It has been my understanding we could use webrtc library to detect speakers. Is it not the case?
For one speaker in a conference it works, start and end of speech are detected OK. With several speakers in a conference one of them who is silent (confirmed by asterisk monitor app) can get multiple frames in a row detected as speech by webrtc library when some other member is talking. Tried appKonference releases 2.2 and 2.5.
Again, should this library be used for detecting speech in a conference, or it still can be used to detect speech coming from a specific member? Thank you.