[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [MiNT] Anything happening
on 8/31/06 5:30 AM, Petr Stehlik wrote:
> Well as we both know there are four steps in the text-to-speech routine.
> And first step "preprocess" the English to some easier-to-speak
> language. And IMHO this first step should be configurable - either
> simple skip or better yet - let the caller provide its own translation
> table? - if it is to be usable for non-English as well. IIRC the
> original routine allows for skipping the first step?
Correct. The actual translation is completely separate from the final
spoken word. It first calls a translator, this breaks it down into
phonemes, then this is handed off to be spoken. The translator could be
replaced for various languages I suppose.
The current state of my driver allows one to skip the English translator
entirely if you desire. One can construct phrases with the phonetic syntax
and make is say anything you want. Although it might lack a few phonemes
that are common to other languages that English doesn't have.
Lonny Pursell http://www.bright.net/~gfabasic/