Recherche avancée

Sur d’autres sites (4122)

• How to optimize ffmpeg w/ x264 for multiple bitrate output files

  10 octobre 2013, par Jonesy

  The goal is to create multiple output files that differ only in bitrate from a single source file. The solutions for this that were documented worked, but had inefficiencies. The solution that I discovered to be most efficient was not documented anywhere that I could see. I am posting it here for review and asking if others know of additional optimizations that can be made.

  Source file       MPEG-2 Video (Letterboxed) 1920x1080 @>10Mbps
  
                    MPEG-1 Audio @ 384Kbps
  
  Destiation files  H264 Video 720x400 @ multiple bitrates
  
                    AAC Audio @ 128Kbps
  
  Machine           Multi-core Processor

  The video quality at each bitrate is important so we are running in 2-Pass mode with the 'medium' preset

  VIDEO_OPTIONS_P2 = -vcodec libx264 -preset medium -profile:v main -g 72 -keyint_min 24 -vf scale=720:-1,crop=720:400

  The first approach was to encode them all in parallel processes

  ffmpeg -y -i $INPUT_FILE $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 250k -threads auto -f mp4 out-250.mp4 &
  ffmpeg -y -i $INPUT_FILE $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 500k -threads auto -f mp4 out-500.mp4 &
  ffmpeg -y -i $INPUT_FILE $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 700k -threads auto -f mp4 out-700.mp4 &
  

  The obvious inefficiencies are that the source file is read, decoded, scaled, and cropped identically for each process. How can we do this once and then feed the encoders with the result ?

  The hope was that generating all the encodes in a single ffmpeg command would optimize-out the duplicate steps.

  ffmpeg -y -i $INPUT_FILE \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 250k -threads auto -f mp4 out-250.mp4 \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 500k -threads auto -f mp4 out-500.mp4 \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 700k -threads auto -f mp4 out-700.mp4

  However, the encoding time was nearly identical to the previous multi-process approach. This leads me to believe that all the steps are again being performed in duplicate.

  To force ffmpeg to read, decode, and scale only once, I put those steps in one ffmpeg process and piped the result into another ffmpeg process that performed the encoding. This improved the overall processing time by 15%-20%.

  INPUT_STREAM="ffmpeg -i $INPUT_FILE -vf scale=720:-1,crop=720:400 -threads auto -f yuv4mpegpipe -"
  
  
  
  $INPUT_STREAM | ffmpeg -y -f yuv4mpegpipe -i - \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 250k -threads auto out-250.mp4 \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 500k -threads auto out-500.mp4 \
  
  $AUDIO_OPTIONS_P2 $VIDEO_OPTIONS_P2 -b:v 700k -threads auto out-700.mp4

  Does anyone see potential problems with doing it this way, or know of a better method ?

• libav/ffmpeg : avcodec_decode_video2() returns -1 when separating demultiplexing and decoding

  26 avril 2013, par unbekannt

  I'm using libav (from a C++ program on Linux and Windows) to decode video streams from a file, which works fine (decoding various formats like H264 and MPEG2) using avformat_open_input(), av_read_frame() and avcodec_decode_video2().

  Now I have to separate demultiplexing and decoding. One class will call avformat_open_input() and av_read_frame() and then pass the AVPackets into a queue that is read by another class. There I use avcodec_alloc_context3() to get the AVCodecContext needed for avcodec_decode_video2(). I've tested that with a MPEG2 video stream and it works.

  Problems arise if I try to decode a H264 stream : avcodec_decode_video2() always returns -1 and outputs "no frame". I understand that additional data (SPS/PPS) is needed to decode this stream, so I've tried to replicate the original AVCodecContext from the demultiplexer in the decoder, but it won't work :

  • Copying the content of the extradata field and setting all other values that differ from the default ones in the decoder : -1 is returned
  • Using the same context (i.e. passing along the pointer) results in a crash

  I also tried to set CODEC_FLAG2_CHUNKS. avcodec_decode_video2() then always returns packet.size - 3 (??) and frameFinished is never set to 1.

  In my opinion I have a general problem here that will arise whenever settings from the original CodecContext are needed to decode the AVPackets. I'd be grateful for any hints on how to solve that problem !

  EDIT : Sometimes writing down your problem helps solving it... Using a copy of the context struct (avcodec_copy_context) and opening the codec only after receiving the copy results in decoded frames. Does anyone know if that is safe or the best way to do it ?

• Which is a better expression for this if statement ?

  3 novembre 2012, par Philos

  I've been told of a patch for ffmpeg that fixes an issue when streaming to FLV players.

  Around line 2314 of rtmpproto.c there is the following expression :

  if (rt->flv_header_bytes < 11)
  
              break;

  The fix is to change it to include this additional requirement :

  if (rt->flv_header_bytes < 11 && !rt->flv_off)
  
              break;

  However I'm curious about the logic of this statement. The first statement checks to see if the FLV header is less than 11 bytes, but wouldn't it be more efficient to check to see if the flv stream is turned on first before it checks how many bytes big the header is ?

  if (!rt->flv_off && rt->flv_header_bytes < 11)
  
              break;

  Alternatively what about changing the statement to be like so ?

  if (rt->flv_off || (!rt->flv_off && rt->flv_header_bytes < 11)
  
                  break;

  Would that be more efficient or could it break the code somehow ?