Recherche avancée

Sur d’autres sites (14126)

• FFmpeg image2pipe write buffer wait until done

  20 décembre 2020, par Michael Joseph Aubry

  I'm extracting frames from an external source and passing it in as a buffer to FFMPEG using image2pipe and -i -

  


  const childProcess = spawn(ffmpeg, [
  "-y",
  "-f",
  "image2pipe",
  "-i",
  "-",
  "-vcodec",
  "libx264",
  "-pix_fmt",
  "yuv420p",
  output
]);


  


  Then I have a loop that does the job.

  


  for (let i = 0; i < 250; i++) {
  // ...await
}


  


  Inside the promise

  


  // ... do the job to get buffer

childProcess.stdin.write(frame); // frame === buffer

// frame done
resolve("success!");


  


  The problem is in some videos the frames jump and is janky. This is because FFmpeg is not fully done writing to the file before moving onto the next frame.

  


  Is there a way to write a buffer to a file through FFmpeg and make sure the frame is done writing before moving on ?

  


  Some more information

  


  Here is the source file https://s3.us-west-2.amazonaws.com/storycreator.v2.uploads/ckigi4kro00160vlfjmt74afp

  


  Here is the rendered file https://s3.us-west-2.amazonaws.com/storycreator.testing/607715f0-3ab9-11eb-a139-3bb84618c6c5.mp4?t=1607585343922

  


  Here are logs

  


  2020-12-10T07:28:48.942Z    0ae0c435-54d3-416f-9d1a-8ddf595a7e83    INFO    frame=  130 fps= 16 q=-1.0 Lsize=      83kB time=00:00:05.08 bitrate= 133.6kbits/s speed=0.641x    video:80kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 2.928688%


  


  2020-12-10T07:28:48.942Z    0ae0c435-54d3-416f-9d1a-8ddf595a7e83    INFO    [libx264 @ 0x640d6c0] frame I:1     Avg QP:15.47  size: 68227[libx264 @ 0x640d6c0] frame P:33    Avg QP:15.07  size:   246[libx264 @ 0x640d6c0] frame B:96    Avg QP:18.75  size:    56[libx264 @ 0x640d6c0] consecutive B-frames:  1.5%  0.0%  0.0% 98.5%


  


  2020-12-10T07:28:48.943Z    0ae0c435-54d3-416f-9d1a-8ddf595a7e83    INFO    [libx264 @ 0x640d6c0] mb I  I16..4: 24.6% 60.5% 14.9%[libx264 @ 0x640d6c0] mb P  I16..4:  0.0%  0.2%  0.0%  P16..4:  0.8%  0.0%  0.0%  0.0%  0.0%    skip:98.9%[libx264 @ 0x640d6c0] mb B  I16..4:  0.0%  0.0%  0.0%  B16..8:  0.3%  0.0%  0.0%  direct: 0.0%  skip:99.7%  L0:26.7% L1:73.3% BI: 0.0%[libx264 @ 0x640d6c0] 8x8 transform intra:62.0% inter:83.6%[libx264 @ 0x640d6c0] coded y,uvDC,uvAC intra: 49.2% 44.3% 32.1% inter: 0.0% 0.2% 0.0%[libx264 @ 0x640d6c0] i16 v,h,dc,p: 53% 38%  7%  2%[libx264 @ 0x640d6c0] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 26% 24% 29%  3%  3%  3%  5%  3%  4%[libx264 @ 0x640d6c0] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 37% 26% 14%  3%  4%  5%  4%  4%  3%[libx264 @ 0x640d6c0] i8c dc,h,v,p: 57% 27% 13%  3%[libx264 @ 0x640d6c0] Weighted P-Frames: Y:0.0% UV:0.0%[libx264 @ 0x640d6c0] ref P L0: 95.1%  1.2%  3.1%  0.6%[libx264 @ 0x640d6c0] ref B L0: 48.1% 51.3%  0.7%[libx264 @ 0x640d6c0] ref B L1: 97.3%  2.7%[libx264 @ 0x640d6c0] kb/s:125.75


