How physical models of the human vocal tract contribute to the field of speech communication

Abstract

Several types of physical models of the human vocal tract have been developed previously by our group. Even though they were originally designed for education purposes in acoustics and speech science, some of the models can also be applied to research, pronunciation training and clinical purposes. For example, a model for the English /r/ was originally designed to teach how the sound is produced, but we have also found it to be affective when applied to practicing English vowels for non-native speakers. Another model for lateral approximant was originally designed to teach how lateral sounds are produced. The model was then tested to measure differences in sounds radiated from the center and lateral directions with the possibility of evaluating misarticulation in a clinical situation. A recent model with a movable lower lip and rotating tongue to imitate the retroflex gesture was used to simulate the English /br/ cluster, a particularly difficult speech sound for Japanese native speakers. By using this model, users can observe each articulators' movement visibly with individually adjustable parameters to produce different speech sounds. Thus, the vocal-tract models potentially contribute to the field of speech communication.