Streaming Audio Input.

[Praj-17]

Hi, I am generating Audio streams, and I want to pass the generated audio stream directly to the MuseTalk real-time inference script. and get realtime streams generated. Is it possible? If Yes, How?

[Viktoo]

I need this as well. But I don't think it's possible

For example, when inference begins, the code extracts features from audio found in a local audio_path. This implementation assumes the full audio file is ready at the start.

MuseTalk/musetalk/whisper/audio2feature.py

Line 97 in 058f7dd

def audio2feat(self,audio_path):