March 29, 2013
Activision Brings Next-Gen Video Game Characters To Life
Peter Suciu for redOrbit.com — Your Universe Online
This week at the Game Developers Conference (GDC) some of the new video game animations looked a lot less like video games and a lot more real. Activision Blizzard offered a demonstration on Wednesday, which included next-generation rendering of characters that featured skin, facial expressions and eyes that were incredibly detailed and convincing.
Part of the reason for this is that the characters' designs feature those flaws that exist with most people, and thus the designs are actually less than perfect, which in turn could make them all the more realistic looking.
These character designs could also be the first to actually break past the so-called Uncanny Valley, a term coined by early robotics designer Masahiro Mori in 1970. It essentially describes the range of sophistication in animated graphics, where as human figures at one end look unrealistic and much more like cartoon characters.
As animation or rendering moves to the middle it becomes just realistic enough to seem real, but off-putting at the same time for its level of creepiness. Activision Blizzard may have successfully crossed to the other end with animation that is essentially indistinguishable from reality.
Mashable explains that the 2004 movie Polar Express is often evoked as an example of animation that is squarely within the center of the Uncanny Valley. A lot has happened in the eight years, including greater computing power and developments with graphics technology.
What has also happened is that greater understanding of creating the “less than perfect” animated character. This includes animating facial movement in a more realistic way, but also breaking up uniform surfaces and refining the way light plays off of skin surfaces.
This also doesn´t require vast amounts of extra processing power either, but is being handled in what the team calls “cavity mapping” that can simulate the behavior of human pores and skin in relation to the light that hits it.
“ For us, the challenge goes beyond entertaining; it´s more about creating a medium for better expressing emotions, and reaching the feelings of the players,” Van der Pahlen wrote in a blog post. “We believe this technology brings current generation characters, into next generation life. At 180 fps in a Geforce GTX 680 [sic].”
Just as beauty is in the eye of the beholder, and it will require seeing to be believed, the team has focused a great detail of attention on the eyes in the animations.
Van der Pahlen told Polygon that the eyes were the most difficult part of the body to render, in part because eyes are made of different tissues, but also because eyes reflect that aforementioned light so differently. For now this could still use 3D texturing to simulate the way the light behaves.
The Activision Blizzard team further said much of the believability could come down to face time — that is when a character´s face fills the screen it should be made more realistic and thus believable, with more resources devoted to it. When the character faces are farther away from the player´s camera resources should be reduced, and focus on the other elements to continue the magic.
The GDC demonstration reportedly involved a facial capture that utilized 30,000 points of information along the actor´s face and head, or approximately 30 times the number used in-game now. Those digital animations might finally and truly be read for their close-up.