How do we get sounds that feel expressive out of a computer device completely devoid of human emotion?
