Use [S1] and [S2] tags in your script to indicate which voice should speak each line. Processing typically takes 1-3 minutes.