open source text to speech library [closed]

问题:

Does any one know of a good text to speech library. It needs to be open source and provide C API?

PS: I've already done a search, but I'd like recommendations from people who have actually used these APIs

回答1:

you can go for Festival worked nice for me.

回答2:

Festival is an open source text-to-speech system.

Stanford uses it for their Natural Language Processing class, and they have up-to-date instructions about installation on this cs224s homework page. Installation on Mac OS X requires a couple patches, which they've wrapped into a handy install script.

There are alternate voices you can use which sound noticeably better than the stock ones. You can find information on these voices in this forum post: How to setup more realistic voices in Festival. Those instructions are for Ubuntu, but the voices work with any Festival installation.

回答3:

I have used flite in an embedded server. It has a small footprint and comes with a single voice

回答4:

eSpeak is another lightweight TTS. More robotty than Festival.

回答5:

Hey what about MARY? It looks so awesome to me which one should we push further for open source enterprise computer or androids or whatever? We need to strenghten efforts by shifting all open source resources to the best/most promising we have so far.

Anyone experiences with MARY TTS? Or does it have a flaw (because noone mentioned it before)? More links:

http://www.babelfish.org/tts-free.htm. https://www.cereproc.com/en/support/live_demo http://www.digitalfuturesoft.com/dfttssdk.php (also provides ARM version, like MARY + Festival)