Emru Townsend: Eric, what would you say are some of the limitations of working with a computer for animation?
ED: It is not unusual for an animation student to get excited about the computer because "anything is possible". In an abstract sense this may be right, but this perspective is not held by anyone who has spent any time actually trying to do "anything" with a computer. I recently took a Story Structure class given by screenwriter Robert McKee and one of his many bits of wisdom was, "The most foolish thing that an artist can do is to think that he can do anything, because if he does then he will end up doing nothing." In practice, the limitations of using a computer as a tool to do art and entertainment become all too clear. I'll try to break the idea of limitations into a few categories.
The first category, and perhaps the most basic, concerns the limitations that the computer user brings to the table. Obviously, the limitations associated with using any medium are defined by the experience and abilities of the person using that medium. I suppose these could be considered "fixed cost" limitations. There is no reason why a person with no CG experience, but with the right set of interactive tools and a good manual, could not create a wonderful piece of animation, but that person would certainly be working under some pretty significant limitations. There are people who have been involved with CG for years that are geniuses when it comes to writing programs and generating tools for creating computer imagery, but many (though not all) of these people couldn't create an entertaining piece of computer animation if their lives depended on it. This is not necessarily a negative thing. These people have spent their time developing certain skills in the art of computer programming instead of certain skills in filmmaking art and design.
On the other hand, there are people who are architects of wonderful computer animation who know little about using a computer. We have great art directors at PDI who never touch an Indigo computer, but they are able to bring their art, design and directing skills to a project to help guide those who are actually producing the images on the computer. There was a time in computer graphics when one person really did have to do it all (and it usually showed), but even today, with large teams of experts and specialists, we are a long way from being able to use the computer to do "anything" we can dream up.
Another set of limitations are the ones that are consciously applied to any CG project (or any other type of production). This includes things such as how much time and money is available, and what resources are at hand, but even more important are the limitations imposed by design, story and art direction. CG people are often seduced by things that computer technology tends to make easy, like flying a camera around, and they end up designing their work based upon the technology instead of upon a good idea. A friend of mine teaches college-level computer character animation and the first limitation he imposes on his students work is that there can be no moving cameras. In the long run, this limitation actually becomes liberating for the students because they are free to concentrate on character animation without worrying about what to do with the camera (another art). In a way, the right set of consciously imposed limitations can allow us to do "anything", whatever the medium.
Of course there are more concrete, day-to-day limitations that we run up against when computer animation tries to "compete" with other types of animation. It's hard to imagine a facial animation system that would give an animator the same flexibility that a traditional animator has with a pencil and paper. If a job suddenly requires a new unplanned expression on a character's face, a traditional animator can just draw the new face the way she wants it. A computer animator might have to spend hours or even days figuring out how to move the points in her 3D face model into the right configuration. "Collision detection" is another thing that is usually a pain with CG. Basically, unless someone programs it into the system, computer generated objects do not collide with each other and recoil. Instead, they are happy to overlap and share the same space. Computers are dumb as posts. They don't know anything unless they are told and they must be told in no uncertain terms. If you don't tell a computer anything about physics, then don't expect the computer to figure physics out for itself. The physics that we take for granted in the real world is, in most cases, impractical to program into a computer animation system. The calculations can be very involved and it is often less time consuming (but time consuming nonetheless) for an animator to just make sure his character doesn't walk through any walls or stick his hand through his chest. Consequently, things that "should" be easy to deal with, like characters shaking hands or trying on clothes, can be a much bigger ordeal for the CG animator than for animators working in other media.
I've really just scratched the surface. I guess that the notion of "limitations" is really the defining characteristic of just about anything that we do, especially art, and I feel like I'm just beginning to get a feel for how to manipulate a set of limitations to get good results in what I do with computer animation.