Slashdot Mirror
Image Metrics May Revolutionize Facial Animation
iStorm writes, "I've been interested in computer animation for a long time and have recently started cracking down on my studies in an effort to eventually move myself from hobbyist to professional... then I find this article about Image Metrics, which can map an actor's emoting onto a generated face or onto the image of another actor, living or dead. How does a seasoned animator view this sort of push ahead in technology? If so much of the creative process is made so easy, where's the need for traditional animators spending exponentially larger amounts of time to create work of equal or lesser quality? How did animators view motion capture when it first appeared? Will there still be room for creativity if this tech comes to fruition?" The article doesn't say what kind of time or processing power Image Metrics's "high-fidelity, performance-driven facial animation" requires.
My roommate is a digital animator and if his comments are worth anything then Mo-Cap is not all it's cracked up to be. This new great thing may end up there, where it can map facial expression but does it in a way that isn't quite right looking to the human eye, thus requiring hours and hours of cleanup afterwards.
I think it'll be a while before the industry starts putting out photo realistic digital animations of people.
A blog about stuff.
Humans hold human characters to a higher standard than other CGI. It's part of the reason why Toy Story, Monsters, Inc., A Bug's Life, and Over the Hedge have focused mainly on non-human characters or a cartoon-like story. When a more "adult" story is tried, as with Final Fantasy, the technology still comes up short. It's a step in the right direction, but until you can't discriminate between a CGI actor and a real one, this isn't going to be used in "serious" movies.
An earlier version of the software "needs slightly less than 5 ms per iteration and parameter (>50 ms for all 12 parameters) on a PC with a 500 MHz Pentium III processor." I'm not sure if that's 50ms per frame or what. That site also has a nice mpeg of the software mapping to a face.
wow?? I've been doing character animation professionally (read: features) for 10yrs on the computer and I didnt see anything new there...and certainly nothing that made me say wow. I sure didnt think that African Warrior was very far beyond Polar Express. Every couple of years this same old hype gets dragged up and people get all excited and then the studios hit the bandwagon and make another Final Fantasy or Polar express. It sucks, the eyes are dead, and its creepy. Uncanny Valley revisited. The imporvements each time are incremental at best (to my eye.) But the real bottom line of all this is that its kind of gimmicky. Character animation is meant to *EXAGGERATE* reality, not replicate it. What they describe here is really more visual effects than character animation. Studios love this because they hope someday they can stop dealing with all those annoying actors, animators, and artists. If you want to make a movie about James Dean dating Marilyn Monroe, go right ahead. But if you want to make a movie about a fish thats in love with a bug, mocap (facial or gestural) isn't going to be the most appealing solution. The only photoreal CG Character that ever made me say wow was Gollum and there were many, many animators (whom i know personally) who both animated and cleaned up facial motion capture for thousands of hours to make Gollum. No slight to Andy Serkis who is a very talented performer, but it just ain't that simple. Animators will be around as long as there are children (and adults) who want to see thier imaginations taken somewhere its never been before. (i.e. forever)