Activision showed off the state of the art of real-time graphics on Wednesday, releasing this mind-boggling character demo. The character’s skin, facial expressions and eyes look so real, it’s uncanny.
When you watch this video, see if you think this character has reached the other side of what’s commonly called the “uncanny valley,” a term first uttered by early robotics guru Masahiro Mori in 1970. It describes the range of sophistication of animated graphics, from one side of the valley where human figures simply look unrealistic, to the middle of the valley — where they look just realistic enough to be creepy — to our side of the valley, where animation is indistinguishable from reality.
Whenever the uncanny valley is mentioned, the animation techniques from the November, 2004 movie Polar Express come to mind. Most viewers noticed the characters weren’t quite photorealistic enough to keep them out of the creepy zone. But that was nearly 8 years ago, and graphics technology has made spectacular progress since then.
This newest character edges ever-so-closely to our own side of the uncanny valley, where art is indistinguishable from reality. That’s thanks to what game developer Activision calls “next-generation character rendering.” This character is part of a presentation Activision’s real-time graphics R&D expert Jorge Jimenez gave on Wednesday at GDC 2013, the annual Game Development Conference.
According to Jimenez, this face and others like it represent “the culmination of many years of work in photorealistic characters.” At the presentation, Jimenez showed “how each detail is the secret for achieving reality.”
The trick for these gaming developers is not only to create photorealistic animated characters, but to ensure that such animations can play in real time on video cards and computers that ordinary people actually own. And that’s what Activision has done, using standard bone animation, facial scanning, performance capture and lots of intricate artwork to make skin look real.
Jimenez wrote in his blog about how this type of animation has made tremendous progress. “We believe this technology will bring current-generation characters into next-generation life.” Like baseball great Babe Ruth pointing at the center-field fence, before his presentation Jimenez declared this animation would be rendered live, and “we will show it running on our two-year-old laptop.”
I think this is impressive, but not perfect yet. For instance, look at the guy’s teeth. And what about hair (not that there’s anything wrong with that clean-headed look)?
Do you think this character looks real enough to fool people if it were used in a motion picture or video game? Or is it still creepy enough to be considered a resident of the uncanny valley? Let us know what you think in the comments.