Currently CH listens for 60+ sounds and translates them into the 11 different viseme shapes. In general, these should support multiple languages and dialects. So an "ah" type sound, regardless of language, will probably have an open mouth with the teeth and tongue showing, for example. But yes, we are currently limited to those 11 mouth shapes, but we are always working on making the lip sync better. Are there particular sounds that you find are not working for you very well?
Yes, if I may rally to the question.
In German language we have lots of "sh"s, as in shift, shamrock, or, shotgun.
The first to "sh" translate well with a standard "s"viseme, while "shotgun", that runs into an "o", doesn't really work well with that viseme.
We have that way more often in German:
Sch (pronounce "sh") is commonly followed by "o" and "u" and most people make mouth shapes that look a loot like a standard "W-Oo" viseme.
So why not simply use a viseme similar to that for the "s" viseme and translate it into "sch" that way? Because, of course, that would get rid of the normal "s" and is not ideal.
The ultimate thrill would be this: Having the possibility to add new visemes based on the 60 sounds that ChAn recognizes, thereby any language could be custom-made for by the respective community or user.
Thanks Stefan, great feedback!
Thank you. I posted this in the "feature request" page.