Recherche avancée

Sur d’autres sites (9371)

• What's the reason of failed to read from input stream while playing in ffmpeg

  5 février 2016, par Long Bai

  I’m using ffmpeg library in qt compiled with mingw.
  I’m going to read from rtmp stream, process video frame and write to another rtmp stream.
  I refered ffmpeg examples and developed it.
  It works well with local video stream.
  But on the internet stream, it is failed to read from input stream after a few seconds playing.
  What’s the reason of this problem ?
  Thanks for your answer.

  Here is my code.
  as you can see the following, on the line "ret=av_read_frame(m_pInContext,&inPacket)" return negative value.

  while (!m_bAborted)
  
  // read stream.
  int64_t tstart = av_gettime_relative() ;
  ret = av_read_frame(m_pInContext, &inPacket) ;
  if (ret < 0)
  SetError("Failed to read input stream.") ;
  break ;
  //continue ;
  

      if (inPacket.stream_index == m_nVideoStreamIdx)
  
      {
  
          do
  
          {
  
              // decode frame.
  
              int decodedLen = DecodePacket(m_pInFrame, &inPacket, &got_frame, 1);
  
              if (decodedLen < 0)
  
                  break;
  
  
  
              if (got_frame)
  
              {
  
                  // convert frame from yuv420 to rgb24.
  
                  if (sws_ctx1)
  
                  {
  
                      sws_scale(sws_ctx1, m_pInFrame->data, m_pInFrame->linesize, 0, m_nHeight, pRawData, arrRawDataLineSize);
  
                  }
  
  
  
                  // image processing to remove the code and ads on video.
  
                  if (m_pListener)
  
                      m_pListener->OnFrame(pRawData[0], m_nWidth, m_nHeight, arrRawDataLineSize[0], AV_PIX_FMT_RGB24, 24);
  
  
  
                  // convert from rgb24 to yuv420 to write to the output stream.
  
                  if (sws_ctx2)
  
                  {
  
                      sws_scale(sws_ctx2, pRawData, arrRawDataLineSize, 0, m_nHeight, m_dst_picture.data, m_dst_picture.linesize);
  
                  }
  
  
  
                  // increase frame number.
  
                  frameCount ++;
  
  
  
                  // write to the output stream.
  
                  ret = WriteVideoFrame(m_pOutContext, m_pOutVideoStream, &frameCount);
  
                  if (ret < 0) {
  
                      if (sws_ctx2)
  
                          sws_freeContext(sws_ctx2);
  
                      sws_ctx2 = NULL;
  
                      CloseOutputStream();
  
  
  
                      SetError("Failed to write video stream.");
  
                  }
  
  
  
                  // estimate frame time and sleep.
  
                  //clock_t tend = clock();
  
                  int64_t tend = av_gettime_relative();
  
                  frameDuration = tend-tstart;//(long)(1000 * (tend - tstart) / CLOCKS_PER_SEC);
  
                  if (1000*frameTime - frameDuration > 0)
  
                      av_usleep(1000*frameTime - frameDuration);
  
                      //QThread::msleep(frameTime - frameDuration);
  
                  tstart = tend;
  
              }
  
  
  
              // if there is another frame in the packet, then loop.
  
              inPacket.size -= decodedLen;
  
              inPacket.data += decodedLen;
  
          }while(inPacket.size > 0);
  
      }
  
  
  
      else if (m_pOutContext)
  
      {
  
  
  
      }
  
  
  
      av_packet_unref(&inPacket);
  
  }

• lavf/segment : add new option segment_clocktime_wrap_duration

  18 janvier 2016, par Marton Balint

  lavf/segment : add new option segment_clocktime_wrap_duration
  This option can force the segmenter to only start a new segment if a packet
  
  reaches the muxer within the specified duration after the segmenting clock
  
  time, which makes it more resilient to backward local time jumps, such as leap
  
  seconds or transition to standard time from daylight savings time.
  Reviewed-by : Stefano Sabatini <stefasab@gmail.com>
  
  Signed-off-by : Marton Balint <cus@passwd.hu>
  

  • [DH] doc/muxers.texi
  • [DH] libavformat/segment.c

• x86 : Avoid some bypass delays and false dependencies

  11 octobre 2015, par Henrik Gramner

  x86 : Avoid some bypass delays and false dependencies
  A bypass delay of 1-3 clock cycles may occur on some CPUs when transitioning
  
  between int and float domains, so try to avoid that if possible.
  

  • [DH] common/x86/deblock-a.asm
  • [DH] common/x86/mc-a.asm
  • [DH] common/x86/pixel-a.asm
  • [DH] common/x86/predict-a.asm
  • [DH] common/x86/sad-a.asm
  • [DH] common/x86/x86util.asm