Puzzle Shots, the 50 Percent Rule, and Deep, Fully CG Environments
By Barbara Robertson / Feb 10, 2014
This year’s Oscar nomination is the first for visual effects supervisor Tim Alexander, although far from his first major award. Alexander received a BAFTA award in 2001 for The Perfect Storm followed by two BAFTA nominations for two Harry Potter films: The Goblet of Fire and The Half-Blood Prince. In addition, he received a VES nomination for Half-Blood Prince and, later, a VES award for supervising the visual effects in Verbinski’s Rango, which won an Oscar for best animated feature.
Alexander began his career at Buena Vista Visual Effects where he was a lead digital compositor on the 1996 film James and the Giant Peach. He joined Industrial Light & Magic to help composite the 1996 film Star Trek: First Contact, and has remained at ILM since. He became a CG supervisor on Star Wars: Episode 1 – The Phantom Menace (1999), an associate VFX supervisor on The Perfect Storm (2000), and soon after, VFX supervisor. His work as a VFX supervisor has taken him from Hidalgo to Sky Captain and the World of Tomorrow, from Harry Potter to The Spiderwick Chronicles, and from Rango to, now, The Lone Ranger.
Alexander shares his nomination for The Lone Ranger with visual effects supervisor Gary Brozenich of the Moving Picture Company, visual effects supervisor Edson Williams of Lola Visual Effects, and special effects supervisor John Frazier. The Walt Disney film also received three Visual Effects Society nominations, for Outstanding Models and Miniatures, Supporting Visual Effects, and Outstanding Models.
StudioDaily: You must be really excited about receiving your first Oscar nomination.
Tim Alexander: It has been just awesome. The number of people who have reached out and congratulated us. The surprise factor. It is hard for people to see what was done in this movie, and at the bake-off you can’t really show anything. You describe the work and show the reel. So, I’m happy and honored that the other members of the Academy recognized what we had done and voted for us.
All 10 films up for the nomination had great visual effects. Why do you think your peers choose to honor the effects in The Lone Ranger?
I think one of the things is that we were definitely different that night. We were the only movie that takes place on earth. All the others were sci-fi, fantasy, [or] space movies. We were the only one that was grounded, and we presented in the middle, so I think it stood out. Our approach to the effects was that we were trying to hide everything. When you look at what Gore [Verbinski] did with the shot designs and the shots we implemented, it’s pretty innovative. We were transitioning between four or five plates in one shot, so we had to make that seamless. And then we had the whole third act with the virtual environments. It’s aggressive, hard work, and it’s hard to see.
Also, people like to see a special-effects component. We tried to use the right tool for every shot, so we relied on special effects. That probably stood out from the other films at the bake-off, too.
How many visual effects shots were in the film?
We had 1350 shots. It was actually a big visual-effects film.
So, it was a mix of live-action special effects and digital visual effects?
We had a 50 percent rule. For every shot, we tried to figure out how to shoot 50 percent in camera. I was involved in the planning stages before we started shooting, and Gore [Verbinski] insisted that, because it’s a Western, it had to look real. That was his mantra: “This has to look real. If it looks fake we lost.” We didn’t want people to know we couldn’t shoot people on a train.
But the old Westerns often had people on top of a train.
We looked at a lot of that stuff. But in those old movies the train didn’t have the speed and they were kinda boring. Some of the fastest trains were moving only 25 or 35 miles per hour and that does not look fast on top of the train. That’s why we put the train on a truck.
The train was on a truck?
John Frazier’s team built two 1860s, full-scale locomotives from the ground up. One could run under its own power and pull 10 train cars behind it around a track. We mostly used those for the desert environments. When the train moves toward the forest, we put the same trains on flatbed trucks and drove down the twisty roads. That’s where the CG came into play. We could have only one car on the flatbed, so we had CG train extensions. And when we realized how difficult it was to drive through curves at 35 miles per hour with a train on a flatbed and actors on top, we parked the truck and created digital environments. The 50 percent rule fell down when we couldn’t get the performance for safety reasons.
Did making it real make post-production especially difficult?
This was one of the most complex films I’ve worked on. During the planning phases, we’d look at the animatics, figure out how to shoot them and how to implement them at ILM. It was complex not only in planning, but later in implementation. We shot the film in pieces with large CG components that we had to integrate. We had to match things in close-ups and to bridge plates that didn’t match each other. Our attempt to get as much live action as possible put a burden on the special-effects team that had to do large-scale train effects, and then it put a burden on ILM to patch everything together.
