Future Work

Designing models that have yet to be built for the American English language allows for a lot of expansion off of this project. For instance, testing the models with more complex tests than simple identification could lead to a better understanding of how the models work and help with improvements in the design. Testing a wider variety of people could also shed some light on human perception of sound and specifically speech.

One improvement that would require some more research and experimentation would be to change the design to allow for the input vibration to be produced with an electro-larynx with a oral cap and straw. The cap and straw combination is used by patients with too much scar tissue on their vocal cords for normal electro-larynx use. However, an added bonus is a reduction in background buzzing produced by the device. To implement this set up with models, the placement of the input must be considered because the straw is typically inserted into the mouth as opposed to pressed against the throat. So this leads to the questions of where to design the “port” or entry point for this input signal.

Of course, a project such as this could be inspiration for a number of different types of models for a number of different types of phonemes. Diphthongs, consonants, or any of the other vowel sounds I did not focus on could be the central point of another project.