update libfaad2 to CVS snapshot 2004-09-15

CVS patchset: 7777
CVS date: 2005/10/29 23:57:06

1 files changed, 301 insertions, 86 deletions

diff --git a/src/libfaad/common.c b/src/libfaad/common.c
index 954d5dee6..8c0d93e95 100644
--- a/src/libfaad/common.c
+++ b/src/libfaad/common.c
@@ -1,19 +1,19 @@
 /*
 ** FAAD2 - Freeware Advanced Audio (AAC) Decoder including SBR decoding
 ** Copyright (C) 2003-2004 M. Bakker, Ahead Software AG, http://www.nero.com
-**  
+**
 ** This program is free software; you can redistribute it and/or modify
 ** it under the terms of the GNU General Public License as published by
 ** the Free Software Foundation; either version 2 of the License, or
 ** (at your option) any later version.
-** 
+**
 ** This program 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 General Public License for more details.
-** 
+**
 ** You should have received a copy of the GNU General Public License
-** along with this program; if not, write to the Free Software 
+** along with this program; if not, write to the Free Software
 ** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 **
 ** Any non-GPL usage of this software or parts of this software is strictly
@@ -22,7 +22,7 @@
 ** Commercial non-GPL licensing of this software is possible.
 ** For more info contact Ahead Software through Mpeg4AAClicense@nero.com.
 **
-** $Id: common.c,v 1.7 2004/12/03 01:15:30 tmattern Exp $
+** $Id: common.c,v 1.8 2005/10/29 23:57:06 tmmm Exp $
 **/
 
 /* just some common functions that could be used anywhere */
@@ -33,67 +33,6 @@
 #include <stdlib.h>
 #include "syntax.h"
 
-#ifdef USE_SSE
-__declspec(naked) static int32_t __fastcall test_cpuid(void)
-{
-    __asm
-    {
-        pushf
-        pop eax
-        mov ecx,eax
-        xor eax,(1<<21)
-        push eax
-        popf
-        pushf
-        pop eax
-        push ecx
-        popf
-        cmp eax,ecx
-        mov eax,0
-        setne al
-        ret
-    }
-}
-
-__declspec(naked) static void __fastcall run_cpuid(int32_t param, int32_t out[4])
-{
-    __asm
-    {
-        pushad
-        push edx
-        mov eax,ecx
-        cpuid
-        pop edi
-        mov [edi+0],eax
-        mov [edi+4],ebx
-        mov [edi+8],ecx
-        mov [edi+12],edx
-        popad
-        ret
-    }
-}
-
-uint8_t cpu_has_sse()
-{
-    int32_t features[4];
-
-    if (test_cpuid())
-    {
-        run_cpuid(1, features);
-    }
-
-    /* check for SSE */
-    if (features[3] & 0x02000000)
-        return 1;
-
-    return 0;
-}
-#else
-uint8_t cpu_has_sse()
-{
-    return 0;
-}
-#endif
 
 /* Returns the sample rate index based on the samplerate */
 uint8_t get_sr_index(const uint32_t samplerate)
@@ -146,7 +85,7 @@ uint8_t max_pred_sfb(const uint8_t sr_index)
 uint8_t max_tns_sfb(const uint8_t sr_index, const uint8_t object_type,
                     const uint8_t is_short)
 {
-    /* entry for each sampling rate	
+    /* entry for each sampling rate
      * 1    Main/LC long window
      * 2    Main/LC short window
      * 3    SSR long window
@@ -230,14 +169,13 @@ int8_t can_decode_ot(const uint8_t object_type)
     return -1;
 }
 
-/* common malloc function */
-void *faad_malloc(int32_t size)
+void *faad_malloc(size_t size)
 {
 #if 0 // defined(_WIN32) && !defined(_WIN32_WCE)
     return _aligned_malloc(size, 16);
-#else
+#else   // #ifdef 0
     return malloc(size);
-#endif
+#endif  // #ifdef 0
 }
 
 /* common free function */
@@ -247,18 +185,18 @@ void faad_free(void *b)
     _aligned_free(b);
 #else
     free(b);
-#endif
 }
+#endif
 
 static const  uint8_t    Parity [256] = {  // parity
-	0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
-	1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
-	1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
-	0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
-	1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
-	0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
-	0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
-	1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0
+    0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
+    1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
+    1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
+    0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
+    1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,
+    0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
+    0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0,1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,
+    1,0,0,1,0,1,1,0,0,1,1,0,1,0,0,1,0,1,1,0,1,0,0,1,1,0,0,1,0,1,1,0
 };
 
