Recherche avancée

Sur d’autres sites (6251)

• Split a movie into 1000+ shots using PyAV in a single pass ?

  3 mai 2019, par Andrew Klaassen

  I need to split a 44 minute MP4 into 1000 shots (i.e. separate MP4s) with ffmpeg. I want to do it quickly (i.e. in a single pass rather than 1000 passes), I need perfect frame accuracy, and I need to do it in Windows.

  The Windows command-line length limit is stopping me from doing this, and I’m wondering if someone could show me an example of how to do this using a library like PyAV or Avpy. (Libraries like ffmpeg-python and ffmpy won’t help, since they simply construct an ffmpeg command line and run it, leading to the same Windows command-line length issue that I already have.)

  After much testing and gnashing of teeth, I’ve learned that the only way to get perfect frame accuracy from ffmpeg, 100% of the time, is to use the "select" filter. ("-ss" in the newest versions of ffmpeg is frame accurate 99% of the time ; unfortunately, that’s not good enough for this application.)

  There are two ways to use "select" for this. There’s the slow way, which I’m doing now, and which requires having ffmpeg open the file 1000 times :

  for (start, end, name) in shots:
  
      audio_start = start / frame_rate
  
      audio_end = end + 1 / frame_rate
  
      cmd = [
  
          path_to_ffmpeg,
  
          '-y',
  
          '-i', input_movie,
  
          '-vf', r'select=between(n\,%s\,%s),setpts=PTS-STARTPTS' % (start, end),
  
          '-af', 'atrim=%s:%s,asetpts=PTS-STARTPTS' % (audio_start, audio_end),
  
          '-c:v', 'libx264',
  
          '-c:a', 'aac',
  
          '-write_tmcd', '0',
  
          '-g', '1',
  
          '-r', str(frame_rate),
  
          name + '.mp4',
  
          '-af', 'atrim=%s:%s' % (audio_start, audio_end),
  
          name + '.wav',
  
      ]
  
      subprocess.call(cmd)

  And there’s the fast way, which causes the Windows command line to explode when there are too many shots. The long command line leads to a failure to run :

  cmd = [
  
      path_to_ffmpeg,
  
      '-y',
  
      '-i',
  
      input_movie,
  
  ]
  
  for (start, end, name) in shots:
  
      audio_start = start / frame_rate
  
      audio_end = end + 1 / frame_rate
  
      cmd.extend([
  
          '-vf', r'select=between(n\,%s\,%s),setpts=PTS-STARTPTS' % (start, end),
  
          '-af', 'atrim=%s:%s,asetpts=PTS-STARTPTS' % (audio_start, audio_end),
  
          '-c:v', 'libx264',
  
          '-c:a', 'aac',
  
          '-write_tmcd', '0',
  
          '-g', '1',
  
          '-r', str(frame_rate),
  
          name + '.mp4',
  
          '-af', 'atrim=%s:%s' % (audio_start, audio_end),
  
          name + '.wav',
  
      ]
  
  subprocess.call(cmd)

  I’ve looked through the documentation of PyAV and Avpy, but I haven’t been able to figure out whether the second form of my function is something I could do there, or how I’d go about doing it. If it is possible, would someone be able to write a function equivalent to my second function, using either library ?

• Cutting movie with ffmpeg result in audio/video desync

  15 mai 2020, par T4ng10r

  I've concate long ago set of movies taken during some lecture. Now I want to cut them for each question/answer.

  



  I do it like this.

  



  


  ffmpeg -ss 00:00:34.7 -t 00:10:44.6 -y -i input_movie.mp4 -vcodec copy -acodec copy output_1.mp4

  
 


  ffmpeg -ss 00:11:22.2 -y -i input_movie.mp4 -vcodec copy -acodec copy output_2.mp4

  


  



  Yet, for the second part I can't set proper starting point so audio and video would be in sync.
  
Usually I could fix it with small tweeks in cut start time (like .1, .2, and so on). For this case this doesn't work.
  
When I play second cut in mplayer video is few second behind audio (where audio is cut properly). When I jump forward and back - all is again in sync.

  



  Where's the problem ? How to fix it ?

  


• Handling an arbitrary number of start and stop time pairings to cut a movie file down

  9 juin 2019, par Kieran

  I am writing a function that takes a list of tuples and a file path string as arguments and outputs a cut down video that only includes the frames that fall inside the start/stop pairings provided.

  I’m getting stuck because I am not sure whether the .trim() method of the ’infile’ object is altering the existing object or or creating a new one or doing something else entirely.

  the list of start/stop frame pairings can be arbitrarily long, every example I have found has been for a specific number of start and stop pairings and I can’t find anything describing what data structure needs to be passed back to ffmpeg.concat().

  My code is displayed below :

  import ffmpeg
  
  
  
  frameStamps = [(50,75),(120,700),(1250,1500)]
  
  videoFilePath = 'C:/Users/Kieran/Videos/testMovie.mp4'
  
  outputFolder = 'C:/Users/Kieran/Videos/'
  
  
  
  def slice_video(frameStamps, videoFilePath, outputFolder):
  
  
  
      originalFile = ffmpeg.input(videoFilePath)
  
  
  
      for stamp in frameStamps:
  
          ffmpeg.concat(originalFile.trim(start_frame=stamp[0], end_frame=stamp[1]))
  
  
  
  
  
      ffmpeg.output(outputFolder + 'testoutput.mp4')
  
      ffmpeg.run() 
  
  
  
  slice_video(frameStamps, videoFilePath, outputFolder)

  Now I am able to get the following when I individually print out originalFile.trim() which are getting recognised in the console as "FilterableStream" objects

  trim(end_frame=75, start_frame=50)[None] <29b4fb0736ec>
  
  trim(end_frame=700, start_frame=120)[None] <c66c4e1a48f5>
  
  trim(end_frame=1500, start_frame=1250)[None] &lt;13e0697a5288>  
  
  </c66c4e1a48f5>

  and I have tried passing them back as a list, dictionary and tuple and haven’t been able to get it working

  Output Errors :

    File "C:/Users/Kieran/Example.py", line 21, in slice_video
  
  ffmpeg.output(outputFolder + 'testoutput.mp4')
  
  
  
    File "C:\ProgramData\Anaconda3\lib\site-packages\ffmpeg\_ffmpeg.py", line 94, in output
  
  return OutputNode(streams, output.__name__, kwargs=kwargs).stream()
  
  
  
    File "C:\ProgramData\Anaconda3\lib\site-packages\ffmpeg\nodes.py", line 282, in __init__
  
  kwargs=kwargs
  
  
  
    File "C:\ProgramData\Anaconda3\lib\site-packages\ffmpeg\nodes.py", line 170, in __init__
  
  self.__check_input_len(stream_map, min_inputs, max_inputs)
  
  
  
    File "C:\ProgramData\Anaconda3\lib\site-packages\ffmpeg\nodes.py", line 149, in __check_input_len
  
  raise ValueError('Expected at least {} input stream(s); got {}'.format(min_inputs, len(stream_map)))
  
  
  
  ValueError: Expected at least 1 input stream(s); got 0