antiboredom/videogrep

Ability to recognize/choose speakers?

rwfeather opened this issue · 1 comments

Would it be feasible to choose clips only from a certain speaker (or set of speakers)? Not sure what the workflow would be like.

Vosk supports doing speaker recognition (example here). I haven't tried it yet, so not sure if it's accurate enough for this purpose.

I'd be super interested in implementing this, or at least playing around to see if it's viable!