 static uint32_t  __r1 = 1;
@@ -293,12 +231,289 @@ static uint32_t  __r2 = 1;
  */
 uint32_t random_int(void)
 {
-	uint32_t  t1, t2, t3, t4;
+    uint32_t  t1, t2, t3, t4;
+
+    t3   = t1 = __r1;   t4   = t2 = __r2;       // Parity calculation is done via table lookup, this is also available
+    t1  &= 0xF5;        t2 >>= 25;              // on CPUs without parity, can be implemented in C and avoid unpredictable
+    t1   = Parity [t1]; t2  &= 0x63;            // jumps and slow rotate through the carry flag operations.
+    t1 <<= 31;          t2   = Parity [t2];
+
+    return (__r1 = (t3 >> 1) | t1 ) ^ (__r2 = (t4 + t4) | t2 );
+}
+
+uint32_t ones32(uint32_t x)
+{
+    x -= ((x >> 1) & 0x55555555);
+    x = (((x >> 2) & 0x33333333) + (x & 0x33333333));
+    x = (((x >> 4) + x) & 0x0f0f0f0f);
+    x += (x >> 8);
+    x += (x >> 16);
 
-	t3   = t1 = __r1;   t4   = t2 = __r2;       // Parity calculation is done via table lookup, this is also available
-	t1  &= 0xF5;        t2 >>= 25;              // on CPUs without parity, can be implemented in C and avoid unpredictable
-	t1   = Parity [t1]; t2  &= 0x63;            // jumps and slow rotate through the carry flag operations.
-	t1 <<= 31;          t2   = Parity [t2];
+    return (x & 0x0000003f);
+}
 
