A Facial Animation Model For Expressive Audio-Visual Speech