Not to hijack the post, but I was reading some of the stuff here and I would have to agree with concretefire, again agreeing with his sentiment that I'm not stepping on toes or anything... there is a huge difference between having the ability to write the incredible software and put together the art for it.
I'm currently employed as a 3d artist for an online series of publications. Most of what I do is centered around Daz3d, Poser, 3dsMax, and Photoshop.
Ive been into 3d graphics for 20+ years and have seen quite a bit of the evolution of the technology where that is concerned....
As far as animations for speech, there is an applet that works with Daz Studio and Poser that will automate the mouth moving, the head moving, expressions, etc... To construct a series of anims to output into a 2d format is so simple it's scary. Even easier if you are working with a 3d application.
I've been tinkering with the different AI and speech recognition software for a few years on and off and have just recently come to have a look around again...
In my line of work, over the past two years or so, Ive built up a render farm... which is basically a series of computers that work together to perform large scale renders and multi frame animations. Right now the farm is up to 16 computers ranging from P4s to i3s.
What I want to be able to do is utilize an AI bot that will act something like, yes Im going to say it, a HAL... of course without the psychotic tendencies. I want to get up in the morning and tell her to power up certain computers, turn the blacklights on in my computer room, set the stereo to a certain CD.. etc.
I plan on dedicating one computer to just displaying her headshot and text boxes and have her rigged into the entire system, lighting, enviro controls, and all.
I know mostly how Im going to pull off the external stuff, (lighting, stereo, etc) I have the character made that Im going to use for the representation of the machine, but what Im stuck on is the AI end of it.
Windows 7 has an incredible speech recog system, accurate to an incredible degree, but no return dialog. I love the functionality of Hal, but I guess what I cant seem to figure out is... can you program functions in hal like you can in e-speaking? E-speaking will allow you to create commands, almost like a macro-type thing, but again, no return, no ability to carry on a conversation...
Wow, that was lengthy. Sorry about that... I saw this stuff and teh advances, and jumped on ebay and bought a 22 inch flat panel monitor so my AI comp will have a place for her face...
Just gotta work out the fine details now...
-M-
Show Posts
