DIRECTOR'S CHAIR: HOYT YEATMAN - 'G-FORCE'
HOLLYWOOD — The plot of G-Force may sound silly — an elite team of talking guinea pigs are trained as spies by the US government — but producer Jerry Bruckheimer has the Midas touch, and the 3D live-action/animated project quickly attracted top-notch talent, including Penelope Cruz, Nicolas Cage, Will Arnet and Bill Nighy (who starred in another little Bruckheimer project, Pirates of the Caribbean).
At the helm, and making his feature directorial debut, is another frequent Bruckheimer collaborator — Hoyt Yeatman, one of the original founders of Dream Quest Images, an Academy Award-winning visual effects company. His miniatures and underwater bluescreen photography in the 1989 film The Abyss won Yeatman an Oscar for Best Achievement in Visual Effects, and he was also visual effects supervisor on the 1998 Academy Award-nominated Mighty Joe Young, as well as The Rock, Crimson Tide and Armageddon, all produced by Bruckheimer.
Here, in an exclusive interview with Post, he talks about making the film and his love of post and visual effects.
POST: What sort of film did you set out to make?
HOYT YEATMAN: "In my mind it was always a hybrid: a mix of live action and animation, and my background in visual effects and having worked on many of Jerry's films — all the explosions in Con-Air, underwater stuff in Crimson Tide and so on — was a big help.
"When Disney bought Dream Quest back in '96, I moved to feature animation, where I was introduced to character animation, and I worked on films like Mighty Joe Young, and they were called visual effects, but they really revolved around a main character that was digitally created. I really felt that animation could be much bigger and broader than how it was defined by the major animation studios. We did Kangaroo Jack for Jerry, which was a perfect example of where a digital background character turned out to be one of the main characters. I think that's when Jerry first got the idea of character animation being a central element in a film.
"So when I developed this, I took it around as an animation project and found very quickly that there are very few companies interested in something like this, and most deemed it not animation. Then we went to visual effects houses, and they said, 'Gosh, look at all the animation!' So we were stuck… until we went to Jerry. He had a very open mind and no preconceived ideas about what it should or shouldn't be, and embraced the concept."
POST: When did you decide to go 3D?
YEATMAN: "We were going to begin a very intensive hybrid film, but the twist was that Disney and Jerry both wanted to go 3D, and most of the 3D films done to date are fully animated. And in a CG world, it's not that hard to add another virtual camera and render a second eye, but in the real world, you're dealing with a lot of physical things.
"Our problem is you're down at nine-inches high with our main characters, and Mooch the Fly is only 1/4-inch big, so we're talking about sizes that don't fit comfortably into 3D. So we did some preliminary tests using some advanced 3D camera rigs. I shot two days with the Pace rig that Cameron built for Avatar, and it looked great, but I just felt the technology was limiting us. There was no way I could use a fisheye lens or travel three-inches off the deck at 50MPH. So the studio at that point deemed it not feasible to go 3D and said, 'Do it 2D,' which I was very happy to do. So we began shooting with the new lightweight Arri 235, which gave us the angles and mobility we needed, and then half-way through the shoot, Jerry came to the set and said, 'We need to do this 3D.'"
POST: That must have been a shock. What were the biggest challenges facing you?
YEATMAN: "I'd done the visual effects for Underdog right before this, and I'd experimented a lot with 'dimensionalization,' a process coined by In-Three (http://thedimensionalists.com). Much like you can colorize a film, you can dimensionalize a film, and I'd done tests with this process and found it worked really well. When you do a lot of visual effects in true stereo, it's extremely hard, since a lot of the trickery we use to create the illusion works well in 2D but it's often given away in 3D, because you can tell the perspective's not right. So you have to be very careful, and doing a whole show like this with as many shots as we have is very tough. So we decided to use different basic techniques to generate the 3D aspect.
"One is In-Three's, which is basically a 2D approach where they trace different parts and dimensionalize it. Two is Sony Imageworks', which did both the visual effects and the animation. We build phantom objects that go into the scene that the live action is projected onto, like the bus in the haunted mansion at Disneyland, and again, it works pretty well. And by doing that, it allows us to control the 3D-ness of a scene. We know where everything fits in perspective. So the addition of a CG character, like a guinea pig, is very straightforward. And one key thing is that we can now control all the stereo decisions in post."
POST: That must make a big difference?