-	return (__r1 = (t3 >> 1) | t1 ) ^ (__r2 = (t4 + t4) | t2 );
+uint32_t floor_log2(uint32_t x)
+{
+#if 1
+    x |= (x >> 1);
+    x |= (x >> 2);
+    x |= (x >> 4);
+    x |= (x >> 8);
+    x |= (x >> 16);
+
+    return (ones32(x) - 1);
+#else
+    uint32_t count = 0;
+
+    while (x >>= 1)
+        count++;
+
+    return count;
+#endif
 }
+
+/* returns position of first bit that is not 0 from msb,
+ * starting count at lsb */
+uint32_t wl_min_lzc(uint32_t x)
+{
+#if 1
+    x |= (x >> 1);
+    x |= (x >> 2);
+    x |= (x >> 4);
+    x |= (x >> 8);
+    x |= (x >> 16);
+
+    return (ones32(x));
+#else
+    uint32_t count = 0;
+
+    while (x >>= 1)
+        count++;
+
+    return (count + 1);
+#endif
+}
+
+#ifdef FIXED_POINT
+
+#define TABLE_BITS 6
+/* just take the maximum number of bits for interpolation */
+#define INTERP_BITS (REAL_BITS-TABLE_BITS)
+
+static const real_t pow2_tab[] = {
+    REAL_CONST(1.000000000000000), REAL_CONST(1.010889286051701), REAL_CONST(1.021897148654117),
+    REAL_CONST(1.033024879021228), REAL_CONST(1.044273782427414), REAL_CONST(1.055645178360557),
+    REAL_CONST(1.067140400676824), REAL_CONST(1.078760797757120), REAL_CONST(1.090507732665258),
+    REAL_CONST(1.102382583307841), REAL_CONST(1.114386742595892), REAL_CONST(1.126521618608242),
+    REAL_CONST(1.138788634756692), REAL_CONST(1.151189229952983), REAL_CONST(1.163724858777578),
+    REAL_CONST(1.176396991650281), REAL_CONST(1.189207115002721), REAL_CONST(1.202156731452703),
+    REAL_CONST(1.215247359980469), REAL_CONST(1.228480536106870), REAL_CONST(1.241857812073484),
+    REAL_CONST(1.255380757024691), REAL_CONST(1.269050957191733), REAL_CONST(1.282870016078778),
+    REAL_CONST(1.296839554651010), REAL_CONST(1.310961211524764), REAL_CONST(1.325236643159741),
+    REAL_CONST(1.339667524053303), REAL_CONST(1.354255546936893), REAL_CONST(1.369002422974591),
+    REAL_CONST(1.383909881963832), REAL_CONST(1.398979672538311), REAL_CONST(1.414213562373095),
+    REAL_CONST(1.429613338391970), REAL_CONST(1.445180806977047), REAL_CONST(1.460917794180647),
+    REAL_CONST(1.476826145939499), REAL_CONST(1.492907728291265), REAL_CONST(1.509164427593423),
+    REAL_CONST(1.525598150744538), REAL_CONST(1.542210825407941), REAL_CONST(1.559004400237837),
+    REAL_CONST(1.575980845107887), REAL_CONST(1.593142151342267), REAL_CONST(1.610490331949254),
+    REAL_CONST(1.628027421857348), REAL_CONST(1.645755478153965), REAL_CONST(1.663676580326736),
+    REAL_CONST(1.681792830507429), REAL_CONST(1.700106353718524), REAL_CONST(1.718619298122478),
+    REAL_CONST(1.737333835273706), REAL_CONST(1.756252160373300), REAL_CONST(1.775376492526521),
+    REAL_CONST(1.794709075003107), REAL_CONST(1.814252175500399), REAL_CONST(1.834008086409342),
+    REAL_CONST(1.853979125083386), REAL_CONST(1.874167634110300), REAL_CONST(1.894575981586966),
+    REAL_CONST(1.915206561397147), REAL_CONST(1.936061793492294), REAL_CONST(1.957144124175400),
+    REAL_CONST(1.978456026387951), REAL_CONST(2.000000000000000)
+};
+
+static const real_t log2_tab[] = {
+    REAL_CONST(0.000000000000000), REAL_CONST(0.022367813028455), REAL_CONST(0.044394119358453),
+    REAL_CONST(0.066089190457772), REAL_CONST(0.087462841250339), REAL_CONST(0.108524456778169),
+    REAL_CONST(0.129283016944966), REAL_CONST(0.149747119504682), REAL_CONST(0.169925001442312),
+    REAL_CONST(0.189824558880017), REAL_CONST(0.209453365628950), REAL_CONST(0.228818690495881),
+    REAL_CONST(0.247927513443585), REAL_CONST(0.266786540694901), REAL_CONST(0.285402218862248),
+    REAL_CONST(0.303780748177103), REAL_CONST(0.321928094887362), REAL_CONST(0.339850002884625),
+    REAL_CONST(0.357552004618084), REAL_CONST(0.375039431346925), REAL_CONST(0.392317422778760),
+    REAL_CONST(0.409390936137702), REAL_CONST(0.426264754702098), REAL_CONST(0.442943495848728),
+    REAL_CONST(0.459431618637297), REAL_CONST(0.475733430966398), REAL_CONST(0.491853096329675),
+    REAL_CONST(0.507794640198696), REAL_CONST(0.523561956057013), REAL_CONST(0.539158811108031),
+    REAL_CONST(0.554588851677637), REAL_CONST(0.569855608330948), REAL_CONST(0.584962500721156),
+    REAL_CONST(0.599912842187128), REAL_CONST(0.614709844115208), REAL_CONST(0.629356620079610),
+    REAL_CONST(0.643856189774725), REAL_CONST(0.658211482751795), REAL_CONST(0.672425341971496),
+    REAL_CONST(0.686500527183218), REAL_CONST(0.700439718141092), REAL_CONST(0.714245517666123),
+    REAL_CONST(0.727920454563199), REAL_CONST(0.741466986401147), REAL_CONST(0.754887502163469),
+    REAL_CONST(0.768184324776926), REAL_CONST(0.781359713524660), REAL_CONST(0.794415866350106),
