Here's what didn't suck about the Oscars: seeing the wizards from Digital Domain accept the visual-effects award for "Benjamin Button."
DD's work here provides a glimpse of what's next in feature and commercials image making. I wrote about the film's visual-effects process a while back in Creativity. In the hopes that you're over your "Benjamin Button" and Oscar fatigue, here's a condensed version.
The central technical challenge for DD and director David Fincher, a longtime collaborator, was to ensure that Brad Pitt (rather than an animator's interpretation of Pitt) was actually playing the character throughout his life, while avoiding the dreaded "uncanny valley" (the creeps people get when looking at a being that appears almost but not quite human) that would have resulted had time-tested visual-effects techniques been employed.
The first thing to know: After the initial shots of the animatronic infant Button, the next 52 minutes of Pitt do not contain any actual Brad Pitt -- it's animation (until the real Pitt appears in makeup on the tugboat bound for Russia. A total of 325 shots of animation. Visual-effects shops Hydraulx and Asylum also contributed many elements to the film).
Fincher shot the film using small actors wearing blue hoods. Makeup artist Rick Baker created painstakingly detailed maquettes (silicone, plastic and hair models of Pitt's face) that represented the actor at ages 60, 70 and 80, and DD digitized the maquettes. Visual-effects supervisor Eric Barba and character supervisor Steve Preeg turned to what's known as FACS (Facial Action Coding System, a body of research that categorizes the full set of universal human expressions) as a starting point for capturing Pitt's head's performance. DD used the Mova Contour system to capture Pitt performing the gamut of facial expressions. "When we first looked at doing this movie in 2004, we had something in our original R&D plan that was not dissimilar to what they built," says Barba, who accepted DD's Oscar at the show. "When they presented their system we looked at each other like, 'They did it.'"
Kreeg worked with the company to develop the system to fit Button's requirements. Mova's marker-less system represented a significant advance in performance capture, creating a high-def, "volumetric" representation of a face (or other surface). Performers' faces are covered in phosphorescent makeup and captured by a multi-camera rig -- this eliminates the sort of interpolation that happens with a system that uses markers placed on the face to capture data.
Pitt (the whole Pitt) was also shot by multiple cameras while performing the role of Button and DD artists matched the CG expressions to his overall live-action performance. And then? Well, they retargeted Pitt's captured expressions onto CG versions of the appropriately-aged maquettes -- hand animating where necessary to get the nuances right. There was more -- DD faced monumental challenges in things such as tracking, compositing and lighting, in many cases building proprietary systems to do so.
DD is working on projects that will advance the tech used on Button. "It'll be difficult to do a less realistic character now," says DD Exec VP-Production Ed Ulbrich.
And yes, you will see the technology used in commercials. "With computer-based visuals there's a belief that anything is possible as long as you have enough time, money and resources," says Ulbrich. "In advertising there's just never enough. It's a big advantage for our advertising clients to have tools available to them that would otherwise be cost- and time-prohibitive to develop for a single spot or campaign."
~ ~ ~
Teressa Iezzi is the editor of Creativity magazine and Creativity-Online.com.