Recherche avancée

Sur d’autres sites (9466)

• Modify video frames but keep their timestamp in python (maybe with ffmpeg ?)

  3 août 2018, par bendaf

  I have .ts videos with variable frame rate and I would like to modify the frames of the videos and keep the timestamps belonging to the frames. I am using python and experiencing with ffmpeg, but I am open to any solution.

  The output of

  $ffmpeg -i myvideo.ts -hide_banner

  Input #0, mpegts, from 'myvideo.ts':
  
    Duration: 00:02:37.84, start: 1.400000, bitrate: 13799 kb/s
  
    Program 1 
  
      Metadata:
  
        service_name    : Service01
  
        service_provider: FFmpeg
  
      Stream #0:0[0x100]: Video: h264 (High) ([27][0][0][0] / 0x001B), yuv420p, 3840x1080, 25 fps, 25 tbr, 90k tbn, 50 tbc

  What I have tried so far is :

  import skvideo.io
  
  
  
  outpudict = {"-map": "0",
  
               "-vcodec": "libx264",
  
               "-pix_fmt": "yuv420p",
  
               "-f": "mpegts"
  
               }
  
  
  
  videogen = skvideo.io.vreader(args.input_video)
  
  writer = skvideo.io.FFmpegWriter(filename + file_extension, outputdict=outpudict)
  
  for i, frame in enumerate(videogen):
  
      # modify the frame
  
      writer.writeFrame(frame)
  
  
  
  writer.close()

  But so far I am not able to keep the pts and dts parameters of the frames only the format and encoding.

  Note : I have tried to use skvideo.io.FFmpegReader or pass the outputdict parameter also to vreader but the script crashed because of this bug.

  Note2 : I am checking the pts of the frames with

  $ffprobe -hide_banner -show_entries packet=pts,dts myvideo.ts

  Thank you for reading through my question, any idea would be appreciated !

• ffmpeg demux into audio and video resets PTS

  30 juillet 2018, par Mukund Manikarnike

  Demuxing

  I am demuxing TS segments into audio and video as follows.

  ffmpeg -y -i input.ts -vcodec copy -an output_video.ts
  
  ffmpeg -y -i input.ts -acodec copy -vn output_audio.aac

  Inspecting Input

  The start_pts and start_time on input.ts are as shown below. I was able to inspect these values using ffprobe -show_streams -print_format json input.ts

  "start_pts": 8306558438,
  
  "start_time": "92295.093756",

  Inspecting output video

  The output .ts has some default start_pts and start_time values as shown below. These were also obtained using the same ffprobe command as indicated above.

  "start_pts": 126000,
  
  "start_time": "1.400000",

  Inspecting output audio

  The same ffprobe command on output_audio.aac shows that the output aac has invalid codec_tag and codec_tag_string as shown below. The start_pts and start_time are not present in the output_audio.aac.

  "codec_tag_string": "[0][0][0][0]", (should have been [15][0][0][0])
  
  "codec_tag": "0x0000", (should have been 0xf000)

  Questions

  1. Wondering if this difference in the start_pts, start_time, codec_tag is expected ?
  2. If it is expected, what can I do to ensure that the all of these parameters get retained on the output ?
  3. If it is not expected, is there some more information I can share to track this down ?

  Note

  There were other outputs that I found inconsistent in the ffprobe command for the output_audio.aac like duration etc.. I shared what I thought are most valuable at this point. If required I can share complete outputs from all of the above executions.

  [EDIT 07/30/2018 - 08:00 MST]
  logs for ffmpeg -y -i input.ts -vcodec copy -an output_video.ts -acodec copy -vn output_audio.aacare as shown below.

  ffmpeg version 4.0.2 Copyright (c) 2000-2018 the FFmpeg developers
  
    built with Apple LLVM version 9.0.0 (clang-900.0.39.2)
  
    configuration: --prefix=/usr/local/Cellar/ffmpeg/4.0.2 --enable-shared --enable-pthreads --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-gpl --enable-ffplay --enable-frei0r --enable-libass --enable-libfdk-aac --enable-libfreetype --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopus --enable-librtmp --enable-libspeex --enable-libtheora --enable-libvorbis --enable-libvpx --enable-libx264 --enable-libxvid --enable-opencl --enable-videotoolbox --disable-lzma --enable-libopenjpeg --disable-decoder=jpeg2000 --extra-cflags=-I/usr/local/Cellar/openjpeg/2.3.0/include/openjpeg-2.3 --enable-nonfree
  
