Reported by aleksey_s on 2008-10-27 14:00
for newfon synthesizer (which has 4 voices) i need to make some text preprocessing. currently it is done with regexprs inside the module, but many of them (and many, which will be added later) might be expanded to the nvda speech dictionaries mechanizm. if developers not opossite to this i can write patch myself. Blocked by #219
The text was updated successfully, but these errors were encountered:
Comment 1 by pvagner on 2008-10-31 09:37
I guess it is not difficult to add another synth specific dictionary. But I am afraid we do need to shrink it somehow. This would mainly clash with sapi4 and sapi5 synthesizers. We can use a lot of various engines via sapi4 and sapi5 speaking various languages. Imagine loading russian specific dictionary entries to the sapi4 would preprocess the text even with english, french, german and other voices.
Some ideas from the top of my head:
Lets do something similar to what we are doing with the current voice filtering by language instead of voice name.
Or let's filter by manufacturer's name or some other engine identifier.