+    REAL_CONST(0.807354922057604), REAL_CONST(0.820178962415188), REAL_CONST(0.832890014164742),
+    REAL_CONST(0.845490050944375), REAL_CONST(0.857980995127572), REAL_CONST(0.870364719583405),
+    REAL_CONST(0.882643049361841), REAL_CONST(0.894817763307943), REAL_CONST(0.906890595608519),
+    REAL_CONST(0.918863237274595), REAL_CONST(0.930737337562886), REAL_CONST(0.942514505339240),
+    REAL_CONST(0.954196310386875), REAL_CONST(0.965784284662087), REAL_CONST(0.977279923499917),
+    REAL_CONST(0.988684686772166), REAL_CONST(1.000000000000000)
+};
+
+real_t pow2_fix(real_t val)
+{
+    uint32_t x1, x2;
+    uint32_t errcorr;
+    uint32_t index_frac;
+    real_t retval;
+    int32_t whole = (val >> REAL_BITS);
+
+    /* rest = [0..1] */
+    int32_t rest = val - (whole << REAL_BITS);
+
+    /* index into pow2_tab */
+    int32_t index = rest >> (REAL_BITS-TABLE_BITS);
+
+
+    if (val == 0)
+        return (1<<REAL_BITS);
+
+    /* leave INTERP_BITS bits */
+    index_frac = rest >> (REAL_BITS-TABLE_BITS-INTERP_BITS);
+    index_frac = index_frac & ((1<<INTERP_BITS)-1);
+
+    if (whole > 0)
+    {
+        retval = 1 << whole;
+    } else {
+        retval = REAL_CONST(1) >> -whole;
+    }
+
+    x1 = pow2_tab[index & ((1<<TABLE_BITS)-1)];
+    x2 = pow2_tab[(index & ((1<<TABLE_BITS)-1)) + 1];
+    errcorr = ( (index_frac*(x2-x1))) >> INTERP_BITS;
+
+    if (whole > 0)
+    {
+        retval = retval * (errcorr + x1);
+    } else {
+        retval = MUL_R(retval, (errcorr + x1));
+    }
+
+    return retval;
+}
+
+int32_t pow2_int(real_t val)
+{
+    uint32_t x1, x2;
+    uint32_t errcorr;
+    uint32_t index_frac;
+    real_t retval;
+    int32_t whole = (val >> REAL_BITS);
+
+    /* rest = [0..1] */
+    int32_t rest = val - (whole << REAL_BITS);
+
+    /* index into pow2_tab */
+    int32_t index = rest >> (REAL_BITS-TABLE_BITS);
+
+
+    if (val == 0)
+        return 1;
+
+    /* leave INTERP_BITS bits */
+    index_frac = rest >> (REAL_BITS-TABLE_BITS-INTERP_BITS);
+    index_frac = index_frac & ((1<<INTERP_BITS)-1);
+
+    if (whole > 0)
+        retval = 1 << whole;
+    else
+        retval = 0;
+
+    x1 = pow2_tab[index & ((1<<TABLE_BITS)-1)];
+    x2 = pow2_tab[(index & ((1<<TABLE_BITS)-1)) + 1];
+    errcorr = ( (index_frac*(x2-x1))) >> INTERP_BITS;
+
+    retval = MUL_R(retval, (errcorr + x1));
+
+    return retval;
+}
+
+/* ld(x) = ld(x*y/y) = ld(x/y) + ld(y), with y=2^N and [1 <= (x/y) < 2] */
+int32_t log2_int(uint32_t val)
+{
+    uint32_t frac;
+    uint32_t whole = (val);
+    int32_t exp = 0;
+    uint32_t index;
+    uint32_t index_frac;
+    uint32_t x1, x2;
+    uint32_t errcorr;
+
+    /* error */
+    if (val == 0)
+        return -10000;
+
+    exp = floor_log2(val);
+    exp -= REAL_BITS;
+
+    /* frac = [1..2] */
+    if (exp >= 0)
+        frac = val >> exp;
+    else
+        frac = val << -exp;
+
+    /* index in the log2 table */
+    index = frac >> (REAL_BITS-TABLE_BITS);
+
+    /* leftover part for linear interpolation */
+    index_frac = frac & ((1<<(REAL_BITS-TABLE_BITS))-1);
+
+    /* leave INTERP_BITS bits */
+    index_frac = index_frac >> (REAL_BITS-TABLE_BITS-INTERP_BITS);
+
+    x1 = log2_tab[index & ((1<<TABLE_BITS)-1)];
+    x2 = log2_tab[(index & ((1<<TABLE_BITS)-1)) + 1];
+
+    /* linear interpolation */
+    /* retval = exp + ((index_frac)*x2 + (1-index_frac)*x1) */
+
+    errcorr = (index_frac * (x2-x1)) >> INTERP_BITS;
+
+    return ((exp+REAL_BITS) << REAL_BITS) + errcorr + x1;
+}
+
+/* ld(x) = ld(x*y/y) = ld(x/y) + ld(y), with y=2^N and [1 <= (x/y) < 2] */
+real_t log2_fix(uint32_t val)
+{
+    uint32_t frac;
+    uint32_t whole = (val >> REAL_BITS);
+    int8_t exp = 0;
+    uint32_t index;
+    uint32_t index_frac;
+    uint32_t x1, x2;
+    uint32_t errcorr;
+
+    /* error */
+    if (val == 0)
+        return -100000;
+
+    exp = floor_log2(val);
+    exp -= REAL_BITS;
+
+    /* frac = [1..2] */
+    if (exp >= 0)
+        frac = val >> exp;
+    else
+        frac = val << -exp;
+
+    /* index in the log2 table */
+    index = frac >> (REAL_BITS-TABLE_BITS);
+
+    /* leftover part for linear interpolation */
+    index_frac = frac & ((1<<(REAL_BITS-TABLE_BITS))-1);
+
+    /* leave INTERP_BITS bits */
+    index_frac = index_frac >> (REAL_BITS-TABLE_BITS-INTERP_BITS);
+
+    x1 = log2_tab[index & ((1<<TABLE_BITS)-1)];
+    x2 = log2_tab[(index & ((1<<TABLE_BITS)-1)) + 1];
+
+    /* linear interpolation */
+    /* retval = exp + ((index_frac)*x2 + (1-index_frac)*x1) */
+
+    errcorr = (index_frac * (x2-x1)) >> INTERP_BITS;
+
+    return (exp << REAL_BITS) + errcorr + x1;
+}
+#endif