Speech production model based on articulatory movements

Speech is generated by articulating speech organs such as jaw, tongue, and lips according to their motor commands. We have developed speech production model that converts a phoneme-specific motor task sequence into a continuous acoustic signal. This paper describes a computational model of the speech production process which involves a motor process to generate articulatory movements from the motor task sequences and an aero-acoustic process in the vocal tract to produce speech signals. Simulation results on continuous speech production show that this model can accurately predict the actual articulatory movements and generate natural-sounding speech.