A Digital-Computer Simulation of the Human Speech-Production Mechanism

A simulation of the human speech-production mechanism is described and its application to the area of biomedical teaching is indicated. For the group of students interested in the clinical and scientific areas of speech work, the simulation is particularly useful for showing how various parameters affect speech outputs, and also for helping estimate parameter values that cannot be measured directly. For students of bioengineering, the model and simulation is valuable for illustrating techniques of simplifying complicated physiological systems and actually carrying out the details of a real-world simulation. The model used in our work divides the speech generation into four distinct portions: 1) the lungs-bronchi-trachea which serve as a pressure source; 2) the larynx which provides the oscillatory mechanism; 3) the vocal tract which serves to shape the audible signal produced by the larynx; and 4) the mouth opening which serves to interface the vocal tract with the air outside. Since the above model has been implemented using the IBM System/360 Continuous System Modeling Program (CSMP), it is very easy to demonstrate effects of changing various parameters on the waveshape produced by the system, and to introduce structural changes in the model corresponding to newly obtained physical data or refinements in the mathematical models of the various subsystems. Because CSMP is an application-oriented language, the user can concentrate his efforts on the details of the system he is studying rather than worrying about subtleties of numerical analysis and programming.