Recherche avancée

Sur d’autres sites (8734)

• how to fill AvFrame pixel by pixel

  18 mai 2017, par masoud khanlo

  i try to encode opengl output to a mp4 file using ffmpeg but when i want to simply fill AvFrame with pixel colors with this code snippet :

  for (y = 0; y < c->height; y++) {
  
  for (x = 0; x < c->width; x++) {
  
      int offset = 3 * (x + y * c->width);
  
  
  
      if (x > 2* c->width / 3 && x width )
  
      {
  
          frame->data[0][offset + 0] = 0; // B
  
          frame->data[0][offset + 1] = 0; // G
  
          frame->data[0][offset + 2] = 255; // R
  
      } 
  
      else if(x < 2 * c->width / 3 && x > c->width / 3) {
  
          frame->data[0][offset + 0] = 0; // B
  
          frame->data[0][offset + 1] = 255; // G
  
          frame->data[0][offset + 2] = 0; // R
  
      }
  
      else {
  
          frame->data[0][offset + 0] = 255; // B
  
          frame->data[0][offset + 1] = 0; // G
  
          frame->data[0][offset + 2] = 0; // R
  
      }
  
  }

  then i have this video output :

  

  and i had this options for ffmpeg :

  codec : libx264rgb
  & pix_fmt : AV_PIX_FMT_BGR24

  in fact i expect something like this video output :

  

  i’m also new in ffmpeg and i realy dont know advanced stuffs about video encoding ,so anybody knows what is problem ?

• Ffmpeg libav with multiple inputs crashes when cleaning memory

  2 mars 2021, par Expressingx

  I'm streaming multiple cameras with ffmpeg libav

  


  for (int i = 0; i < CameraList.Count; i++)
            {
                var camera = CameraList[i];
                    AVDictionary* optionalOpts;
                    if (!string.IsNullOrEmpty(camera.InputResolution))
                        ffmpeg.av_dict_set(&optionalOpts, "video_size", camera.InputResolution, 0);

                    ffmpeg.av_dict_copy(&optionalOpts, options, 0);

                    AVInputFormat* inputFormat = null;

                    if (!string.IsNullOrEmpty(camera.Format))
                    {
                        inputFormat = ffmpeg.av_find_input_format(camera.Format);
                        if (inputFormat == null)
                            // exception
                    }

                    if (ffmpeg.avformat_open_input(&pInputFmtCtx, camera.Url, inputFormat, &optionalOpts) != 0)
                    {
                        if (ffmpeg.avformat_open_input(&pInputFmtCtx, camera.Url, inputFormat, &options) != 0)
                            //exception
                    }
            }


  


  Every camera is streamed to own file, having its own output AVFormatContext. Everything is working fine till the recording is ended and the memory has to be freed. And the problem is here

  


  if (InputFormatContext != null)
            {
                fixed (AVFormatContext** p = &InputFormatContext)
                    ffmpeg.avformat_close_input(p);

                InputFormatContext = null;
            }


  


  If I use only 1 camera everything is fine, but if they are two the application crashes. And I don't know why.

  


• Reading RTMP streams using FFMPEG returns AVERROR_EOF randomly

  9 octobre 2013, par user2628781

  I am using FFMpeg to receive RTMP streams. This logic is placed in a custom video player I am building.

  I managed to successfully connect to the RTMP stream and display the video correctly. However, after a period in time, the stream terminates prematurely with an AVERROR_EOF when I perform av_read_frame().

  This indicates to me that the file has ended so my demux, video and audio threads terminates thinking that the video has ended.
  However, the video playback hasn't yet reached its end (in the case of a file streamed through rtmp) or from a live stream (which runs forever). The EOFs are received very randomly so it may run for say 7 min before this occurs or after 3 mins.

  Is this a characteristic behaviour of RTMP or am I doing something incorrectly ?
  I am also having a similar problem with Http Live Streams using FFMpeg.

  A small snippet of the code is provided below :

   AVPacket packet;
  
  //start timeout timer and timeout
  
  __interrupt_timer.start();
  
  
  
  int ret = av_read_frame(format_context, &packet); //0: ok, <0: error/end
  
  
  
  
  
  //invalidate the timer
  
  __interrupt_timer.invalidate();
  
  
  
  if (ret != 0) {
  
      if (ret == AVERROR_EOF) { //end of file. FIXME: why no eof if replaying by seek(0)?            
  
          if (!eof) {
  
              eof = true;
  
              started_ = false;
  
              pkt->data = QByteArray(); //flush
  
              pkt->markEnd();
  
              qDebug("End of file. %s %d", __FUNCTION__, __LINE__);
  
              emit finished();
  
              return true;
  
          }
  
          pkt->data = QByteArray(); //flush
  
          //return true;
  
          return false; //frames after eof are eof frames
  
      } else if (ret == AVERROR_INVALIDDATA) {
  
          qWarning("AVERROR_INVALIDDATA");
  
      } else if (ret == AVERROR(EAGAIN)) {
  
          return true;
  
      }
  
      qWarning("[AVDemuxer] error: %s", av_err2str(ret));
  
      return false;
  
  }