VFX Supervisor Scott Farrar on Transformers: Dark of the Moon
The film’s Oscars VFX bake-off reel, as uploaded by MichaelBay.com.
In the world of cinema, motion pictures, visual effects, stunts, physical effects, and all the other departments required to do movies, it doesn’t get much bigger than this. And it’s a heckofa lot of fun. I’ve worked with the same guys for six years now, all working together to improve our craft. Each film is new and different, and Michael Bay gives us tremendous freedom to explore. If you want to be involved in creativity in the world of visual effects, this is the place to be. It doesn’t get much better than this. And that’s true for Matt Butler and the great team at Digital Domain, too. The relationship with Michael Bay, producer, Lorenzo di Bonaventura, and all the people at the studio, is the best I’ve had with anyone bar none.
F&V: Can you give me an example of “tremendous freedom to explore?”
[ILM's] Scott Benza [animation director] and Rick O’Connor [associate animation supervisor], who knows the entire Transformers universe, pitched ideas to Michael Bay that Ehren Kruger incorporated into the script. That’s collaboration at the highest level.
F&V: What were your goals for the film?
A big goal for me has been making the computer graphics look better, to improve the look, to stop the feeling that these things are fake because they look “CG”. We can talk about the big numbers, but the artistry is in the lighting, the dexterity we show in the robot movements, the great composition, the shot design, the soaring, free camera movement that we exhibit, all that’s important. I went for a darker mood where sometimes you don’t see the whole character, and we used super slow motion in some cases to really show off the transformations and to see the nooks and crannies, the hundreds of parts. We have robots expressing as much emotion as possible with rigid pieces of metal and we got to explore that more than in the last film. We have cool animation and cool action, but beyond that, it’s always all about the light.
F&V: These sound like the kinds of things a director of photography would think about.
Exactly right. I want to make our work seamless and glued into the rest of the movie, so I use the same logic in how to set up shots. We can choose things like a technocrane and have dolly moves, but we can also have a really free camera that can do things we couldn’t before. We anchor to the tools used in the film business, but we can go freer. Some people don’t like a freeform camera. I personally like it. These are philosophical approaches and I think about that. But, no matter what the camera is doing, you really have to pay attention to the lighting.
I teach classes where I have the CG artists go down to our stage and have them move the lights around to light a car or a doll. I want them to control the lights to get a gleam off the hood, a nice bounce light off a face. It’s an eye opener.
[On location], we always record the environment on spheres, but after we come back, I might want to flip the sun to the other side of the street. Someone might say, “Wait. That’s not in the script.” But, if a director of photography were on set filming an actor and didn’t like the light, he or she would change it – block out the sun, add reflective boards Ã¢Â€Â“ and that’s what we do. I try to adopt the theatrical style of lighting people see in movies.
F&V: Do you have special tools for working with lights in that way?
We’re always building new tools, but it isn’t like we have a nice movie package that automatically puts reflective boards in to get the right angles. We’re still working on that.
F&V: In each film, the biggest robot is bigger than in the film before, with Colossus in this third film more than twice the size of the ginormous Devastator in the last film. By now, the crew must be very good at building big bots.
Everyone is getting better, but as they get better, the work gets harder. We had 24 new robots and each one takes about 12 to 15 weeks to build. At that point we have the beginning stages for paint and textures and can demo it on a turntable. Then, it takes another 15 weeks to rig. I think this was one of the largest modeling crews ILM has had. They started seven months before we began shooting.
F&V: Was Colossus the most difficult robot?
Sentinel Prime, the red guy with Leonard Nimoy’s voice that transforms from a big red fire truck, was really hard. And, the Ferrari character. Boy, red is hard to capture in CG. It can go purple, pink, orange, anywhere except where you want it to go from shot to shot. This is what’s curious about CG. Something as simple as trying to get a color that matches a vehicle is hard. You have to go back to really basic things. No matter how good you are, a red car photographed during different times of day under different types of lights will adopt a different color, and without a real car to photograph, it gets crazy hard. We made all these little car body shapes for every color of the robots. We’d run around with these shapes and have the camera photograph them for every shot in a location; every camera in every location for each color. You can’t believe the specular change of light for something as simple as a red car.
F&V: What was the hardest shot in the film?
We did a digital double of Shia [LaBeouf] for quite a few shots where he needed to be seen in close up and that was certainly difficult. To duplicate human skin, to make a CG model look like a real person is so complicated. The more you know, the more you realize you don’t know. Even when we had perfect video reference, you always have to go back to the source to make sure your mind isn’t fooling you.
We also had a lot of shots that were difficult because they were stereo 3D. Some were fun and great Ã¢Â€Â“ especially because you can see the density of the robots. It isn’t like a human face. You could see inside, all the nooks and crannies and shapes. It was amazing subject matter to put in 3D. But, it made things difficult because of the precision.
F&V: The robots are hugely obvious effects. Did you create less obvious effects as well?
We always known that whatever we shoot on location might have to be augmented or changed, and we’ve gotten adept at that, so we document every location with high megapixel stills to have photographic information. There’s a shot where Shia flies at the camera in super slow motion while Bumblebee whacks things out of the way. At the last minute, Bumblebee grabs him, rolls into a ball, bumbles along on the pavement, and then transforms into a car. We shot that in Long Beach using a high-speed camera car. As it turned out, we needed to alter the path of the camera later. We’d done our photography at the right time of the day, so we could rebuild the environment and change the path of the move. We do that quite a bit.
On a larger level, we created a mile of downtown Chicago on both sides of the river. It’s astonishing. It’s photoreal. If they ever tear it down, we have historical documentation of that riverfront.
F&V: How did you create this section of downtown Chicago?
We did a lot of photomodeling. We had a crew of about four people shooting every building in that area from top to bottom with stills morning, noon, and night in all weather; two camera rigs shooting all the time. They spent several weeks doing that. We didn’t do much Lidar scanning on this film. We created geometry from the photographs. We could fly a virtual camera up and down that corridor any place we wanted to, and we did. For the tilted building sequence, I shot aerials over the span of about six weeks and then we built an exact model of the building at 155 Wacker Drive. You see it in the film in its true location and in a couple other places, too. It’s really exciting.
F&V: Why is this so exciting to you?
We can do shots that couldn’t be done before and we can do them safely. If Michael [Bay] and Roger [Barton, editor] see that they’ve missed something after they have a cut, we can make that shot. It’s an awesome thing. It gives directors new freedom and possibilities.
F&V: What trends do you see for visual effects artists?
More freedom of expression. I think people are just starting to spread their wings. We’re becoming more nimble and agile in the things we’re trying. Sometimes, you’re stuck with something Ã¢Â€Â“ if you have a superhero with blue lightning, there aren’t many places you can take that. Fantasy is open to the suggestion and whims of designers. That’s why I like the photoreal stuff where you really fool the audience, where the audience isn’t sure what they’re looking at. That’s fun. I like magic. I like illusion. I like improving and polishing the illusion. That’s the attraction of Transformers for me. There is all kinds of stuff we can do. That’s what makes it great.
F&V: So, you’d do another one?
Absolutely. It’s a lot of fun. It’s fun because of how much creativity we get to utilize.
F&V: This is your sixth visual effects nomination. Do you remember the first?
It was for Cocoon, and we won. Oh, my gosh, it was like a shot out of nowhere. Talk about a bonus on top of everything else. I was just really happy to be in the business doing this crazy job. It was such a thrill. There’s nothing like it. And, it’s wonderful at this stage. It’s very rewarding to be recognized by your own group. It means an awful lot. I look at the array of work and they’re all great looking pieces and that’s cool.