Pixar's Tony DeRose Talks Math and Hollywood Prior to JMM MAA Invited Address
Tony DeRose is Senior Scientist and lead of the Research Group at Pixar Animation Studios. Before joining Pixar in 1996 he was a professor in the Computer Science and Engineering department at the University of Washington. In 1998, he was a major contributor to the Oscar (c) winning short film "Geri's game", in 1999 he received the ACM SIGGRAPH Computer Graphics Achievement Award, and in 2006 he received a Scientific and Technical Academy Award (c) for his work on surface representations.
DeRose will be presenting the MAA Invited Address “How Mathematics has Changed Hollywood” on Wednesday, January 9, 3:20 p.m. - 4:10 p.m., at the 2013 Joint Mathematics Meetings in San Diego, California. He recently answered some questions from MAA staff about his upcoming talk.
What do you do as the senior scientist and lead of the Research Group at Pixar Animation Studios and what kind of problems do you get to tackle?Our group was formed in 2005 with the mission of getting out in front of the needs of future films and developing new technology in anticipation of those films. There’s a lot of innovation going on in different parts of the company, but, in some sense, it’s kind of “innovation on demand,” and our group is designated to basically be innovating all the time.
The way we think about the difference between projects that we tackle and projects that other groups in the company tackle is we look for the risky bits. A project might be risky because it has a long time horizon—it’s going to take a while to work on. On a particular production, there is some innovation that is done for each show, but they only have about a year to do their preproduction development and then the production pressures themselves really kick in. Any problems that have a time horizon of more than a year aren’t going to be picked up by individual productions. Those are some of the opportunities we look for, those longer-term, harder problems.
We also work on problems that have a shorter time to delivery, but the potential for failure is high enough that a producer probably isn’t going to pay for it on their budget [laughs]. So that’s another kind of risk that our group is really set up to work on.
A recent example is the hair that Merida in Brave had. Her hair is basically another character in the film. Getting that full-bodied, soft head of hair that behaved the way the director wanted was simply out of our reach at the time that Brave began production. Some members of my team put together a completely new hair simulation system for her, and it looks like that simulation system is going to continue on well beyond Brave and will pretty much be used for hair in all our future films.
[Read Inside Science’s article “Brave Features Hair-Raising Animations”]
We’re working on a number of different things. One of the things we’re working on is to try to make the motion of our characters more physical, especially human characters. Developing more of a sense of an underlying musculature, especially for fat characters [laughs]. It turns out to be quite difficult for us using current, more-nonphysical techniques to get some of that jiggle and volume preservation and really physically plausible deformation. We’ve got a project going to try to bring more physics into our character movement, and one of the challenges there is getting the algorithms to run fast enough. We’re about a factor of 1,000 away from being able to do it in real time.
Another place where we’ve got a big computational gap is simulating the lighting effects that go into the final images. There are a lot of really intricate physical phenomena having to do with the way light scatters around in an environment that is computationally very challenging. We’ve been working on this for a number of years. Typical final frames take about four to five hours, and we’d like to make that real time. We’re very close to being able to do that now.
In 2009, you gave a talk about “Math in the Movies” as part of MAA’s Distinguished Lecture Series. Without giving away too much about your upcoming JMM talk, what significant changes have you seen in the time since?
In 2009, our ability to capture physical effects was pretty limited. We had fairly simple attempts at things like physically simulating cloth and hair, but the lighting simulations we were doing at the time were pretty primitive by standards today. Some of the biggest advances have been in the area of physical animation.
[Read about DeRose’s 2009 MAA Distinguished Lecture]
Have you seen a growth in these techniques outside of Pixar?
Rewind 15 years or so, special effects in films would sometimes be done with actual physical devices and explosions. That’s almost completely gone now. Just about all special effects are digital because it’s cheaper and safer and faster. There are a lot of films that don’t even seem like special effects films in which computer graphics techniques are being used to a really great extent. Looking forward I think that trend is going to continue to where it may become rare to even have physical sets where the acting takes place. You know, maybe virtually all the acting gets done on a green screen stage and all the physical world around the actors is added digitally later.
I assume that the story of every Pixar film is the starting point in the creative process, but have the advances that you and your team made influenced what the story writers consider when conceptualizing a new film?
That happened to a greater extent in the earlier days of graphics. These days, I think the directors have become so accustomed to the technical staff being able to do whatever they ask for that the directors feel like they have pretty free creative rein. It occasionally happens that we can’t completely realize the director’s vision so it has to be scaled back or modified in some regards. The extent to which the technology influences the creative decisions has diminished. There are some interesting places inside Pixar where there is some potential for new technology to really let directors branch out in a new direction, but it’s a little too early for me to talk about that.
What’s your favorite Pixar movie?
I’m particularly fond of The Incredibles. I’m a James Bond fan, and that kind of runs like a Bond film. I love them all, but I kind of have a special place for that one.
What’s your favorite Pixar movie based on the mathematics behind its animation?
I would have to say Monsters, Inc. because that was the first time that a lot of techniques that I had developed saw widespread use for surface modeling, creating the geometry of their skin, faces, hands, and clothing.
What advice would you give a college math major who is interested in animation?
We need people that are mathematically sophisticated, especially in areas like linear algebra, numerical analysis, differential geometry, and differential equations. We also need people that can implement their ideas in software. It’s really important that candidates are well versed in reasonably modern software development, being able to implement in languages like Python and C++.
Have you ever attended a JMM? Are there any sessions or talks you’re looking forward to?
This will be my third. I particularly enjoy the student sessions. I’m really looking forward to the Who Wants to Be a Mathematician game. I’m really interested in educational outreach, and that’s one of the reason I do these talks. Those are the kinds of activities I tend toward at JMM.
DeRose will present his MAA Invited Address "How Mathematics has Changed Hollywood" on Wednesday, January 9, 3:20 p.m. - 4:10 p.m., Room 6AB, Upper Level, San Diego Convention Center during the 2013 Joint Mathematics Meetings.