Hanging out undisturbed on the internet for a good number of months was this most exciting demonstration of a new animation technique developed at Pixar. As you’ll see, this was presented at last year’s Siggraph conference.
And it’s quite stunning.
Thanks to Cartoon Brew for the embed.
Now, let’s think about what we saw. So, the system is smart enough to take two different types of input and then synthesise them to create a hybrid. On the one hand there’s character animation in the standard Pixar style and on the other, there’s painted keyframes.
What this would mean, if it could work efficiently, is that any style the artists could desire could be input as the keyframes and the system would then re-skin, as it were, the animation to take on that aesthetic. It’s not just simple texture mapping, but actually mimicking the style of the keyframes within the parameters of the full animation.
The final output would have “temporal coherence” in comparison to the “base” animation and a final look that was controlled by the style of the keyframes. This really is a new frontier and the possibilities are amazingly far reaching.
Well… consider my mind blown. Genuinely. Now let’s see who uses this, and to what end.
You can read Pixar’s full paper on the technique on their website.
(And, yes, I know some folk will have put 2+2 together and thought of animated Drew Struzan already)