Recherche avancée

Sur d’autres sites (3130)

• Burning subtitles into video with ffmpeg with specific margin not working

  17 mai 2024, par Anthony

  I am trying to build a tool where people can position their subtitles on an HTML5 video and then I will burn the subtitles into the video for them.

  


  I can easily get the position of the subtitles, in whatever format I need, that's easy. However, no combination of values that I feed to ffmpeg are working.

  


  My holy grail is an x and y offest (aka offset 5% from the left, 10% from the bottom, in terms of video height/width).

  


  To achieve this, I would like the subtitles to start in the very bottom-left corner.

  


  ffmpeg -y -i english.mp4 -vf "subtitles=english.srt:force_style='Alignment=1,OutlineColour=&H100000000,BorderStyle=3,Outline=1,Shadow=0,FontName=Arial,FontSize=24,MarginL=140,MarginV=0'"  -c:v libx264 -crf 23 -c:a copy output_video.mp4


  


  Here is something that is approaching working, but it's behaving very strangely. The video is 850x480 but when I do a margin of 140 it goes way past the midway point of the video.

  


  Am I missing something ? How can I start the margin in the bottom left, and then apply the margin I want to push it away from the left and away from the bottom ? I can do it either as a percentage of the video height/width if that API is supported or just pixels is fine too (but clearly doesn't seem to be working)

  


• The duration information of the extracted truehd audio stream is lost when demux m2ts file with ffmpeg

  23 novembre 2019, par Bo Wang

  I would like to extract an truehd audio stream from m2ts files of a blu-ray with ffmpeg. All these m2ts files actually have 7 audio streams, which can be detected by eac3to :

  M2TS, 1 video track, 7 audio tracks, 5 subtitle tracks, 1:40:03
  
  1: Chapters, 34 chapters
  
  2: h265/HEVC, 2160p24 /1.001 (16:9), 10 bits
  
  3: TrueHD/AC3 (Atmos), English, 7.1 channels, 48kHz, dialnorm: -28dB
  
     (embedded: AC3, 5.1 channels, 640kbps, 48kHz, dialnorm: -27dB)
  
  4: E-AC3, English, 7.1 channels, 1024kbps, 48kHz, dialnorm: -28dB
  
     (core: AC3, 5.1 channels, 576kbps, 48kHz, dialnorm: -28dB)
  
  5: AC3, English, 5.1 channels, 640kbps, 48kHz, dialnorm: -27dB
  
  6: AC3, English, 2.0 channels, 320kbps, 48kHz, dialnorm: -27dB
  
  7: AC3, English, 2.0 channels, 320kbps, 48kHz, dialnorm: -27dB
  
  8: AC3, Chinese, 2.0 channels, 448kbps, 48kHz
  
  9: AC3, Chinese, 5.1 channels, 384kbps, 48kHz, 300ms
  
  10: Subtitle (PGS), English
  
  11: Subtitle (PGS), Chinese
  
  12: Subtitle (PGS), Chinese
  
  13: Subtitle (PGS), Chinese
  
  14: Subtitle (PGS), Chinese

  When I use ffmpeg to do the demux work for every m2ts, problem arises.
  By runningffmpeg.exe -hide_banner -i "F:\xxx\BDMV\STREAM\00055.m2ts" -map 0:1 01.thd, an file 01.thd is obtained. However, by running ffprobe 01.thd, it is shown that this file is with fault as its duration information is absent :

  Input #0, truehd, from '0x1100-00055.thd':
  
    Duration: N/A, start: 0.000000, bitrate: N/A
  
      Stream #0:0: Audio: truehd, 48000 Hz, 7.1, s32 (24 bit)

  If I run ffmpeg.exe -hide_banner -i "F:\xxx\BDMV\STREAM\00055.m2ts", it can be found that ffmpeg shows an additional audio stream as it recognizes the embedded 5.1 AC3 stream inside the 7.1 TrueHD audio stream, as independent :

  Input #0, mpegts, from 'F:\xxx\BDMV\STREAM\00055.m2ts':
  
    Duration: 00:00:48.62, start: 4198.333333, bitrate: 50005 kb/s
  
    Program 1
  
      Stream #0:0[0x1011]: Video: hevc (Main 10) (HDMV / 0x564D4448), yuv420p10le(tv, bt2020nc/bt2020/smpte2084), 3840x2160 [SAR 1:1 DAR 16:9], 23.98 fps, 23.98 tbr, 90k tbn, 23.98 tbc
  
      Stream #0:1[0x1100]: Audio: truehd (AC-3 / 0x332D4341), 48000 Hz, 7.1, s32 (24 bit)
  
      Stream #0:2[0x1100]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, 5.1(side), fltp, 640 kb/s
  
      Stream #0:3[0x1101]: Audio: eac3 (AC-3 / 0x332D4341), 48000 Hz, 7.1, fltp, 1024 kb/s
  
      Stream #0:4[0x1102]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, 5.1(side), fltp, 640 kb/s
  
      Stream #0:5[0x1103]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, stereo, fltp, 320 kb/s
  
      Stream #0:6[0x1104]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, stereo, fltp, 320 kb/s
  
      Stream #0:7[0x1105]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, stereo, fltp, 448 kb/s
  
      Stream #0:8[0x1106]: Audio: ac3 (AC-3 / 0x332D4341), 48000 Hz, 5.1(side), fltp, 384 kb/s
  
      Stream #0:9[0x12a0]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090)
  
      Stream #0:10[0x12a1]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090)
  
      Stream #0:11[0x12a2]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090)
  
      Stream #0:12[0x12a3]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090)
  
      Stream #0:13[0x12a4]: Subtitle: hdmv_pgs_subtitle ([144][0][0][0] / 0x0090)

  How can I make ffmpeg generate correct .thd file with duration information for the stream 0:1 ?
  And if possible, how can I make ffmpeg keep the embedded 5.1 core stream as well (test result shows that this core stream is stripped out) ?

• FFMpeg concatenation with large list of inputs — how to minimize command line length

  27 août 2020, par huitlacoche

  I have a python tool that concatentates several mp4s together into a single movie file using FFMPEG run from a subprocess call. The issue is that an arbitrarily large number of inputs in the command string will hit the windows CLI character max.

  


  ffmpeg -y -i a.mp4 -i b.mp4 -i c.mp4 -f lavfi -t 0.1 -i anullsrc  -filter_complex '[0]scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2: (oh- ih)/2,setsar=1[0:v];[1]scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(owiw)/2:(oh-ih)/2,setsar=1[1:v];[2]scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow- iw)/2:(oh-ih)/2,setsar=1[2:v]; [0:v][3:a][1:v][3:a][2:v][3:a]concat=n=3:v=1:a=1[v][a]' -c:v libx264 -map [v] -map [a] output.mp4

  


  Assuming 3 input movies that may or may have audio (hence the null audio source).
This works fine assuming there are no more than 30-50 input files. Often the input files have rather long filepaths and can get long very quickly.

  


  After doing a bit of reading I tried a simple :

  


  ffmpeg -f concat -i files.txt -c copy output.mp4

  


  Which works if I can guarantee that all the files are the same codec/resolution/etc, but this cannot be guaranteed.
I'm not clear on how to translate my filter_complex to work with a text file of inputs. Finding a specific enough solution has proven elusive.

  


  I'd be perfectly fine assigning the input paths to a temporary environment variable but in my use case this doesn't seem to work either. The subprocess call doesn't seem to translate wildcard %MYVAR% calls to environment variables.

  


  I feel like there has to be an answer to this problem. If anyone can suggest a solution I would greatly appreciate it.

  


  Thanks !