Re: Making a speech synthesizer
This is a bit of a long shot, but have you seen the wilderness project from CMU?
http://www.festvox.org/cmu_wilderness/
It's sort of a spinoff from the Festvox project. Its purpose is to produce text to speech voices automatically from datasets, with no prior knowledge of the language in question. As was said above, you do need a lot of data so if you don't have that for Uzbek, this might not be a good route for you. But I wanted to bring it to your attention just in case. They produced voices for some 700+ languages based on recordings of the bible, if memory serves. Being a native Swedish speaker, I can say that the results obtained for Swedish are reasonable considering there was no manual labor involved. Certainly a dedicated team will beat the automated approaches out of the water (at least for now), but the Wilderness project is a good backup plan. You need some technical skills to be able to generate a voice with the toolchain, but it's not that bad. I've done it multiple times with Festival/Festvox/Flite.
Kind regards,
Philip Bennefall
-- Audiogames-reflector mailing list Audiogames-reflector@sabahattin-gucukoglu.com https://sabahattin-gucukoglu.com/cgi-bin/mailman/listinfo/audiogames-reflector