Presently the mouth shapes for "ah" and "oh" are inserted as still frames as determined by the speech file.

How about we have intermediate frames for these mouth shapes? Like the transition from "ah" to "oh" could have just a few frames of transitionary shapes between these two (which we could upload ourselves). The program could then insert these transitionary phases in automatically depending on the order of the shapes. It would be very easy for the program to do this.

It would make the whole thing look so much smoother, and if it worked, I think would be an absolute game changer.