Say there is a sound recording of Mark Twain reading something. It is processed and comes out as a set of parameters that is used to run a speech synthesizer. Then you feed the text of one of his books into the machine as well and there is Mark Twain's voice reading his story-- Iridium7, Oct 18 2007 Text-To-Speech http://www.research...ttsweb/tts/demo.phpOur Demo Speaks Your Text [baconbrain, Oct 18 2007] A panphonic poem for Mission Impossible III http://literalminde...ssion-impossible-3/This idea in movie fiction. [jutta, Oct 18 2007] Voice fonts for singing (re: xenzag) http://www.guardian...19/onlinesupplementOne application: helping people who can't sing? I doubt that untrained singers could handle the recording process. [jutta, Oct 18 2007] Model Talker http://www.modeltalker.com/Looking for beta testers. [jutta, Oct 18 2007] and if you used Al Gore's voice and made it undulate melodically, would that be an algorithm too?-- xenzag, Oct 18 2007 Good one, [xen].
Link to state-of-the-art in text-to-speech. It has various accents, but still sounds odd.
This is a wish.-- baconbrain, Oct 18 2007 This is called a "voice font", and exists, although not to the degree that you'd really want. (I can't help the feeling that I'm really saying the same thing as baconbrain before me.) AT&T's research, which he links to, is a good place to start.
The FAQ from "ModelTalker", which tries to make this technology into an end-user usable product, is an interesting read that reveals much about the very real applications and limits of the technology.
This isn't just about funny celebrity voices; if you're suffering from an illness like ALS (what Stephen Hawking has) that slowly destroys your ability to move your muscles, including your throat, this may be your best chance at preserving a bit of "yourself" in your future prosthetics. (Of course, most people by now think of TTS as Stephen Hawking's voice, so for him, uh. Recursion! Recursion!)-- jutta, Oct 18 2007 random, halfbakery