  


  2020-12-10T07:28:48.944Z    0ae0c435-54d3-416f-9d1a-8ddf595a7e83    INFO    [cache @ 0x5f57940] Statistics, cache hits:0 cache misses:3551


  


• Getting "Buffer queue overflow, dropping" while concatenation in ffmpeg

  19 août 2020, par Yevhenii Molodtsov

  I have a python script that generates the FFmpeg command to concatenate the video clips.
Here is the script :

  


          ffmpeg 
        -i /tmp/media/import-2020-Aug-19-Wednesday-08-08-26/temp_eba59811-bb5c-4e36-aa7d-b1731e70432c.mp4 
        -i /tmp/media/import-2020-Aug-19-Wednesday-08-08-26/temp_2a6d7907-7cad-4f38-ae82-4b475546d3e0.mp4 
        -i /tmp/media/import-2020-Aug-19-Wednesday-08-08-26/temp_d7371ecb-01df-4199-951f-336fa475998f.mp4 
        -i /tmp/media/import-2020-Aug-19-Wednesday-08-08-26/temp_cc5de0f1-84b4-47bc-b390-45b5f8711914.mp4 
        -i /tmp/media/import-2020-Aug-19-Wednesday-08-08-26/temp_bf796e0d-135f-40fd-ab8e-4780da6973fd.mp4  
        -filter_complex 
        "[0:v]trim=start=0:end=8.032,setpts=PTS-STARTPTS[clip0];
        [1:v]trim=start=2:end=13.047,setpts=PTS-STARTPTS[clip1];
        [2:v]trim=start=2:end=13.558,setpts=PTS-STARTPTS[clip2];
        [3:v]trim=start=2:end=13.186,setpts=PTS-STARTPTS[clip3];
        [4:v]trim=start=2,setpts=PTS-STARTPTS[clip4];
        [0:v]trim=start=9.032:end=10.032,setpts=PTS-STARTPTS[out0];
        [1:v]trim=start=14.047:end=15.047,setpts=PTS-STARTPTS[out1];
        [2:v]trim=start=14.558:end=15.558,setpts=PTS-STARTPTS[out2];
        [3:v]trim=start=14.186:end=15.186,setpts=PTS-STARTPTS[out3];
        [1:v]trim=start=0:end=2,setpts=PTS-STARTPTS[in1];
        [2:v]trim=start=0:end=2,setpts=PTS-STARTPTS[in2];
        [3:v]trim=start=0:end=2,setpts=PTS-STARTPTS[in3];
        [4:v]trim=start=0:end=2,setpts=PTS-STARTPTS[in4];
        [in1]format=pix_fmts=yuva420p,fade=t=in:st=0:d=2:alpha=1[fadein1];
        [in2]format=pix_fmts=yuva420p,fade=t=in:st=0:d=2:alpha=1[fadein2];
        [in3]format=pix_fmts=yuva420p,fade=t=in:st=0:d=2:alpha=1[fadein3];
        [in4]format=pix_fmts=yuva420p,fade=t=in:st=0:d=2:alpha=1[fadein4];
        [out0]format=pix_fmts=yuva420p,fade=t=out:st=0:d=2:alpha=1[fadeout0];
        [out1]format=pix_fmts=yuva420p,fade=t=out:st=0:d=2:alpha=1[fadeout1];
        [out2]format=pix_fmts=yuva420p,fade=t=out:st=0:d=2:alpha=1[fadeout2];
        [out3]format=pix_fmts=yuva420p,fade=t=out:st=0:d=2:alpha=1[fadeout3];
        [fadein1]fifo[fadein1fifo];
        [fadein2]fifo[fadein2fifo];
        [fadein3]fifo[fadein3fifo];
        [fadein4]fifo[fadein4fifo];
        [fadeout0]fifo[fadeout0fifo];
        [fadeout1]fifo[fadeout1fifo];
        [fadeout2]fifo[fadeout2fifo];
        [fadeout3]fifo[fadeout3fifo];
        [fadeout0fifo][fadein1fifo]overlay[crossfade0];
        [fadeout1fifo][fadein2fifo]overlay[crossfade1];
        [fadeout2fifo][fadein3fifo]overlay[crossfade2];
        [fadeout3fifo][fadein4fifo]overlay[crossfade3];
        [clip0][crossfade0][clip1][crossfade1][clip2][crossfade2][clip3][crossfade3][clip4]concat=n=9[output];
         [0:a][1:a]acrossfade=d=10:c1=tri:c2=tri[A1];
         [A1][2:a]acrossfade=d=10:c1=tri:c2=tri[A2];
         [A2][3:a]acrossfade=d=10:c1=tri:c2=tri[A3];
         [A3][4:a]acrossfade=d=10:c1=tri:c2=tri[audio] " -map "[output]" -map "[audio]" /tmp/media/final/some_filename_587463a2-e3ca-436a-874b-a72e295e46ac.mp4


