It’s Friday and I’m tired of stupid politics. Here, watch this amazing piece of animation.
I’ve been hearing about this short film for a while, how it’s like nothing anybody’s seen before, it’s signaling the future of animation, etc. Apparently it’s been playing before Wreck-It Ralph, and Disney just put it on YouTube a few days ago. Even at YouTube quality, it’s fantastic.
It looks a lot like traditional hand-drawn animation, but it’s actually CGI. Graeme McMillan at Wired explains how they did it:
The first thing you notice about Paperman is how different it seems from most modern cartoons, not just because of the limited color palette and retro styling of the characters and the world they live in, but because it doesn’t look like the generic, quasi-photo-realistic CGI animation of everything from Pixar’s Brave to, well, Disney’s Wreck-It Ralph…
Paperman‘s seemingly seamless way of blending the personality of hand-drawn animation with CGI in the physical space of the story is the result of new in-house software called Meander, a vector-based drawing program that allows for manipulation of the line after the fact — something that Kahrs described as “just like painting on the surface of the CG.”
In practice, it successfully blends the best of both forms of animation together in way they’ve never been seen before. Depicting George and Meg as flat, drawn characters keeps them safely out of the uncanny valley that even the best CGI sometimes can’t avoid and somehow makes them seem more real; other sequences, like the multiple paper airplanes zooming through the air, would be far less convincing and far more time-consuming if rendered without the help of computer generated imagery.
“Uncanny valley” is the term for the phenomenon where the more “realistic” a CGI-animated person looks, the more our minds tell us it isn’t real. I always think of the Robert Zemeckis version of Beowulf. All the supposedly “lifelike” and “realistic” characters creeped me out. Whereas the character who was supposed to creep me out, Grendel, seemed more “real” because he looked less human. The only thing I’ve ever seen that has managed to overcome the uncanny valley effect is the game L.A. Noire, and even that skeeves me out a little bit.
That reminds me of an upcoming game whose designers are obviously working on the uncanny valley problem. Check out the “acting” at the 3:30 mark in this demo of Watch Dogs. Be advised that there’s some naughty language and graphic violence and whatnot, so if you don’t like that sort of thing, don’t click the play button:
Videogames don’t cause violence any more than guns do, by the way.
Anyhow, for all the pyrotechnics, the most impressive thing to me is how they captured the characters’ facial nuances and body language in that brief conversation in the club. It’s still not quite there yet, and it doesn’t make you forget you’re watching computer-animated characters, but they’re working on it.
If you remember when this was cutting-edge CGI, you are old: