What’s Animating CG Scientists?
Studios turn to scientists to keep the technological cutting edge for the tools their animators use to create art. By Ellen Wolff.
It’s been 75 years since visitors to Oz were told: “Pay no attention to that man behind the curtain!” But in the modern world of computer-animated features, it’s increasingly evident that those scientific wizards are vital to the dazzling images seen on screen. Storms of fire and ice, raging floods and characters with windblown hair and feathers are now plentiful — thanks to the way animation artists are empowered by the latest CG tools.
“It seems like the amount of simulation — and even things that are traipsing towards photorealism — are things that audiences are liking more,” says Pixar general manager Jim Morris. “Even if characters are stylized, there’s something about the surfaces and lighting and textures that is not so much ‘realism’ as authenticity.”
Morris should know, having previously presided over Industrial Light & Magic during years of technical innovation. But with many of today’s visual-effects houses needing to focus on short-term problem solving for their current clients, it’s probably not surprising that feature animation studios are becoming even more noticeable drivers of digital R&D.
“Animation companies have to be more long-term focused,” says Ray Feeney, a longtime member the Academy’s Sci-Tech Committee and a four-time winner himself. “These studios require multiple years to produce their films. The internal architecture of animation companies allows for a certain amount of scientific brainpower to be brought to bear.”
This year, DreamWorks Animation scientist Ron Henderson and Pixar’s Thomas Lokovic and Eric Veach earned Sci-Tech honors for research that has enabled imagery in their studios’ movies. Alongside Henderson, notes Lincoln Wallen, DreamWorks’ chief technology officer, there are more than 100 people in the R&D group focused on movies.
“There’s not a great deal of visibility of the high-performance engineering that actually goes into making animated movies,” says Wal- len. “But one reason that major technology corporations like Intel or HP work closely with us is that we’re solving problems that other industries have yet to encounter. For example, DreamWorks moved its entire production process to a cloud-computing model in 2002, even before the word ‘cloud’ had been identified.”
The need for scientific innovation has been essential to CG animation’s success since its inception. When DreamWorks’ PDI branch entered the feature business with 1998’s Antz, the fluid simulation expertise of Sci-Tech winner Nick Foster helped achieve the film’s climatic flood. The following year, Blue Sky Studios won the Animated Short Film Oscar for Bunny, which showcased gorgeous images rendered with its proprietary ray-tracer, CGI Studio.
Blue Sky co-founder Carl Ludwig, who with Eugene Troubetzkoy spearheads the studio’s 14-person R&D group, credits animators for continually challenging his technical team.
“The nice thing about being in the movie business is that it’s the greatest beta site in
the world,” says Ludwig, who’s also a SciTech winner. “Artists come up with things we wouldn’t dream of. We’re constantly upgrading, averaging about a software release per week.”
He estimates there’s a 50-50 split between working on future-oriented R&D and solving immediate production problems. “It’s important to identify things we don’t know how to do yet, because in research you have to be able to fail. You have to take on risks because that’s the unexplored territory where you find new things.”
Having a research mentality as part of a studio’s DNA also famously describes Pixar, which has been led since its inception by Oscar-winning CG pioneer Ed Catmull. That Pixar has a sister software company — RenderMan, which monitors the technical pulse of the industry — puts the studio in a unique position. As Pixar senior scientist Tony DeRose says: “We’re constantly communicating with the RenderMan folks. It’s almost like getting ‘matching funds’ to work on projects.”
While DeRose’s in-house R&D group typically numbers fewer than 10 people, Pixar also employs around a 100-person technical team. DeRose, who won Sci-Tech honors for 1997’s Gerry’s Game, sees no shortage of challenges. “In simulation, we’re not there yet. We want to get to a stage where, ultimately, we can do physics fast enough that animators can see things essentially in real time while they’re animating. Right now, when characters are animated, it’s often with impoverished clothing — if any clothing. There are really no effects and very little lighting. One of our big technical challenges is to get to a stage where animators can see a final performance in context.”
Making computer animation more interactive for artists is a goal shared by all studio technologists. DreamWorks’ Wallen says, “We’ve gone from hand-drawn animation, where animators could immediately express ideas, into the engineering hell of CG manipulation. Now they’re back to an analog penon-screen, which bodes extremely well for the medium going forward.”
Ludwig agrees. “I don’t want Blue Sky’s artists to be clerks twiddling dials. The tools have to be intuitive.”
One key area where this is happening is lighting. As Ludwig explains, “Our code is highly threaded, so an artist can take a halfres frame, basically change one light and all the shadow and lighting calculations on that half-res frame are recalculated in two to five seconds.”
At Pixar, notes Morris: “One of our big projects is a setup where you can move a light and it immediately shows you the effect at low res. As you let it sit there, it up-reses so you can see all the details. Real time everything is the Holy Grail.”
How soon CG scientists could achieve that goal is an open question, but all the major animation studios expect to maintain staffs dedicated to that level of research. “It continues be a cost-intensive part of our operation,” Morris says.
But being able to attract top young scientists to the movie business might be more challenging at a time when Google and Facebook are also seeking talent, says Feeney. “The allure of this industry is up against some fairly strong competitive pulls. Studios will need to keep their voices active in the ‘sciences’ part of Motion Picture Arts and Sciences. If they do that, then they’ll be able to foster scientific creativity going forward.”