  


  But I get the following error : Buffer queue overflow, dropping.
I read that to solve this I need to use fifo, but I can't figure out where should I place it. I addition, it seems like fifo works only with video streams.
I would much appreciate any help !

  


• FFmpeg : concatenate m4v-files while preserving vobsub subtitles ?

  27 octobre 2020, par Nick

  I'm trying to concatenate two h.264-m4v files using ffmpeg. At least one of the source files (at least the second, not necessarily the first) contains subtitles in the vobsub format. The source files were created using HandBrake which allows for vobsub inclusion in m4v-containers. The files are exactly the same codec, same resolution, same type.

  


  Let's suppose the source files are called 1.m4v and 2.m4v. The 2.m4v-file contains the following streams which I want to have included in the output file :

  


      Stream #0:0(und): Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt470bg), 702x574 [SAR 16:15 DAR 1872:1435], 607 kb/s, 25 fps, 25 tbr, 90k tbn, 180k tbc (default)
    Metadata:
      creation_time   : 2020-10-11T16:02:38.000000Z
      handler_name    : VideoHandler
    Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 350 kb/s (default)
    Metadata:
      creation_time   : 2020-10-11T16:02:38.000000Z
      handler_name    : Stereo
    Stream #0:2(eng): Subtitle: dvd_subtitle (mp4s / 0x7334706D), 720x576, 3 kb/s (default)
    Metadata:
      creation_time   : 2020-10-11T16:02:38.000000Z
      handler_name    : SubtitleHandler
    Stream #0:3(eng): Data: bin_data (text / 0x74786574)
    Metadata:
      creation_time   : 2020-10-11T16:02:38.000000Z
      handler_name    : SubtitleHandler


  


  The two files are included in a sources.txt file containing the following lines :

  


  file '1.m4v'
file '2.m4v'


  


  The command in ffmpeg is then executed as follows :

  


  $ ffmpeg -f concat -safe 0 -i sources.txt -c copy output.m4v


  


  The resulting file correctly combines 1 and 2 with all their audio and video streams. However, the subtitle streams are omitted by ffmpeg. The output.m4v file does not contain them (Stream #0:2(eng) and Stream #0:3(eng) are not present) :

  


      Stream #0:0(und): Video: h264 (Main) (avc1 / 0x31637661), yuv420p(tv, bt470bg), 702x576 [SAR 16:15 DAR 13:10], 626 kb/s, 25 fps, 25 tbr, 90k tbn, 180k tbc (default)
    Metadata:
      handler_name    : VideoHandler
    Stream #0:1(eng): Audio: aac (LC) (mp4a / 0x6134706D), 48000 Hz, stereo, fltp, 349 kb/s (default)
    Metadata:
      handler_name    : Stereo


  


  This also happens if I alter the command to use -scodec copy as an option. In this case, the joining-process fails with the following error :

  


  [ipod @ 0x7f7faa00c000] Could not find tag for codec dvd_subtitle in stream #2, codec not currently supported in container
Could not write header for output file #0 (incorrect codec parameters ?): Invalid argument
Error initializing output stream 0:0 -- 


  


  Any idea on how to join the two files while preserving the vobsubs in the resulting m4v-container ?