A Mathematical Model for Speech Processing