Recherche avancée

Sur d’autres sites (10198)

• Why use sysconf(_SC_NPROCESSORS_CONF) to get the number of CPU cores in x264 ?

  5 novembre 2017, par bknieven

  I find x264 use sysconf(_SC_NPROCESSORS_CONF) instead of sysconf(SC_NPROCESSORS_ONLN) to get the number of CPU cores.

  Here is the code snippet in x264/common/cpu.c :

  ...
  
  #ifdef __ANDROID__
  
      // Android NDK does not expose sched_getaffinity
  
      return sysconf( _SC_NPROCESSORS_CONF );
  
  #else
  
      ...
  
  ....

  And I find the difference between _SC_NPROCESSORS_CONF and _SC_NPROCESSORS_ONLN in this manual :

  sysconf (_SC_NPROCESSORS_CONF) which returns the number of processors
  the operating system configured. But it might be possible for the
  operating system to disable individual processors and so the call
  sysconf (_SC_NPROCESSORS_ONLN) returns the number of processors which
  are currently online (i.e., available).

  I have two questions :

  1. What is the difference between "the number of processors the operating system configured" and "the number of processors which are currently online" ?
  2. Why x264 use _SC_NPROCESSORS_CONF in Android to get the number of CPU cores ?

• How to decode H264 stream to image byte array with ffmpeg in C# ?

  1er juin 2022, par massivemoisture

  I have an application in C# that receives H264 stream from a socket.

  


  The stream is in Annex B format. I want to decode the stream to individual image as a byte array.

  


  I saw this code snippet using PyAV, courtesy of leng-yue (https://github.com/leng-yue/py-scrcpy-client/blob/main/scrcpy/core.py) :

  


  raw_h264 = self.__video_socket.recv(0x10000)
                packets = codec.parse(raw_h264)
                for packet in packets:
                    frames = codec.decode(packet)
                    for frame in frames:
                        frame = frame.to_ndarray(format="bgr24")
                        if self.flip:
                            frame = cv2.flip(frame, 1)
                        self.last_frame = frame
                        self.resolution = (frame.shape[1], frame.shape[0])


  


  How can I do something similar with ffmpeg but in C# and the output is a byte array in BGR24 ? Thank you !

  


• ffmpeg complex filter pip loop not working

  4 août 2021, par Dzseti

  I am using an ffmpeg complex filter to :

  


  

  1. turn a number of images into a slideshow

  


  2. scale the slideshow to 1/6th of its original size

  


  3. loop the slideshow infinitely

  


  4. set the looping slideshow as a picture-in-picture over another looping video

  


  


  then add audio from another file which determines the maximum length of the final video.

  


  Everything is working except step 3.

  


  Here is my code (broken down into individual lines) :

  


  ffmpeg -y 
-loop 1 -t 5.5 -i /uploads/2021/07/slide04-scaled.jpg 
-loop 1 -t 5.5 -i /uploads/2021/07/Slide05-scaled.jpg 
-loop 1 -t 5.5 -i /uploads/2021/07/Slide06-scaled.jpg 
-i /slowwmo-videos/intro.mp4 
-i /uploads/2021/07/busy_bees_01.mp3 
-filter_complex 
"
[1]fade=d=0.5:t=in:alpha=1,setpts=PTS-STARTPTS+5/TB[a1];
[2]fade=d=0.5:t=in:alpha=1,setpts=PTS-STARTPTS+10/TB[a2];
[0][a1]overlay[b2];
[b2][a2]overlay[slides];
[slides]loop=loop=-1,scale=iw/6:ih/6,format=yuv420p,fps=fps=25[pip];
[3][pip]overlay=W-w-100:200[v]
" 
-map [v]:v:0 -map 4:a:0 -vcodec libx264 -profile:v main -video_track_timescale 25000 -pix_fmt yuv420p -s 1280x720 -acodec aac -af aresample=44100 -shortest /slowwmo-videos/busy_bees/7494.ts