When headphones are plugged in, stream cannot capture the system output sound.

[zhuozl]

I want to record everyone's remarks during online meetings while wearing headphones.

[jmalfara]

This isn't an issue with whisper.cpp. This is a OS limitation. You can't route an output device into an input without some type of audio loopback.

MacOS:
https://github.com/ExistentialAudio/BlackHole

Windows:
https://vb-audio.com/Cable/index.htm?gclid=CjwKCAiAxreqBhAxEiwAfGfndEpHbfPuAMoqRqrUhmdQgIXZxomaHV5FLMcmBD21OOzu-5SN1zOF7BoCj8sQAvD_BwE

Linux: Im sure there is a way without custom software.

I wouldn't recommend these if your building a standalone / distributed app thought. They are all a little sketchy

[tartakynov]

Here I briefly described how to transcribe real-time audio from system output on a Mac using blackhole loopback driver
https://github.com/tartakynov/whispercpp-scripts/tree/main#transcribing-audio-from-system-output

[zhuozl]

This isn't an issue with whisper.cpp. This is a OS limitation. You can't route an output device into an input without some type of audio loopback.

MacOS: https://github.com/ExistentialAudio/BlackHole

Windows: https://vb-audio.com/Cable/index.htm?gclid=CjwKCAiAxreqBhAxEiwAfGfndEpHbfPuAMoqRqrUhmdQgIXZxomaHV5FLMcmBD21OOzu-5SN1zOF7BoCj8sQAvD_BwE

Linux: Im sure there is a way without custom software.

I wouldn't recommend these if your building a standalone / distributed app thought. They are all a little sketchy

@jmalfara When whipser recognizes the input voice device, can it get the output voice device?