Recherche avancée

Sur d’autres sites (6840)

• Boto3 Video Upload 0 Bytes from Heroku

  14 juillet 2017, par genghiskhan

  I have a small Flask api that takes a video and an image, overlays the image on the video and uploads the result to Amazon S3. I am using ffmpeg to do the actual overlaying. Here is that code :

  command = "ffmpeg -i {0} -i {1} -filter_complex \"overlay=0:0\" {2}".format(background_name, overlay_name, output_name)
  
  subprocess.getoutput(command)

  Then I simply upload it via Boto3 :

  s3.upload_file(output_name, VIDEO_BUCKET_NAME, output_name)

  This code works fine when I run on localhost ; however, when I test in while deployed to Heroku, it always uploads a file with 0 bytes. I suspect that it may be a problem with Heroku’s transient filesystem, but the file is being used immediately after it is created.

• Converting AWS Polly Audio Stream with fluent-ffmpeg

  27 juillet 2017, par Joel

  I am trying to convert an audio stream from Amazon AWS Polly in Node.js using fluent-ffmpeg. The documentation says that I can convert a stream, which is what the output of Polly provides, but I am getting an "Invalid input" error.

  polly.synthesizeSpeech(pollyParams, function (err, data) {
  
  if (err) {
  
      console.log(err)
  
  } else {
  
      console.log('Audio')
  
      console.log(data)
  
      ffmpeg().input(data.AudioStream).inputOptions(['-ac 2', '-codec:a libmp3lame', '-b:a 48k', '-ar 16000']) 
  
  }  

  Results in :

  AudioStream : }
  2017-07-27T14:07:09.335Z dd75614c-72d4-11e7-b7cd-5d4425c782fc Error : Invalid input
  at FfmpegCommand.proto.mergeAdd.proto.addInput.proto.input (/var/task/node_modules/fluent-ffmpeg/lib/options/inputs.js:34:15)

  I know the output from Polly is a valid audio stream, because I am able to save it to an S3 bucket. I would prefer to convert the stream before saving to S3, rather than saving it, picking it up from S3, converting it, and then saving it again.

  Thanks for your help !

• How to merge multiple video files into one video only

  1er août 2017, par shamaleyte

  I have multiple webm video files of a conference call.
  However, each participant joined the call at a different time which resulted in the fact that each video file has a different startTimeOffset values.

  Video Start Time

  Video1 : 00:00

  Video2 : 00:10

  Video3 : 01:40

  My purpose is to play back this conference. However, I do not record the conference as 1 video, it is recorded with multiple video files, instead.
  Is there any best practice to stitch such videos accordingly ?
  Maybe by ffmpeg library ?

  There is also a paid solution ; https://aws.amazon.com/about-aws/whats-new/2016/11/amazon-elastic-transcoder-adds-support-for-clip-stitching/ ) to merge video fragments to a single clip, this will make the client-side much simpler. But any free practice of doing it ?

  The expected outcome is to have 1 video showing 3 videos in a grid.
  When ffmpeg stitches the videos, it should consider their start time values properly so that the videos are played accordingly.