Recherche avancée

Sur d’autres sites (9826)

• Just set the English audio stream on output provided that it exists on video

  16 mai 2017, par parsa

  How can I have the english audio stream on output file where it exists on video ? and If not exists, I set all streams on output.
  Is there anyway for do that at one command ?
  I’m using this code

     ffmpeg -i input.mkv -f mp4 -map a -map -m:language:eng -map v -c:v copy -c:a copy output.mp4

• Stopping rq worker doesn't stop underlying ffmpeg process

  23 mars 2020, par sqr

  I am fairly new to python and rq, and have come to a point I can’t solve by myself.

  I am using ffmpeg-python to encode livestreams, this is distributed in rq workers and displayed on a web app using flask, but since the livestreams can go on forever, I need some way to stop this process while it is still in execution. Opening the terminal where the rq worker is executing the task and pressing ’q’ (ffmpeg shortcut to quit) works, and marks the job as OK, but I need to be able to do this from my web app.

  I have tried getting the worker ID and sending it a SIGKILL, this stops the worker but the task continues running, which is something I don’t understand at all. It’s as if the actual ffmpeg process was being executed somewhere else and stopping the worker didn’t stop ffmpeg. Note that I am not using ffmpeg.run_async, I am using ffmpeg.run which as far as my limited knowledge goes, should not be executed asynchronously. While the streaming is being encoded the worker is marked as busy and has the Job ID properly assigned, so I really don’t understand why, when the worker is killed, the underlying process is still in execution.

  If instead of sending a SIGKILL I send a SIGTERM, the worker says it’s waiting for a warm exit and is never closed, as the ffmpeg process is still doing it’s thing.

  One of my ideas was trying to send a ’q’ keystroke to the worker (which I have no idea how to do even though i’ve been doing some research) or trying to switch from rq to celery, that supposedly supports the cancellation of tasks that are being executed.

  This is my routes file

  @app.route('/streamings', methods=['GET', 'POST'])
  
  @login_required
  
  def streamings():
  
  ...
  
      if form2.submit_stop.data and form2.validate():
  
          conn1 = Redis.from_url('redis://')
  
          queue = rq.Queue('tasks-q', connection=Redis.from_url('redis://'))
  
          workers = rq.Worker.all(queue=queue)
  
  
  
          for worker in workers:
  
              peine = worker.get_current_job_id()
  
              if peine == form2.fld1.data:
  
                  os.kill(worker.pid, signal.SIGKILL)

  and this is my tasks file

  def restream(origin, server, stream_key):
  
      stream_server = generate_url(server, stream_key)
  
      try:
  
          stream_map = None
  
          stream1 = ffmpeg.input(get_manifest(origin), re=None)
  
          stream2 = ffmpeg.input('mosca_66.png')
  
          stream_ol = ffmpeg.overlay(stream1, stream2, x='main_w-overlay_w-50', y='50')
  
          a1 = stream1.audio
  
          stream = ffmpeg.output(stream_ol, a1, stream_server, format='flv', vcodec='libx264', acodec='aac', preset='medium', g='120', crf='23', maxrate='4M', bufsize='5M', channel_layout='stereo')
  
          print(stream.get_args())
  
          ffmpeg.run(stream)
  
      except:
  
          set_complete() 

  Any insight on possible solutions would be greatly appreciated.

  Thanks

• Get video duration of file hosted on Amazon S3

  10 janvier 2020, par Michi

  I’m starting a portal which distributes videos. The idea is to upload the videos to Amazon S3 and gather the necessary data using PHP from my server. So far everything works fine... the only thing I could not manage to get is the duration of the video :-( Could anybody give me a hint on how to accomplish it ?

  Thanks,
  Miguel

  UPDATE :

  I finally opted to do it using FFmpeg. I have already installed FFmpeg on the server and I’m now trying to execute the command in the shell prior to execute it with PHP. I’m passing it the URL from Amazon (I tried both the cloudfront URL and the S3 URL) but it says that there is not such a directory or file. I’ve seen examples on the web using external files so I expected it to work.

  The command I’m using is

  ffmpeg -i https://s3-eu-west-1.amazonaws.com/path/to/file.m4v

  Is there something I need to configure in order to use external URLs ?