Recherche avancée

Sur d’autres sites (8053)

• Mix two audio streams using ffmpeg in python

  30 octobre 2024, par Bells2025

  I have a code that is written in Typescript that takes in two streams of audio and combines them into one. Below is the TypeScript Code. How can this be implemented in python. We are reading Kinesis Video Streams in real time and we need to achieve the same functionality in python. I'm able to pipe input and pipe output for one stream. But here they are using multiple pipes which I'm not sure how to achieve in python. Any help would be appreciated.

  


   const leftStream = await getKvsStream(`vd-${callId}-AGENT`);
 const rightStream = await getKvsStream(`vd-${callId}-CUSTOMER`);

 logger.info(`############ Begin piping KVS streams for callId ${callId}`);

 const ffmpegProcess = spawn('ffmpeg', [
     '-y', // Overwrite output files without asking
     '-re', // Read input at native frame rate
     '-f', 'matroska', '-i', 'pipe:3', // Input stream from pipe 3 (left stream)
     '-re', // Read input at native frame rate
     '-f', 'matroska', '-i', 'pipe:4', // Input stream from pipe 4 (right stream)
     '-filter_complex', '[0:a][1:a]join=inputs=2:channel_layout=stereo[a]', // Combine streams into separate tracks
     '-map', '[a]', // Map the combined audio to the output
     '-c:a', 'pcm_s16le', // Audio codec
     '-ar', '8000',
     '-f', 's16le', // Output format as PCM RAW
     'pipe:1' // Output stream to pipe
 ], {
     stdio: ['pipe', 'pipe', 'pipe', 'pipe', 'pipe'] // Create extra pipes
 });

 leftStream.pipe(ffmpegProcess.stdio[3] as Writable);
 rightStream.pipe(ffmpegProcess.stdio[4] as Writable);

 const transcribeStream = new PassThrough();
 const combinedStream = new PassThrough();
         
 ffmpegProcess.stdout.pipe(combinedStream);

 combinedStream.pipe(transcribeStream);

 transcribe.transcribeCombinedAudioStream(transcribeStream, callId, 8000)```


  


• avdevice/decklink : 10 Bit support for Decklink input device

  18 janvier 2015, par Georg Lippitsch

  avdevice/decklink : 10 Bit support for Decklink input device
  Example to capture video clip at 1080i50 10 bit :
  
  ffmpeg -bm_v210 1 -f decklink -i ’UltraStudio Mini Recorder@11’ -acodec copy -vcodec copy output.avi
  Signed-off-by : Michael Niedermayer <michaelni@gmx.at>
  

  • [DH] libavdevice/decklink_common_c.h
  • [DH] libavdevice/decklink_dec.cpp
  • [DH] libavdevice/decklink_dec_c.c

• ffmpeg transcoding : Metallic distorsion on macOS

  10 septembre 2024, par Xavi

  I am following ffmpeg's transcoding.c example to record audio and video from my input devices into an mp4 output file with H264 and AAC as codecs. I have no issues with the video and the audio transcoding works correctly on Windows.

  


  However, on macOS I have an issue with any input device other than the internal microphone of my laptop : the audio comes out distorted and sounding noisy and metallic (sample recording claps : https://file.io/nEn8MXHUAUQ2). I went through the input and output parameters for the filter several times, checked the values, but there must be something I'm missing. And I can't explain why it works on Windows with the same devices.

  


  Has anyone run into a similar problem before ? Any idea where I should start looking at ?

  


  Additional comments :

  


  

  1. While the Macbook's internal microphone (the only one that works) has only one channel, the other microphones I have available are stereo, so I can't say for sure that that's where the problem is.

  


  2. I also tested it on a macbook mini, with the same results.

  


  3. ffmpeg 6.1.1, MacOs 14.5

  


  4. Input filter arguments : time_base=1/1000000:sample_rate=48000:sample_fmt=1:channel_layout=stereo

  


  5. Output filter arguments : setpts=PTS-STARTPTS,aresample=48000:async=1:out_sample_fmt=8:out_chlayout=stereo,asetnsamples=n=1024