Oh one more point, no the program doesnt recognize the two different voices at all, its all grist to its mill and so separating the interwoven speech is up to the post production editor - me.

I'm still enchanted with how well this works despite the obvious shortcomings which may have work arounds for all I know