you could also try converting the audio to keyframes and pick whip it to CCSplit. Might not give you the detail you want but has worked well on one of my projects. If the movement is not extreme enough try boosting the audio and playing with the levels then repeat the steps (convert audio to keyframes and pick whip them to CCsplit across mouth area). You can add teeth, tongue etc. as separate layer behind the face layer. Again, might not be what you are looking for but I have had success with it.