    libavutil      56. 14.100 / 56. 14.100
  
    libavcodec     58. 18.100 / 58. 18.100
  
    libavformat    58. 12.100 / 58. 12.100
  
    libavdevice    58.  3.100 / 58.  3.100
  
    libavfilter     7. 16.100 /  7. 16.100
  
    libavresample   4.  0.  0 /  4.  0.  0
  
    libswscale      5.  1.100 /  5.  1.100
  
    libswresample   3.  1.100 /  3.  1.100
  
    libpostproc    55.  1.100 / 55.  1.100
  
  [mpegts @ 0x7f88ed803000] start time for stream 0 is not set in estimate_timings_from_pts
  
  Input #0, mpegts, from 'i7h9456s_media_46185.ts':
  
    Duration: 00:00:06.05, start: 86216.852667, bitrate: 2898 kb/s
  
    Program 1 
  
      Stream #0:0[0x102]: Data: timed_id3 (ID3  / 0x20334449)
  
      Stream #0:1[0x100]: Video: h264 (Constrained Baseline) ([27][0][0][0] / 0x001B), yuv420p(tv, smpte170m, progressive), 640x360 [SAR 1:1 DAR 16:9], 29.97 fps, 29.97 tbr, 90k tbn, 59.94 tbc
  
      Stream #0:2[0x101]: Audio: aac (LC) ([15][0][0][0] / 0x000F), 44100 Hz, stereo, fltp, 190 kb/s
  
  Output #0, mpegts, to '../output_video.ts':
  
    Metadata:
  
      encoder         : Lavf58.12.100
  
      Stream #0:0: Video: h264 (Constrained Baseline) ([27][0][0][0] / 0x001B), yuv420p(tv, smpte170m, progressive), 640x360 [SAR 1:1 DAR 16:9], q=2-31, 29.97 fps, 29.97 tbr, 90k tbn, 90k tbc
  
  Output #1, adts, to '../output_audio.aac':
  
    Metadata:
  
      encoder         : Lavf58.12.100
  
      Stream #1:0: Audio: aac (LC) ([15][0][0][0] / 0x000F), 44100 Hz, stereo, fltp, 190 kb/s
  
  Stream mapping:
  
    Stream #0:1 -> #0:0 (copy)
  
    Stream #0:2 -> #1:0 (copy)
  
  Press [q] to stop, [?] for help
  
  frame=  180 fps=0.0 q=-1.0 Lsize=    2088kB time=00:00:06.03 bitrate=2833.8kbits/s speed= 904x    
  
  video:1918kB audio:142kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 1.349750%

• ffmpeg memory leak in the avcodec_open2 method

  21 août 2019, par unresolved_external

  I’ve developed an application which handles live video stream. The problem is that it should run as a service and over time I am noticing some memory increase. When I check the application with valgrind - it did not find any leak related issues.
  So I’ve check it with google profile tools. This is a result(substracting the one of the first dumps from the latest) after approximately 6 hour run :

     30.0  35.7%  35.7%     30.0  35.7% av_malloc
  
      28.9  34.4%  70.2%     28.9  34.4% av_reallocp
  
      24.5  29.2%  99.4%     24.5  29.2% x264_malloc

  When I check the memory on the graph I see, that these allocations are related to avcodec_open2. The client code is :

  `           g_EncoderMutex.lock();
  
              ffmpeg_encoder_start(OutFileName.c_str(), AV_CODEC_ID_H264, m_FPS, width, height);
  
              for (pts = 0; pts < VideoImages.size(); pts++) {                
  
                  m_frame->pts = pts;
  
                  ffmpeg_encoder_encode_frame(VideoImages[pts].RGBimage[0]);
  
              }
  
              ffmpeg_encoder_finish();
  
              g_EncoderMutex.unlock()

  The ffmpeg_encoder_start method is :

   void VideoEncoder::ffmpeg_encoder_start(const char *filename, int codec_id, int fps, int width, int height)
  
          {
  
              int ret;
  
              m_FPS=fps;
  
              AVOutputFormat * fmt = av_guess_format(filename, NULL, NULL);
  
              m_oc = NULL;
  
              avformat_alloc_output_context2(&m_oc, NULL, NULL, filename);
  
  
  
              m_stream = avformat_new_stream(m_oc, 0);
  
              AVCodec *codec=NULL;
  
