Domain: naturalvoices.com
Stories and comments across the archive that link to naturalvoices.com.
Comments · 9
-
Re:The new Conectiva CL 10 is entering in RC stage
For details on what the developers are planning for Conectiva 11, read the dev plans.
-
Re:Conectiva
Check out the history behind Connectiva... a fascinating read!
-
release notes for version 11
-
Important message from CmdrTaco
-
Re:How can they really stop it?
-
Re:How can they really stop it?I don't understand the "we todd did" part -- it doesn't come across very clearly in the responses you got with wav/mp3 files.
However, I've been playing around with the AT&T demos and I found a string that produces what I think you wanted:
"eye yam sofa king way stud"
-
It's you!
This can be lots of fun, actually!
-
Re:How can they really stop it?
-laughs-
Probably won't carry the same weight as when you trick a human being to do it; its voice synthesis probably isn't on par with AT&T Natural Voices.. it'll probably be kind of choppy, halting, broken English. -
Re:Ananova
I mean, if the entire Titanic can realistically be rendered, why not be able to render a person in real time with realistic voice synth and physics?
Are you serious? If you are, you obviously have absolutely no grasp of how 3D modeling works.
Rendering something like the Titanic is easy. (Not easy in the sense that anyone with a copy of 3D Studio MAX can do it, but easy in the sense that it's just a ship.)
When you create something in 3D like the Titanic, it's based on specifications that do not change. Lighting is constant, shapes stay the same, and moving parts are minimum.
Compare that to attempting to duplicate a person, detailed, in 3D. People are tremendously harder to do than objects, because people automatically scrutinize other people. That's why when you look at a movie like Final Fantasy, you can say "Wow, they sure are realistic, but there's just *something* not right."
With a person, you have to deal with mouth movement (a very difficult thing to model in 3D), eye movement, muscle expansion and contraction based on movement, bending limbs and joints, breating, and a whole host of other factors. Then when you get into voice synthesis (which is still not perfect, but AT&T is making leaps and bounds.), and physics modelling on things such as cloth and water... It's all very hard.
So between rendering something like a ship moving through the water, or creating a realistic person in 3D, the ship is a lot easier to do. It may be painstaking in detail to create, but it's still just basic shapes.)