Thanks for the reply, yeah our next step is definitely some music theory. Current implementation was a fairly naive approach to bpm and beats, but in theory all of the sections should fit inside 20 second chunks for each dialogue snippet. Thanks so much for the feedback again!