• Blogs
• Forums
• Tradeshows
• Partners
• White Papers
• Podcast Show Notes
• Events

 

Water, Water Everywhere

Mar 1, 2008 12:00 PM, By Ellen Wolff

The Sci-Tech Awards honor believable sims.

 Web Expanded
Image Gallery from Sci-tech Winners

If the Academy of Motion Picture Arts and Sciences' Scientific and Technical Awards had a mantra this year, it would be, “Water, water everywhere — and not a drop to drink.” With singular focus, the Academy turned its attention to the digital-fluid simulations that now swirl through movies, honoring six groups of software innovators at its annual February event in Los Angeles.

Fluid sims are proliferating in movies as directors envision ambitious sequences such as the sea battles in Pirates of the Caribbean: At World's End. Industrial Light & Magic (ILM) and Digital Domain contributed to that Oscar-nominated effects film, and teams from each studio received Scientific and Engineering Award plaques for their proprietary systems. The Academy's Technical Achievement certificates also recognized Rhythm & Hues' inhouse code, along with the developers of RealFlow, Autodesk Maya Fluid Effects, and Flowline software.

“One of the cool things this year is that the Academy is recognizing a lot of work done in the decade since Titanic [1997],” says Doug Roble, who shares the award for Digital Domain's Sim code with Nafees Bin Zafar and Ryo Sakaguchi. While Titanic famously employed a physical model, Roble says, “Tanks and models are very expensive when it comes to fluids. The basic rule of thumb is that fluids start to look stupid at about one-third scale. For Dante's Peak [also 1997], they flooded a valley with one-third-size cars, and it cost a bazillion dollars. That's when the bean counters said, ‘There's gotta be a better way.’”

For ILM, it found a better way in collaboration with Stanford University, whose PhysBAM (physically based modeling) engine helped the studio simulate monster waves for Poseidon (2006). That movie tanked, but the ILM-Stanford approach demonstrated that high-resolution simulation could be done on a production schedule. Previously, fluid simulation was so computer-intensive, it could take days to process a single sim. “We worked hard with Stanford to run our simulations across multiple machines,” says ILM's Nick Rasmussen, who shares the Academy prize with Frank Losasso Petterson and Stanford professor Ron Fedkiw. “By splitting sims into pieces and having our computers talking to each other, we were able to transcend what could fit on one machine.”

Poseidon also showcased simulations created with Flowline, the award-winning software developed at Munich-based Scanline Productions by Stephan Trojansky, Thomas Ganshorn, and Oliver Pilarski. That sim tool allowed the London-based shop Moving Picture Company (MPC) to convincingly flood the Poseidon's interior. Scanline then went on to sim the “liquid battlefield” in 300 (2007).

Fluid sims are becoming prevalent today, but moviegoers got early glimpses back in 1995's Waterworld, which reflected the research of Jerry Tessendorf, another of this year's Sci-Tech honorees. Tessendorf, who also contributed to the digital surface water in Titanic, was honored for the fluid dynamics software at Rhythm & Hues (R&H) — co-developed with Jonathan Cohen, Jeroen Molemaker, and Michael Kowalski. “One of the biggest challenges of simulating water is that we know a lot about what it looks like from everyday life,” Tessendorf says. “If you don't get it really close to right, it looks strange.”

Getting it “close to right” is a multipart problem — from the simulation of gross wave motion through secondary simulations of spray and mist. “One technical approach doesn't give you all the answers,” Tessendorf says. “When engineers research fluid dynamics to build bridges, they don't take one approach because it's an extremely complex problem. The Academy awards are very similar. … [They] cover a gamut of techniques.”

Jos Stam, an honoree for Autodesk Maya Fluid Effects, agrees. Stam has been researching simulation problems for a decade, and he has been recognized along with Duncan Brinsmead, Julia Pakalns, and Martin Werner. “Everyone has attacked the problem from different angles, although there are some things in common,” Stam says. Maya's approach defines a fluid as a grid comprised of interdependent cells that determine overall motion. But there's a computational limit to how fine you can make this grid. “That's where you add particles on top to ‘decorate’ the surface. The combination of grids and particles is the thing to do,” Stam says.

Continue the discussion on “Crosstalk” the Millimeter Forum.
© 2009 Penton Media, Inc.