Extraction of individual speaker audio from overlapped speech regions

[PhilipAmadasun]

As my title says, Is there a way for pyannote.audio to extract individual speaker audio from overlapped speech regions?

[hbredin]

Not yet but we are working on it. Stay tuned. In the meantime, you might want to have a look at Asteroid source separation toolkit.

[PhilipAmadasun]

Thanks for the reply. I had some more questions about how you guys calculate cosine similarity and of the speaker_verification model itself. Not sure if I should ask a whole new question as this aspect kinda also relates to this for my use case.

• First:
  from scipy.spatial.distance import cdist distance = cdist(embedding1, embedding2, metric="cosine")[0,0]
  Gives an error, I was following the tutorial here
  I have to do this instead:
  distance = cdist(np.expand_dims(embedding1,axis=0), np.expand_dims(embedding2, axis=0), metric="cosine")[0,0]
  To get results. And now I'm confused if I'm even supposed to do this.

• Second:
  The speaker verification module gives me an error I could not find when searching through issues on the repo. Should I just create another question to ask about that?