Cellular Visions: The Inner Life of a Cell
What can character animators learn from those who render microscopic worlds in 3D? Plenty.
In some instances, that meant sacrificing literal accuracy for visual effect. "What we did in some cases, with the full support of the Harvard team, was subtly change the way things work," Liebler says. "The reality is that all that stuff that's going on in each cell is so tightly packed together that if we were to put every detail into every shot, you wouldn't be able to see the forest for the trees or know what you were even looking at. One of the most common things we did, then, was to strip it apart and add space where there isn't really that much space."
Because of the sheer volume of structures that needed to be visible and functional in each scene, XVIVO was given a longer development cycle than a typical commercial project, to transform the molecular data and storyboards supplied by Dr.Viel, and Dr. Robert Lue, the director of life sciences education and creator of Harvard's BioVisions computer-based learning program for undergraduates. "In a lot of cases, the animation had to be worked out more thoroughly than we would have done on an ordinary project so that we could understand what was going on," says Liebler. "Even though David has a medical background, Mike and I come from art and graphic design backgrounds. We needed to get a grasp of the actual objects and things that were involved by making them; only then could we understand what was happening. But we also continued working on this project while other projects were coming in and out. The Harvard job was a constant that we were building and refining until we had satisfied its unique academic and aesthetic requirements."
Luckily, Harvard's Dr. Lue and Dr. Viel were receptive to the animators' questions along the way and always came back with good advice and relevant resource materials. "The revision process was more organic than a typical job, where we'd have a simple review and revision schedule after we received the storyboards," says Liebler. "In this case, they needed to describe these unseen processes for us and then we went back and forth with what we thought we could show and what we thought we couldn't show. As we did this, we'd discover things that, surprisingly, we realized we could show. It was much more flexible than an ordinary project."
Dr. Lue says the animation has received a "tremendously enthusiastic response from students in both freshman biology and sophomore cell biology courses. Furthermore, preliminary evaluation shows that using animations as a part of their study resource enhances performance on questions requiring data interpretation followed by hypothesis building in the cellular context by almost 30 percent."
Through a Howard Hughes Medical Institute grant awarded to Dr. Lue, Harvard has contracted XVIVO to create several more animations for the same classroom series. "Both the quality of the final product and my experience working with XVIVO has reinforced my interest in a long-term collaboration with this team," says Dr. Viel. Liebler is already at work on the next one. "In the past couple of days I've been digging into the material to once again get my head around the scientific concepts and figure out how much we can carry over from the previous project to make the next one go even more smoothly," he says.
More recently, he says, the team has started bringing SOFTIMAGE|XSI into the mix. "We've figured out how to get the two programs to talk to each really well. There's also something that I learned to do in LightWave during the Harvard job, involving global motions through displacements, that I canÃ¢Â€Âšt do any other way." Liebler discovered what he now refers to as a "magic plug-in for instancing," Happy Digital's HD Instance, while scrolling through an online forum a few years back. "I can fill up a scene with thousands of things that I couldn't render any other way. A lot of the shots in the Harvard project are single-pass render shots, where there's lots of stuff going on in the foreground and the background at once. We got it all in there because of instancing."
For the Harvard project, Liebler also relied heavily on PDB (Protein Data Bank) files, which contain XYZ coordinates for all of the atoms in a protein. "PDB files tell scientists a lot more than they tell me. But as far as getting the shape of a protein, which has a known form, it is indispensable for an animator," he says. Developer Sean Hyde-Moyer has written a PDB reader for LightWave, and though a version of it comes with LightWave, Liebler recommends downloading Hyde-Moyer's free version (v.2.85) from his Web site
Liebler says he's looking forward to seeing the three-minute excerpt of The Inner Life of a Cell, as well as the sheer breadth of work by other animators, together on the big screen at SIGGRAPH. Will he be taking notes? "We probably have a lot to learn from each other," he says. "Character and scientific animation are really more closely related than some might think. The principals of character animation are not always seen in scientific animation, but they definitely can enhance scientific animation, making things look more organic and believable and have weight. Can it go back the other way? You know, I think it certainly would be a good cross-training exercise for character animators to try their hands at scientific animation for a while. They probably would get a lot of new techniques back that they could use in other projects." Liebler certainly has. Because he's so often asked to visualize the previously unknown and unseen, he's on a constant uphill learning curve. "Most of the things you're asked to do you haven't done before. It's really the most challenging thing about scientific and medical animation. And it's hard to even find anybody else out there who's done the things you haven't done before. Character animators know they can go online in the community and find somebody who already solved the problem they're struggling with. But with medical animation, you really have to think on your feet and figure out workarounds for your own problems. That's good practice for any animator."