The intent of this project is twofold: Primarily to develop an improved device for those who use an electrolarynx, and secondarily to develop a musical instrument for submission to the Hackaday 2018 Musical Instrument Challenge. The contest entry is only to gain visibility for the potential improvement that this device can offer. Development will reflect these priorities.
The fundamental principle of operation involves creating an audio waveform that emanates from within the mouth, that is then shaped by the vocal cavities to result in intelligible speech. Unlike the Talk Box and the Electrolarynx, I believe technology has reached a point where bulky and obtrusive equipment to perform this task is no longer necessary.
Introducing the TalkVox: A self contained module that attaches to the palate, and emits an audio waveform directly within the mouth.
(Disclaimer: This project is currently in its infancy, and its name and capabilities are subject to change at any time.)
Upper teeth aligner (or equivalent)
This is the bracket that holds the device to the roof of one's mouth
After researching various combinations of processors, microcontrollers, and the necessary support IC's, I tentatively settled on using the ESP32-PICO-D4. It has integrated Bluetooth and WiFi (along with most of the required components for it) as well as some ram and flash and some other stuff. To be completely honest, I have only the vaguest idea of whether this will be an sufficient to implement all of the project goals.
Nonetheless, a while ago I ordered the ESP-WROOM-32 module, and it finally came in. As far as I'm aware, the module matches the pico-d4 version quite well in terms of specs, so it should suffice as a very cheap development platform.
Additional thanks go to Demetri Katsilometes for excellent photography.