Recherche avancée

Sur d’autres sites (3358)

• An efficient way to use Windows Named Pipe for IPC

  26 juillet 2020, par Ehsan5

  I am using jna module to connect two processes that both perform FFMPEG commands. Send SDTOUT of FFMPEG command on the server side to NampedPipe and receive STDIN from that NampedPipe for other FFMPEG command on the Client side.

  


  this is how I capture STDOUT and send into the pipe in server Side :

  


  InputStream out = inputProcess.getInputStream();
byte[] buffer = new byte[maxBufferSize];
while (inputProcess.isAlive()) {
     int no = out.available();
     if (no > 0 && no > maxBufferSize) {
        int n = out.read(buffer, 0,maxBufferSize);
        IntByReference lpNumberOfBytesWritten = new IntByReference(maxBufferSize);
        Kernel32.INSTANCE.WriteFile(pipe, buffer, buffer.length, lpNumberOfBytesWritten, null);
     }
}


  


  And this is how I capture STDIN and feed it to the Client Side :

  


  OutputStream in = outputProcess.getOutputStream();
while (pipeOpenValue >= 1 && outputProcess.isAlive() && ServerIdState) {
      // read from pipe
      resp = Kernel32.INSTANCE.ReadFile(handle, readBuffer,readBuffer.length, lpNumberOfBytesRead, null);
      // Write to StdIn inputProcess
      if (outputProcess != null) {
          in.write(readBuffer);
          in.flush();
      }
      // check pipe status
      Kernel32.INSTANCE.GetNamedPipeHandleState(handle, null,PipeOpenStatus, null, null, null, 2048);
      pipeOpenValue = PipeOpenStatus.getValue();
      WinDef.ULONGByReference ServerId = new WinDef.ULONGByReference();
      ServerIdState = Kernel32.INSTANCE.GetNamedPipeServerProcessId(handle, ServerId);
}


  


  But I faced two problems :

  


  

  1. High CPU usage due to iterating two loops in Server and Client. (find by profiling resources by VisualVM)

  


  2. Slower operation than just connecting two FFMPEG command with regular | in command prompt. Speed depends on buffer size but large buffer size blocks operation and small buffer size reduce speed further.

  


  


  Questions :

  


  

  1. Is there any way not to send and receive in chunks of bytes ? Just stream STDOUT to the Namedpipe and capture it in Client. (Eliminate two Loops)

  


  2. If I cant use NampedPipe, is there any other way to Connect two FFMPEG process that runs in different java modules but in the same machine ?

  


  


  Thanks

  


• Merge commit ’930ffd46e1e742674aa7cc1c2450020c63b5015b’

  4 novembre 2014, par Michael Niedermayer

  Merge commit ’930ffd46e1e742674aa7cc1c2450020c63b5015b’
  * commit ’930ffd46e1e742674aa7cc1c2450020c63b5015b’ :
  
    aacsbr : change order of operation to prevent out of array read
  Conflicts :
  
  	libavcodec/aacsbr.c
  See : c2340831b8e9032716acb0aab4893d3cc500213a
  
  Merged-by : Michael Niedermayer <michaelni@gmx.at>
  

  • [DH] libavcodec/aacsbr.c

• Is there a way to speed up audio processing (amix and adelay) in FFMPEG ?

  22 mars 2018, par Nadir

  I’m using, in android application, many ffmpeg amix and adelay filter commands over very small mp3 files (not longer than 3 seconds).
  Unfortunately each adelay or amix command takes between 2 and 4 seconds to execute, which is a lot considering that I should run the same operation for a lot of files.
  Here are two example of commands I’m running :

  amix :

  [-i, input1.mp3, -i, input2.mp3, -filter_complex, amix=inputs=2:duration=longest:dropout_transition=0,dynaudnorm=f=100[aout], -map, [aout], -ac, 2, -c:a, libmp3lame, -q:a, 4, output.mp3]

  adelay :

  [-i, input3.mp3, -filter_complex, [0:a]adelay=1|1, -c:a, libmp3lame, output1.mp3]

  I know in video there is an option (-preset ultrafast) to make the execution faster, is there a way to do the same for audio ?