An arbitrary vocal-tract configuration can be simulated by placing a series of thin plates together, which have holes of different diameters. The following is an example with 10-mm thick plates:
This kind of vocal-tract model is called the “plate-type”. With this plate-type vocal-tract model, one can simulate the five Japanese vowels based on the combinations of the diameters listed below (the glottis is on the right and the lips are on the left, and the unit is in mm):
The following video clip shows the five Japanese vowels produced from this plate-type vocal-tract model with a glottal sound of an electroralynx:
Umeda and Teranishi (1966) implemented a vocal-tract model with a series of moving 10-mm or 15-mm thick plastic bars, closely inserted from one side. With this model, one can dynamically change the vocal-tract configuration arbitrarily.
With this model, we mounted an actuator for each bar, so that a computer can control the positions of the bars. This system allows you to dynamically change the vocal-tract configuration in time. In the next video clip, vowels are produced consecutively:
- Arai, T., “The replication of Chiba and Kajiyama’s mechanical models of the human vocal cavity,” Journal of the Phonetic Society of Japan, 5(2), 31-38, 2001.
- Arai, T., “Education system in acoustics of speech production using physical models of the human vocal tract,” Acoustical Science and Technology, 28(3), 190-201, 2007.
- Arai, T., “Mechanical vocal-tract models for speech dynamics,” Proc. of the INTERSPEECH, 1025-1028, 2010.