I saw an interesting article on Slate the other day: The Undead Zone: Why realistic graphics make humans look creepy.
The basic thrust of the article is that when something looks slightly human – say a cartoon, or a C3PO-like robot – we fill in the gaps. But when something looks almost, but not quite human, we start to focus on the things that look wrong instead. This was observed by roboticist Masahiro Mori, who called it the uncanny valley. The term refers to the appearance of a graph plotting emotional response (y) against how closely something resembles normal humans (x). Up to a point – say 90% – the more humanlike something is, the better people respond to it, until it reaches that almost-but-not-quite-there point where instead of responding positively, people start responding with revulsion and active dislike. Eventually, as things get closer to “real,” the curve swings back up again until the reaction is the same as to a normal person.
So what does this mean for video games? At least for some people — including the article’s author — state-of-the-art graphics are in that valley. We can get a very good representation of a lifeless but moving human being. Getting those last few details, pushing up the far side of the valley, is going to be very hard.