I think that windows has somthing like this. I''d check out the MS site. Anyway, I am sure that some one will come along and correct me if I am wrong, or prehaps give an address/help.

ANDREW RUSSELL STUDIOS

quote: Original post by DaA

Okey thanks, But what i don´t understand is whay no one ever used them in games, a text file is a lot smaller then a sound file.

They sound far too monotonous. Plus CDs and the soon to be popular DVDs have ample room for sound. Not to mention that there would be an obvious extra burden on the processor that has to generate this speech. I''d rather have games that come on two CDs but have human sounding voices than a game that comes on one CD, but sound like someone is talking through an air vent *cough* ananova *cough*.

And to answer the question of the original poster, the best place to look for these things is on the net:

www.ananova.com (comes closest to imitating human emotion)
http://www.research.att.com/~mjm/cgi-bin/ttsdemo (funnest)
http://www.bell-labs.com/project/tts/voices.html (also good)

hope that helps

One idea I had is "moduled" speech. A person''s phenomes would be recorded to a file. This file is then used to detect the phenomes in the person''s voice and "write" what they said phonetically. Along with that, tonal inflections would be recorded. At playback the three types of data would be used to reconstruct the voice.

Err, that''s a bit complex. But it would definately keep the file size low. Around half a meg to a meg for the recorded phenomes. After that the phenomes and inflections would be around the size of a MIDI file. I''d guess roughly 100k for five minutes (on the high side.) Hmm, so much for editable text.