[Discussion] Separating clear human voices, like a female and a male, NOT even overlapping.

It appears UNABLE to do this, even when there is ZERO overlap between voices.

I've seen some comments that if I consider one voice as "vocals" and one voice as background/music, etc... it should be able to do something. There does not appear to be even a hint that it does any separation.

Are there some specific models I am supposed to install?

One more problem is that it creates files, from the original WAV file, that are 4X in size. I start with 25MG and end up with 2-4 WAV files that are 92MG EACH? Why?

I used --bitrate 8000 to absolutely no effect of file sizes.