Recherche avancée

Sur d’autres sites (12111)

• Use fluent-ffmpeg to tell if a file is a video or audio

  11 août 2016, par afterglowlee

  I am using node-fluent-ffmpeg module in NodeJS. It is very good that fluent-ffmpeg provides functions to get the metadata of a video and audio file.

  https://github.com/schaermu/node-fluent-ffmpeg#reading-video-metadata

  I have tried on Mac OS to use the "resolution" attribute in the metadata to tell if a file is audio only or video, i.e. if both resolution.w and resolution.h are 0, then this file is an audio. This work fine on Mac OS. But some strange things happened that this doesn’t work on Windows platform (I have tried Windows 7 64bit and Windows 2008) using the latest ffmpeg. Even though I put a .mp3 file through fluent-ffmpeg,the result looks something like this :

  video:
  
  {
  
    container:'mp3',
  
    ...
  
    resolution: {w:300,h:300},
  
    resolutionSquare: {w:300,h:300},
  
    aspectString: '1:1',
  
    ...
  
  }
  
  audio:
  
  {
  
    codec:'mp3',
  
    bitrate:64,
  
    sample_rate:44100,
  
    stream:0,
  
    channels:1
  
  }

  I am not why there is a "resolution" since it is a pure audio file. So is there any solid way to find out if the file is audio only or video from the metadata ? Or should I use ffmpeg commandline to find it out ?

• How to extract frame from video and save it into memory ?

  11 septembre 2012, par Alrick

  I tried extract frame and save it as output.jpg On server side :

  f = "ffmpeg -vframes 1 -y -ss "+time+" -i ./media/somemov.mp4 ./media/output.jpg"
  
  subprocess.call(f, shell=True)

  And on client side used :

  getImage(noOfFrame);  //send the request for call the ffmpeg function from views.py
  
  
  
  document.getElementById("main").style.backgroundImage="url(http://localhost:8000/media/output.jpg#"+ new Date().getTime();+")";

  But its too slow. DIV displays always the old image becose creates image takes a long time.
  for example i want image no 3 and DIV has no 2, i want image no 4, and div has no 3.

  Is there any way how to extract frame from video (by ffmpeg) and send it into Python without creation the output.jpg file ?

  I want get the image data and send it from server (django) to webpage and set it as background-image of any DIV.

  Is the "pure data sending" solution better then mine ?

  Thanks for help.

• OpenCv VideoWriter does no write. ffmpeg library h264 errors

  22 juillet 2012, par ar ar ar

  Hi I am using OpenCv and trying to use the VideoWriter Class.

  But I get an error.

  The error is Could not find encoder for codec id 28 : Encoder not found

  Normally it must be a ffmpeg problem. How can I check which codec is with id 28 and how can I install it ?

  I think it is causing problems to my program and I cannot get the number of frames in videos right.

  update 2: :

  I installed libavcodec-extra-53 ubuntu package and now it gives other errors.

  Any ideas why ??It seems to be pure ffmpeg problem and not OpenCv problem

  libx264 @ 0x1d9bd80] broken ffmpeg default settings detected
  
  [libx264 @ 0x1d9bd80] use an encoding preset (e.g. -vpre medium)                                                                                                                               
  
  [libx264 @ 0x1d9bd80] preset usage: -vpre <speed> -vpre <profile>                                                                                                                              
  
  [libx264 @ 0x1d9bd80] speed presets are listed in x264 --help                                                                                                                                  
  
  [libx264 @ 0x1d9bd80] profile is optional; x264 defaults to high
  
  </profile></speed>