Thats pretty interesting, I've looked into a few tts engines for speech synthesis but that looks like you could potentially sculpt alot better output. Do you have a link to an article, or were you considering sharing the source? Whats the performance, and can it simulate different vocal ranges like male/female/child/etc?
I sent you a PM with the current version. It's javascript, so the source is right there, and there should be some references near the bottom. If you want to start with this stuff, a good place is delay lines to simulate e.g. guitar strings, there's lots of stuff for example here:
https://ccrma.stanford.edu/~jos/pasp/My prototype runs rather well in chrome, but has problems in firefox (but was fine before firefox 39
). It's a rather simple model of the vocal tract, and a long long way from being usable for text to speech stuff, but I think there are people working on that sort of thing. In principle it should be easy to change to, say, a female voice, by raising the pitch at the glottis and shortening the vocal tract, but I didn't really set the code up for it.
A related question: there's a croquet game I would like to make, but it would rely a lot on communicating the feel of the mallet hitting the ball through sound. Is there a good reference for how to do collision sounds like this?
Developer