  
  
              codec =  avcodec_find_encoder(codec_id);    
  
              if (!codec) 
  
              {
  
                  fprintf(stderr, "Codec not found\n");
  
                  return; //-1
  
              }
  
  
  
              m_c=m_stream->codec;
  
  
  
              avcodec_get_context_defaults3(m_c, codec);
  
  
  
              m_c->bit_rate = 400000;
  
              m_c->width = width;
  
              m_c->height = height;
  
              m_c->time_base.num = 1;
  
              m_c->time_base.den = m_FPS;
  
              m_c->gop_size = 10;
  
              m_c->max_b_frames = 1;
  
              m_c->pix_fmt = AV_PIX_FMT_YUV420P;
  
              if (codec_id == AV_CODEC_ID_H264)
  
                  av_opt_set(m_c->priv_data, "preset", "ultrafast", 0);
  
  
  
              if (m_oc->oformat->flags & AVFMT_GLOBALHEADER) 
  
                  m_c->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
  
              avcodec_open2( m_c, codec, NULL );
  
  
  
              m_stream->time_base=(AVRational){1, m_FPS};
  
  
  
              if (avio_open(&m_oc->pb, filename, AVIO_FLAG_WRITE) < 0)
  
              {
  
                  printf( "Could not open '%s'\n", filename);
  
                  exit(1);
  
              }            
  
  
  
              avformat_write_header(m_oc, NULL);
  
              m_frame = av_frame_alloc();
  
              if (!m_frame) {
  
                  printf( "Could not allocate video frame\n");
  
                  exit(1);
  
              }
  
              m_frame->format = m_c->pix_fmt;
  
              m_frame->width  = m_c->width;
  
              m_frame->height = m_c->height;
  
              ret = av_image_alloc(m_frame->data, m_frame->linesize, m_c->width, m_c->height, m_c->pix_fmt, 32);
  
              if (ret < 0) {
  
                  printf("Could not allocate raw picture buffer\n");
  
                  exit(1);
  
              }
  
          }

  The ffmpeg_encoder_encode_frame is :

  void VideoEncoder::ffmpeg_encoder_encode_frame(uint8_t *rgb) 
  
  {
  
      int ret, got_output;
  
      ffmpeg_encoder_set_frame_yuv_from_rgb(rgb);
  
      av_init_packet(&m_pkt);
  
      m_pkt.data = NULL;
  
      m_pkt.size = 0;
  
  
  
      ret = avcodec_encode_video2(m_c, &m_pkt, m_frame, &got_output);
  
      if (ret < 0) {
  
          printf("Error encoding frame\n");
  
          exit(1);
  
      }
  
      if (got_output) 
  
      {
  
  
  
           av_packet_rescale_ts(&m_pkt,
  
                          (AVRational){1, m_FPS}, m_stream->time_base);
  
          m_pkt.stream_index = m_stream->index;
  
          int ret = av_interleaved_write_frame(m_oc, &m_pkt);
  
  
  
          av_packet_unref(&m_pkt);
  
  
  
      }
  
  
  
  }

  ffmpeg_encoder_finish code is :

  void VideoEncoder::ffmpeg_encoder_finish(void) 
  
          {
  
              int got_output, ret;
  
  
  
              do {
  
  
  
                  ret = avcodec_encode_video2(m_c, &m_pkt, NULL, &got_output);
  
                  if (ret < 0) {
  
                      printf( "Error encoding frame\n");
  
                      exit(1);
  
                  }
  
                  if (got_output) {
  
  
  
                      av_packet_rescale_ts(&m_pkt,
  
                                  (AVRational){1, m_FPS}, m_stream->time_base);
  
                      m_pkt.stream_index = m_stream->index;
  
                      int ret = av_interleaved_write_frame(m_oc, &m_pkt);
  
  
  
                      av_packet_unref(&m_pkt);
  
                  }
  
              } while (got_output);
  
  
  
              av_write_trailer(m_oc);
  
              avio_closep(&m_oc->pb);
  
  
  
              avformat_free_context(m_oc);
  
  
  
              av_freep(&m_frame->data[0]);
  
              av_frame_free(&m_frame);
  
  
  
              av_packet_unref(&m_pkt);
  
              sws_freeContext(m_sws_context);
  
          }

  This code runs multiple times in the loop.
  So my question is - what am I doing wrong ? maybe ffmpeg is using some kind of internal buffering ? If so, how to disable it ? Because such an increase in memory usage is unacceptable at all.