1VIDEO, mpeg1video); + REGISTER_ENCDEC (MPEG2VIDEO, mpeg2video); + REGISTER_ENCDEC (MPEG4, mpeg4); + REGISTER_DECODER(MPEGVIDEO, mpegvideo); + REGISTER_ENCDEC (MSMPEG4V1, msmpeg4v1); + REGISTER_ENCDEC (MSMPEG4V2, msmpeg4v2); + REGISTER_ENCDEC (MSMPEG4V3, msmpeg4v3); + REGISTER_DECODER(MSRLE, msrle); + REGISTER_DECODER(MSVIDEO1, msvideo1); + REGISTER_DECODER(MSZH, mszh); + REGISTER_DECODER(NUV, nuv); + REGISTER_ENCODER(PAM, pam); + REGISTER_ENCODER(PBM, pbm); + REGISTER_ENCODER(PGM, pgm); + REGISTER_ENCODER(PGMYUV, pgmyuv); +#ifdef CONFIG_ZLIB + REGISTER_ENCDEC (PNG, png); +#endif + REGISTER_ENCODER(PPM, ppm); + REGISTER_DECODER(QDRAW, qdraw); + REGISTER_DECODER(QPEG, qpeg); + REGISTER_DECODER(QTRLE, qtrle); + REGISTER_ENCDEC (RAWVIDEO, rawvideo); + REGISTER_DECODER(ROQ, roq); + REGISTER_DECODER(RPZA, rpza); + REGISTER_ENCDEC (RV10, rv10); + REGISTER_ENCDEC (RV20, rv20); + REGISTER_DECODER(SMACKER, smacker); + REGISTER_DECODER(SMC, smc); + REGISTER_ENCDEC (SNOW, snow); + REGISTER_DECODER(SP5X, sp5x); + REGISTER_ENCDEC (SVQ1, svq1); + REGISTER_DECODER(SVQ3, svq3); + REGISTER_DECODER(TARGA, targa); + REGISTER_DECODER(THEORA, theora); + REGISTER_DECODER(TIERTEXSEQVIDEO, tiertexseqvideo); + REGISTER_DECODER(TIFF, tiff); + REGISTER_DECODER(TRUEMOTION1, truemotion1); + REGISTER_DECODER(TRUEMOTION2, truemotion2); + REGISTER_DECODER(TSCC, tscc); + REGISTER_DECODER(ULTI, ulti); + REGISTER_DECODER(VC1, vc1); + REGISTER_DECODER(VCR1, vcr1); + REGISTER_DECODER(VMDVIDEO, vmdvideo); + REGISTER_DECODER(VMNC, vmnc); + REGISTER_DECODER(VP3, vp3); + REGISTER_DECODER(VP5, vp5); + REGISTER_DECODER(VP6, vp6); + REGISTER_DECODER(VP6F, vp6f); + REGISTER_DECODER(VQA, vqa); + REGISTER_ENCDEC (WMV1, wmv1); + REGISTER_ENCDEC (WMV2, wmv2); + REGISTER_DECODER(WMV3, wmv3); + REGISTER_DECODER(WNV1, wnv1); +#ifdef CONFIG_X264 + REGISTER_ENCODER(X264, x264); +#endif + REGISTER_DECODER(XAN_WC3, xan_wc3); + REGISTER_DECODER(XL, xl); +#ifdef CONFIG_XVID + REGISTER_ENCODER(XVID, xvid); +#endif + REGISTER_ENCDEC (ZLIB, zlib); +#ifdef CONFIG_ZLIB + REGISTER_ENCDEC (ZMBV, zmbv); +#endif + + /* audio codecs */ +#ifdef CONFIG_LIBFAAD + REGISTER_DECODER(AAC, aac); + REGISTER_DECODER(MPEG4AAC, mpeg4aac); +#endif +#ifdef CONFIG_LIBA52 + REGISTER_DECODER(AC3, ac3); +#endif + REGISTER_ENCODER(AC3, ac3); + REGISTER_DECODER(ALAC, alac); +#if defined(CONFIG_AMR_NB) || defined(CONFIG_AMR_NB_FIXED) + REGISTER_ENCDEC (AMR_NB, amr_nb); +#endif +#ifdef CONFIG_AMR_WB + REGISTER_ENCDEC (AMR_WB, amr_wb); +#endif + REGISTER_DECODER(COOK, cook); + REGISTER_DECODER(DSICINAUDIO, dsicinaudio); +#ifdef CONFIG_LIBDTS + REGISTER_DECODER(DTS, dts); +#endif +#ifdef CONFIG_LIBFAAC + REGISTER_ENCODER(FAAC, faac); +#endif + REGISTER_ENCDEC (FLAC, flac); + REGISTER_DECODER(IMC, imc); +#ifdef CONFIG_LIBGSM + REGISTER_ENCDEC (LIBGSM, libgsm); +#endif + REGISTER_DECODER(MACE3, mace3); + REGISTER_DECODER(MACE6, mace6); + REGISTER_ENCDEC (MP2, mp2); + REGISTER_DECODER(MP3, mp3); + REGISTER_DECODER(MP3ADU, mp3adu); +#ifdef CONFIG_LIBMP3LAME + REGISTER_ENCODER(MP3LAME, mp3lame); +#endif + REGISTER_DECODER(MP3ON4, mp3on4); + REGISTER_DECODER(MPC7, mpc7); +#ifdef CONFIG_LIBVORBIS + if (!ENABLE_VORBIS_ENCODER) REGISTER_ENCODER(OGGVORBIS, oggvorbis); + if (!ENABLE_VORBIS_DECODER) REGISTER_DECODER(OGGVORBIS, oggvorbis); +#endif + REGISTER_DECODER(QDM2, qdm2); + REGISTER_DECODER(RA_144, ra_144); + REGISTER_DECODER(RA_288, ra_288); + REGISTER_DECODER(SHORTEN, shorten); + REGISTER_DECODER(SMACKAUD, smackaud); + REGISTER_ENCDEC (SONIC, sonic); + REGISTER_ENCODER(SONIC_LS, sonic_ls); + REGISTER_DECODER(TRUESPEECH, truespeech); + REGISTER_DECODER(TTA, tta); + REGISTER_DECODER(VMDAUDIO, vmdaudio); + REGISTER_ENCDEC (VORBIS, vorbis); + REGISTER_DECODER(WAVPACK, wavpack); + REGISTER_DECODER(WMAV1, wmav1); + REGISTER_DECODER(WMAV2, wmav2); + REGISTER_DECODER(WS_SND1, ws_snd1); + + /* pcm codecs */ + REGISTER_ENCDEC (PCM_ALAW, pcm_alaw); + REGISTER_ENCDEC (PCM_MULAW, pcm_mulaw); + REGISTER_ENCDEC (PCM_S8, pcm_s8); + REGISTER_ENCDEC (PCM_S16BE, pcm_s16be); + REGISTER_ENCDEC (PCM_S16LE, pcm_s16le); + REGISTER_ENCDEC (PCM_S24BE, pcm_s24be); + REGISTER_ENCDEC (PCM_S24DAUD, pcm_s24daud); + REGISTER_ENCDEC (PCM_S24LE, pcm_s24le); + REGISTER_ENCDEC (PCM_S32BE, pcm_s32be); + REGISTER_ENCDEC (PCM_S32LE, pcm_s32le); + REGISTER_ENCDEC (PCM_U8, pcm_u8); + REGISTER_ENCDEC (PCM_U16BE, pcm_u16be); + REGISTER_ENCDEC (PCM_U16LE, pcm_u16le); + REGISTER_ENCDEC (PCM_U24BE, pcm_u24be); + REGISTER_ENCDEC (PCM_U24LE, pcm_u24le); + REGISTER_ENCDEC (PCM_U32BE, pcm_u32be); + REGISTER_ENCDEC (PCM_U32LE, pcm_u32le); + + /* dpcm codecs */ + REGISTER_DECODER(INTERPLAY_DPCM, interplay_dpcm); + REGISTER_DECODER(ROQ_DPCM, roq_dpcm); + REGISTER_DECODER(SOL_DPCM, sol_dpcm); + REGISTER_DECODER(XAN_DPCM, xan_dpcm); + + /* adpcm codecs */ + REGISTER_ENCDEC (ADPCM_4XM, adpcm_4xm); + REGISTER_ENCDEC (ADPCM_ADX, adpcm_adx); + REGISTER_ENCDEC (ADPCM_CT, adpcm_ct); + REGISTER_ENCDEC (ADPCM_EA, adpcm_ea); + REGISTER_ENCDEC (ADPCM_G726, adpcm_g726); + REGISTER_ENCDEC (ADPCM_IMA_DK3, adpcm_ima_dk3); + REGISTER_ENCDEC (ADPCM_IMA_DK4, adpcm_ima_dk4); + REGISTER_ENCDEC (ADPCM_IMA_QT, adpcm_ima_qt); + REGISTER_ENCDEC (ADPCM_IMA_SMJPEG, adpcm_ima_smjpeg); + REGISTER_ENCDEC (ADPCM_IMA_WAV, adpcm_ima_wav); + REGISTER_ENCDEC (ADPCM_IMA_WS, adpcm_ima_ws); + REGISTER_ENCDEC (ADPCM_MS, adpcm_ms); + REGISTER_ENCDEC (ADPCM_SBPRO_2, adpcm_sbpro_2); + REGISTER_ENCDEC (ADPCM_SBPRO_3, adpcm_sbpro_3); + REGISTER_ENCDEC (ADPCM_SBPRO_4, adpcm_sbpro_4); + REGISTER_ENCDEC (ADPCM_SWF, adpcm_swf); + REGISTER_ENCDEC (ADPCM_XA, adpcm_xa); + REGISTER_ENCDEC (ADPCM_YAMAHA, adpcm_yamaha); + + /* subtitles */ + REGISTER_ENCDEC (DVBSUB, dvbsub); + REGISTER_ENCDEC (DVDSUB, dvdsub); + + /* parsers */ + REGISTER_PARSER (AAC, aac); + REGISTER_PARSER (AC3, ac3); + REGISTER_PARSER (CAVSVIDEO, cavsvideo); + REGISTER_PARSER (DVBSUB, dvbsub); + REGISTER_PARSER (DVDSUB, dvdsub); + REGISTER_PARSER (H261, h261); + REGISTER_PARSER (H263, h263); + REGISTER_PARSER (H264, h264); + REGISTER_PARSER (MJPEG, mjpeg); + REGISTER_PARSER (MPEG4VIDEO, mpeg4video); + REGISTER_PARSER (MPEGAUDIO, mpegaudio); + REGISTER_PARSER (MPEGVIDEO, mpegvideo); + REGISTER_PARSER (PNM, pnm); + + /* + av_register_bitstream_filter(&dump_extradata_bsf); + av_register_bitstream_filter(&remove_extradata_bsf); + av_register_bitstream_filter(&noise_bsf); + av_register_bitstream_filter(&mp3_header_compress_bsf); + av_register_bitstream_filter(&mp3_header_decompress_bsf); + av_register_bitstream_filter(&mjpega_dump_header_bsf); + */ } + #endif void init_once_routine(void) { |