ff_audio_decoder: Use parser for AAC LATM and MPEG. Fixes streams with multiple audio packets in single PES packet and audio packets splitted to multiple PES packets.

1 files changed, 86 insertions, 2 deletions

diff --git a/src/combined/ffmpeg/ff_audio_decoder.c b/src/combined/ffmpeg/ff_audio_decoder.c
index a078821f9..3b810a3fa 100644
--- a/src/combined/ffmpeg/ff_audio_decoder.c
+++ b/src/combined/ffmpeg/ff_audio_decoder.c
@@ -76,6 +76,8 @@ typedef struct ff_audio_decoder_s {
   char              *decode_buffer;
   int               decoder_ok;
 
+  AVCodecParserContext *parser_context;
+
 } ff_audio_decoder_t;
 
 
@@ -165,6 +167,24 @@ static void ff_audio_init_codec(ff_audio_decoder_t *this, unsigned int codec_typ
   this->context->codec_id    = this->codec->id;
   this->context->codec_type  = this->codec->type;
   this->context->codec_tag   = _x_stream_info_get(this->stream, XINE_STREAM_INFO_AUDIO_FOURCC);
+
+  /* Use parser for AAC LATM and MPEG.
+   * Fixes:
+   *  - DVB streams where multiple AAC LATM frames are packed to single PES
+   *  - DVB streams where MPEG audio frames do not follow PES packet boundaries
+   */
+  if (codec_type == BUF_AUDIO_AAC_LATM ||
+      codec_type == BUF_AUDIO_MPEG) {
+
+    xprintf (this->stream->xine, XINE_VERBOSITY_DEBUG,
+             "ffmpeg_audio_dec: using parser\n");
+
+    this->parser_context = av_parser_init(this->codec->id);
+    if (!this->parser_context) {
+      xprintf (this->stream->xine, XINE_VERBOSITY_LOG,
+               "ffmpeg_audio_dec: couldn't init parser\n");
+    }
+  }
 }
 
 static int ff_audio_open_codec(ff_audio_decoder_t *this, unsigned int codec_type) {
@@ -311,12 +331,55 @@ static void ff_handle_header_buffer(ff_audio_decoder_t *this, buf_element_t *buf
   this->size = 0;
 }
 
+static void ff_audio_reset_parser(ff_audio_decoder_t *this)
+{
+  /* reset parser */
+  if (this->parser_context) {
+    xprintf (this->stream->xine, XINE_VERBOSITY_DEBUG,
+             "ffmpeg_audio_dec: resetting parser\n");
+
+    pthread_mutex_lock (&ffmpeg_lock);
+    av_parser_close(this->parser_context);
+    this->parser_context = av_parser_init(this->codec->id);
+    pthread_mutex_unlock (&ffmpeg_lock);
+  }
+}
+
 static int ff_audio_decode(xine_t *xine,
                            AVCodecContext *ctx,
+                           AVCodecParserContext *parser_ctx,
                            int16_t *decode_buffer, int *decode_buffer_size,
                            uint8_t *buf, int size)
 {
   int consumed;
+  int parser_consumed = 0;
+
+  if (parser_ctx) {
+    uint8_t *outbuf;
+    int      outsize;
+
+    do {
+      int ret = av_parser_parse2(parser_ctx, ctx,
+                                 &outbuf, &outsize,
+                                 buf, size,
+                                 0, 0, 0); 
+      parser_consumed += ret;
+      buf             += ret;
+      size            -= ret;
+    } while (size > 0 && outsize <= 0);
+
+    /* nothing to decode ? */
+    if (outsize <= 0) {
+      *decode_buffer_size = 0;
+      xprintf (xine, XINE_VERBOSITY_DEBUG,
+               "ffmpeg_audio_dec: not enough data to decode\n");
+      return parser_consumed;
+    }
+
+    /* decode next packet */
+    buf  = outbuf;
+    size = outsize;
+  }
 
 #if AVAUDIO > 2
   AVPacket avpkt;
@@ -336,9 +399,13 @@ static int ff_audio_decode(xine_t *xine,
   if (consumed < 0) {
     xprintf (xine, XINE_VERBOSITY_DEBUG,
              "ffmpeg_audio_dec: error decompressing audio frame (%d)\n", consumed);
+  } else if (parser_consumed && consumed != size) {
+
+    xprintf (xine, XINE_VERBOSITY_DEBUG,
+             "ffmpeg_audio_dec: decoder didn't consume all data\n");
   }
 
-  return consumed;
+  return parser_consumed ? parser_consumed : consumed;
 }
 
 static void ff_audio_decode_data (audio_decoder_t *this_gen, buf_element_t *buf) {
@@ -376,7 +443,7 @@ static void ff_audio_decode_data (audio_decoder_t *this_gen, buf_element_t *buf)
     xine_fast_memcpy (&this->buf[this->size], buf->content, buf->size);
     this->size += buf->size;
 
-    if (buf->decoder_flags & BUF_FLAG_FRAME_END)  { /* time to decode a frame */
+    if (this->parser_context || buf->decoder_flags & BUF_FLAG_FRAME_END)  { /* time to decode a frame */
 
       offset = 0;
 
@@ -388,6 +455,7 @@ static void ff_audio_decode_data (audio_decoder_t *this_gen, buf_element_t *buf)
 
 	bytes_consumed =
           ff_audio_decode(this->stream->xine, this->context,
+                          this->parser_context,
                           (int16_t *)this->decode_buffer, &decode_buffer_size,
                           &this->buf[offset], this->size);
 
@@ -479,6 +547,7 @@ static void ff_audio_decode_data (audio_decoder_t *this_gen, buf_element_t *buf)
           audio_buffer->num_frames = bytes_to_send / 2 / this->audio_channels;
 
           audio_buffer->vpts = buf->pts;
+
           buf->pts = 0;  /* only first buffer gets the real pts */
           this->stream->audio_out->put_buffer (this->stream->audio_out,
             audio_buffer, this->stream);
@@ -507,15 +576,30 @@ static void ff_audio_reset (audio_decoder_t *this_gen) {
       this->decoder_ok = 0;
     pthread_mutex_unlock (&ffmpeg_lock);
   }
+
+  ff_audio_reset_parser(this);
 }
 
 static void ff_audio_discontinuity (audio_decoder_t *this_gen) {
+
+  ff_audio_decoder_t *this = (ff_audio_decoder_t *) this_gen;
+
+  this->size = 0;
+
+  ff_audio_reset_parser(this);
 }
 
 static void ff_audio_dispose (audio_decoder_t *this_gen) {
 
   ff_audio_decoder_t *this = (ff_audio_decoder_t *) this_gen;
 
+  if (this->parser_context) {
+    pthread_mutex_lock (&ffmpeg_lock);
+    av_parser_close(this->parser_context);
+    this->parser_context = NULL;
+    pthread_mutex_unlock (&ffmpeg_lock);
+  }
+
   if( this->context && this->decoder_ok ) {
     pthread_mutex_lock (&ffmpeg_lock);
     avcodec_close (this->context);