I met up with a Softimage guy last night at a press event. Softimage recently unveiled a new application called Face Robot which creates facial animation for movies and games in a fraction of the time.
The new system essentially uses software to predict facial muscle and skin movement so you don t need to use as many points in motion capture.
Anyway, we started talking about the whole Uncanny Valley concept and why it is video game characters don t look real.
He believes, and I agree, that it
s all in the eyes.
The eyes aren
t just the window to the soul, he thinks, they are the window to believable character animation.
One of major school of thought is that you won t have really believable character animation unless you have a brain behind the eyes. In other words, you need to have some sort of digital puppeteer in control of the character, a person who can animate the character in real time and react to the situation.
The idea is to program artificial intelligence to fulfill that roll, so you have a computer controlled intelligence directing a computer crafted face.
It
s all very bizarre, but this rep says that many of the next gen companies realize the importance of succeeding in getting past the Uncanny Valley. Next-gen consoles, he said, can
t survive without doing that.




















