Visualizing articulation in speech production with physical models of human vocal tract

We have developed several different variations of physical models of the human vocal tract in the past. The original motivation behind these models was educational purposes for teaching. For example, if we wanted to test how different tongue configurations yield different sounds in a classroom setting, it was not that easy until recently. Nowadays, an ultrasound device can be available not only in clinical settings for speech and language therapy but also in teaching. One of our solutions for many years was to apply vocal-tract models for this purpose. With the models, we can visualize, and even auralize, articulators and their movements, such as the lips, tongue, palate including velum, and pharyngeal wall, and easily demonstrate the effects of tongue configurations on vowel qualities, for instance. Recent models can even visualize dynamic movements of articulators. One application is to visualize timing issues of two consecutive movements of articulators: b and r of a consonant cluster, including vowel epenthesis by L2 speakers. Thus, physical models are useful for visualizing not only the static configuration of the vocal tract but also the dynamics of speech timing. Work supported by JSPS KAKENHI Grant No. 24K06423.