Recherche avancée

Sur d’autres sites (12407)

• ffmpeg process stops for no reason at random position

  29 septembre 2021, par Peter Hammi

  I have a encoding script which can re-encode/repack h264, h265, acc, ac3, mp3, flac etc. using ffmpeg version 4.4 into HLS (http-live-stream). Actually the script is working pretty awesome and I have very nice results but for some reason my conversion process breaks if it simply runs long.

  


  Basically the ffmpeg process gets executed within docker using a
simple command call like so :

  


  def exec_command(string):
    """
    Shell command interface

    Returns returnCode, stdout, stderr
    """
    log('DEBUG', f'[Command] {string}')
    output = run(string, shell=True, check=True, capture_output=True)
    return output.returncode, output.stdout, output.stderr


  


  I sadly can't get to much into detail as the code is closed source and more than 1000 lines long, anyways the string parameter that gets passed and containing the command to be executed looks like this :

  


  command += f' -map 0:{stream["index"]} {build_command_encode(stream, job["config"])} -map_metadata -1 -map_chapters -1 -f hls -hls_time 6 -hls_list_size 0 -hls_segment_filename "{path}/f-%04d.m4s" -hls_fmp4_init_filename "init-{name}.m4s" -hls_segment_type fmp4 -movflags frag_keyframe -hls_flags independent_segments "{path}/master.m3u8"'


  


  Can maybe someone imagine why the ffmpeg process break for no reason ?

  


  Thanks in advance

  


• Read and Save rtsp stream using FFMPEG using python with less memory size

  25 août 2022, par Vishak Raj

  I am trying to read a rtsp stream and save in a file, for that I am using the ffmpeg in python

  


  import ffmpeg

stream = ffmpeg.input(rtsp_link, t=10)
print(stream)

file = stream.output("test.mp4")
testfile = file.run()#capture_stdout=True, capture_stderr=True


  


  but this save the video file in high space, for 10 second video, the file occupies around 3 Mb, how to reduce the file size

  


  thanks

  


• How to not add AdaptationSet for audio in cmd file if audio stream is not present in input video file

  15 juin 2019, par Saurabh

  I am using ffmpeg-dash to create mpd files. I am using following command to generate the same :

  ffmpeg -i input/video.mov \
  
          -map 0:v:0 -map 0:v:0 -map 0:v:0 -map 0:a\?:0 \
  
          -b:v:0 350k -c:v:0 libx264 -filter:v:0 "scale=320:-1" \
  
          -b:v:1 1000k -c:v:1 libx264 -filter:v:1 "scale=640:-1" \
  
          -b:v:2 3000k -c:v:2 libx264 -filter:v:2 "scale=1280:-1" \
  
          -use_timeline 1 -use_template 1 -window_size 6  -adaptation_sets "id=0,streams=v id=1,streams=a"  \
  
          -seg_duration 2 -threads 16 -hls_playlist true -f dash output/ffmpeg/output.mpd

  This works with video which has audio streams, but for the case there is no audio stream in the video file, I get empty entry for audio in mpd files like following, which is not valid and fails while playing in Exo Player in Android.

  <adaptationset contenttype="audio" segmentalignment="true" bitstreamswitching="true">
  
  </adaptationset>

  I tried adding ? in option adaptation_sets like : id=0,streams=v id=1,streams=a?, however that also adds above line.

  One solution is to run different commands for different videos by first checking whether there is audio stream or not, but would be better if there is some option which add audio adaptationSet only if audio stream is present in the input file.