Linux Advanced Guides and News!
Green means go, Yellow means faster, Red means don't stop.
Is there such a thing as text-to-dialogue software?
The idea would be that the computer would give the different characters of a play different voices so that you could hear the conversation.
Umm not sure for other TTS engines but SAPI / opensapi can be used with speech XML where you can define how the speech engine processes the speech throughout a document. So you can define different voices / pitch / speech/volume for different parts in theory. Other engines might support similar facilities.
I'll check these out
Does there exist some free text to speech software which uses Unit selection synthesis?
So using a large database of recorded speech to make new speech.
I think both eSpeak and geSpeaker use Diphone synthesis. Which uses only a small dataset of bi phones. While these programs are small, they tend to sound robot like.
The current state of the art of text to speech is so much better then this :-/!
I'm too used to having nice TTS that I can't accept festival/espeak on linux. They sound too much like "microsoft Sam default voooooice" and non-english languages are unacceptable.
I have great nextup/acapela voices which worked great out of the box in Win with coolreader book reading app and few others as well.
In 10.04 ubuntu I can't install them under wine.
Coolreader works fine.
Acapela license manager simply doesn't see the license (reported in wine appdb), and nextup voice is not visible by coolreader. It must have been incorrectly registered in wine registry or something.
Was anyone able to install/use commercial win sapi voices in ubuntu?
Yeah I am, using various SAPI Speech engines. I have not tested all of them though. You might want to pm me and have a read of:Was anyone able to install/use commercial win sapi voices in ubuntu?
Please bear in mind that it's still under heavy development.
Could you report your findings?
I'm looking for a good English text to speech application, better then eSpeak.
I guess I'm willing to pay money for it.
But preferably I would like to use some open source software, so I can make modifications myself.
As I said earlier, not all text to speech applications work in the same way.
These formant based synthesis / di-phone synthesis applications like eSpeak make it really fatiguing to listen to it for more then 5 minutes.
Unit selection synthesis applications take up a thousand times more disk space, but I find them much nicer to listen to.