Keith Lango points out some statistics on the rendering for Cars vs. Toy Story. While you might expect the rendering times to have gone down, apparently the "average" frame in Cars took 15 times longer than an "average" frame in Toy Story. Keith leaves us with this thought:
But I’m left with this question: Eliminating the variable of increased audience expectation for technical sophistication in CG visuals- was the core metric for film quality (the artistic, visual, social and emotional result) of Cars superior to that of Toy Story by such an expansive order of magnitude?
Cars is generally considered to be a much weaker film than Toy Story, so of course the answer to Keith's question is "no." However, I think it would be wrong in this case to come to any conclusions that the good folks at Pixar are losing sight of story in the face of the marvelous technology they have at their fingertips. In fact, I think before we even think about Keith's question we need to take a closer look at what "average" means here. Take a look at the following three shots from the movie:
Note the racing scene in particular, with it's thousands of spectator cars and high-speed racers. Even with a substantial amount of cheating, those shots had to have sky-high polygon counts, plus whatever else went into rendering those scenes. On the other hand, much of what takes place in Radiator Springs is much more tame by comparison. The question I have, then, is what do we mean by "average?" Even if we assume that the modeling, texturing, and animation is more complex in Cars than in Toy Story, I find it hard to believe that the average rendering time means what it seems to here. Instead, I suspect that the racing scenes in particular (and maybe the stuff in the cornfield) took a phenomenally long time to render, throwing off the stats for the whole movie.
So-- whatever the flaws of "Cars," I think the problem here we have is sloppy math and a geeked out writer marveling over the computing power thrown at the movie, not any misplaced priorities on the part of John Lasseter, et al.
Posted by Jason at December 3, 2006 02:17 PM