Recherche avancée

Sur d’autres sites (5105)

• Thread count option in FFmpeg for FASTEST conversion to h264 ?

  5 septembre 2013, par Saptarshi Biswas

  I need to maximize speed while converting videos using FFmpeg to h264

  • Any input format of source videos
  • User's machine can have any number of cores
  • Power and memory consumption are non-issues

  Of course, there are a whole bunch of options that can be tweaked but this question is particularly about choosing the best -thread <count></count> option. I am trying to find an ideal thread count as a function of

  • no. of cores
  • input video format
  • h264-friendly values maybe ?
  • anything else missed above ?

  I am aware the default -thread 0 follows one-thread-per-core approach which is supposed to be optimal. But I am not sure if this is time or space-optimized. Also, on certain testcases, I've seen more threads (say 4 threads on my dual core test machine) finishes quicker than the default.

  Any other direction, say configure options w.r.t. threads, worth pursuing ?

• How can I intercept encoded video data using ffmpeg with C ?

  24 février 2017, par TerNer

  First of all, I hope you’ll understand my poor english skill. I’ll do my best to write question about it.

  I’m currently developing some application. That application will captures desktop and encodes to MP4 video and send mp4 video data/chunks to another clients (all this process should run in memory not file). so another client should can play mp4 video/data/chunks that received from server.

  so I used avio = avio_open(...) and reassigned write function as like avio->write_packet = some_function. but this way seems not "generate" mp4 file/chunk correctly. and It seems the original write_packet callback having actual logic that write packet to a file. but I don’t want to write to a file. so I was thinking about it for a week but I did’t getting any of idea about it.

  so my question is : How can I intercept "encoded" video data that should processed by avio->write_packet ?

• Compression of a video with constant background

  28 juin 2017, par Spirou003

  (sorry if my english is bad, I can read it but I don’t write it quite well...)

  I want to compress some videos, which have two particularities :

  • there is a background that covers 90% of the area, during the whole video
  • most of the others elements can be separately described by a picture in move

  My videos are like this one, and don’t have audio. As you can see, almost everything can be described only by using a fixed background, few small images in move, plus a noise. Moreover, this noise will be almost nul and then an entropic coding would be very efficient. I think it will produce tiny files (< 5 Mo) even if the duration is in hours, a result that is very appreciable since I have actually recorded 30h of game (actual size is 3 Go).

  Is there any way to get new video files, that benefit of these informations ? If yes, what are the implication of a such encoding for watching these videos with Windows Media Player, or for usage with ffmpeg ?

  I searched with Google after anything that can help me, but I don’t know which keyword I can use for this, then I didn’t found anything usefull :-(

  Thanks in advance :-)

  PS : another example, the video is accelerated but shows the interesting moves