YEATMAN: "Exactly. It's the biggest difference between the way we approached this and other films. Most others that shoot in stereo have to make those stereo decisions at the time of shooting, and this has nearly 2,000 cuts in an 85-minute movie, while a Pixar film probably has 1,100 shots. This is much faster-paced — it's a Bruckheimer action film! So if you're making stereo decisions in the field, by the time you cut you'll often have a mess, as it becomes an eye-strain to watch. You're not choreographing in the way the eye moves through free space, so it's much better to wait until you get your cut down and then build a stereo depth grade, like you would if you were doing color in post. We're also releasing a 2D version, so it has to work for both formats. This is where you get into some storytelling issues and concepts that don't work in both formats.
"For instance, if you're doing a close-up of a guinea pig in a 2D film, you have minimal depth of field because he's small, and we were true to that in the way we did our visual effects. So for an emotional scene with the focus on his face, you make sure his eyes are sharp and not worry about the background so much. But in 3D, you find the same scene's very disturbing, as you begin to approximate real vision and your eyes want to see a more focused background. So when we generate our characters, they're up to 60 elements deep, meaning that all the lighting passes are broken out in different pieces, and because we can put the composites together in different ways, we can control where the depth of field or focus is."
POST: So you're essentially making two different movies?
YEATMAN: "Yes, one version for the 2D reel, the other for the 3D reel, so it can get very confusing. And between the two, we had several thousand visual effects shots."
POST: Did you use virtual sets?
YEATMAN: "Yes, as sometimes our characters are in a huge action sequence, or the complete reverse, just a conversation. And if you're plate-based, meaning you actually shot locations and set pieces, many times your lensing or camera motion becomes limiting to the animators in post, so many times part of a sequence is plate-based and part is virtual. When all the characters meet for the first time in the pet store, the whole mid-ground section is virtual, while the store and live action is all plate-based. That's all composited together, which [leaves] our virtual cameras later on in post, when we're animating the sequence, free to do whatever we want."
POST: The film is edited on Avid by Jason Hellmann (Enemy of the State, Bad Boys II) and Mark Goldblatt (Academy-Award nominee for Terminator 2: Judgment Day, Armageddon). Tell us about the editing process.
YEATMAN: "It's a two-year process, and the way Jerry makes a lot of his films is in post. We shoot lots of material and start editing, and when you're doing animation, you shout 'Action!' and nothing happens! You literally have to draw all the frames, so there were unique challenges on this. Before an editor can do anything, you have to draw it all, so there was heavy preproduction, as with most visual effects films, and after production there's also a lot of boarding back in, as certain sections are being redone. So between the two editors, some of it was pre-editing in a virtual previs environment, other stuff was happening after the fact, and again, as we put the scenes together, we may find that we wanted new lines or jokes. And as it's a virtual scene, I can do whatever I want. I'm not limited by re-shoots.
"So out of the two years, maybe a year is preproduction and production, and the rest is all the rendering and animation, and we're also a hybrid in the way we turn our reels over. We're working in animation, which means you're within four frames of the cut. So you're looking at storyboards, a combination of animatics, and Jerry and his group are very used to working in visual effects, and we don't stay in storyboards very long before we get into fairly sophisticated animatics.
"But in animation, it's still better in some cases to actually draw a conversation between two characters than try an animatic. So it's a very mixed media, in the creative phase. I used puppets on set to help block out scenes, I've got a lot of empty plates, drawings and animatics. So as you watch it develop, it takes a lot of creative energy to really judge if the story's working. Same thing with recording the voices. And none of this was motion capture. It was all done through keyframe animation, completely hand-animated in a classic sense."
POST: It's an interesting hybrid.
YEATMAN: "Yes, not just how we shoot but how we post and animate. It's really taking the world of very sophisticated visual effects and the world of classic animation and putting them together to get a very different look."
POST: Where did you do the post?
YEATMAN: "All the visual effects/animation was done at Sony Imageworks. Then we did some matte and optical work at Asylum, and N3 did their thing, and Tim Sassoon did some 3D conversion. It took hordes of people, but the process worked really well."
POST: This is your directorial debut. How did your visual effects background help on this?
YEATMAN: "I'd directed commercials before and done special venue projects, so it helped enormously. I felt very comfortable and I also had a great team, including a friend, two-time Academy Award-winning visual effects supervisor Scott Stokdyk [Spider-Man, Spider-Man 2], and we thought about solving problems exactly the same way. It's been the most difficult thing I've ever done, but I'm so happy with the way it turned out."