How did you start patching it together?
We went through a long layout phase, longer than most, where we massaged the elements into place. Gore [Verbinski] is very precise. He had previs'd all the train sequences, and he wanted the film to match the previs. Every shot was a puzzle. Where do we bridge with CG? Where do we match the background?
Can you give me an example of a “puzzle” shot?
There are a couple shots in the first act with a bunch of buffalo on a range. We start out wide with a helicopter shot; the train is at a distance. The camera flies into the train, down the side, and into a window for a close up of the Lone Ranger sitting in his seat. The close-up on the Lone Ranger at the end had to be rock solid.
We shot it in pieces. We flew as many passes from the helicopter as we could, and that took us 20 feet from the train. We shot a second plate of the Lone Ranger sitting in a parked train car. We were in a vehicle. We started close-up on him, pulled out and quickly accelerated in reverse trying to get as far out from the train as we could.
So we had those two plates, one a wide shot, one a reverse shot. We had to create a big chunk in the middle to hook up those plates. We tied the two plates together using a CG train to help. And in the second plate, the environment was static, so we had to move the ground, put the environment in the windows, add the reflections and the buffalo. We replaced the environment, made it move, the whole shebang. That was a two-plate situation. We had another shot in the first act that was a five-plate situation.
A five-plate situation?
It was similar in that we started wide with the train coming at us. The camera was sitting on the track. Then, the camera runs up the side of the train. We lock in with Tonto and the Lone Ranger, and we travel with them.
That’s four plates.
They found another plate in editorial that wasn’t supposed to be in the shot at all. The camera and lighting were completely different. There was obviously a lot of compositing time, but a lot of the heavy lifting happened in our layout phase with our match movers and layout artists figuring out where everything went.
You said the train was sometimes CG?
For hundreds of shots. We had extreme close-ups. On set, they couldn’t pull the brake and have the wheels spark. It was dangerous. So we did close-ups of wheels braking, squealing. We had train extensions and a train crash in the first act, so we saw our train quite a bit there. It was primarily in the third act with the digital environment. In the third act when the characters are running on the train, probably 80 percent of the time they were digital doubles. They did use a lower height on a platform that we comp’d on top of a train for some shots, but not that many.
On set, they had a stuffed horse they’d ride for close-ups when we would see only the neck of the horse. Otherwise, we also had digital doubles for the horses. And we did the buffalo [and] the killer cannibal rabbits.
Was the shot of the train crashing off the bridge all CG?
No. All the punch-ins are miniatures we did with 3210 Studios. We try to use the right tool for the right thing — and we had the bridge, water, and fire. We looked at CG and miniatures and decided to go with miniatures and digital backgrounds. The super-wide shot is all CG.
Why do you consider the environmental work innovative?
It’s the scale and the look. These are probably the most realistic, close-up environments we’ve done. We’re on earth in a forest and everyone knows what that looks like. There’s no room for sci-fi-ing it up. We often put trees in environments, but we’ve never created an entire digital environment that we travel through with the camera five feet from the trees. We didn’t use cards. We put full 3D trees in the background because, frankly, we could tell the difference. We made fully deep environments with fully CG trees, rocks, things that go with it. And, the trees had to move.
What tools did you use for the environments?
Our set dressing tools are based on layout tools that we started working on for Rango. We created tools for set dressing that let us scatter thousands of trees in an organic way and add rocks and other elements. If you just go in and separate elements by one foot, you get a grid and it looks like a grid. So, we could control all that. We used a lot of off-the-shelf packages to create the trees. We rendered the environments with VRay, rendered the train with Arnold, and rendered the rabbits with RenderMan.
How did working with Gore Verbinski on The Lone Ranger differ from working with him on Rango?
This film was so different from Rango. Rango was hard in different ways. With Rango, we were in the studio where we could control all the environments. Because you could control everything, it became a very precision piece. There was a lot of stuff to move around and make perfect, but once it was set dressed and the animation done, it was a cohesive thing.
This film is live action. We were out in the wilds in Four Corners — New Mexico, Colorado, Arizona, Utah — shooting for nine months, dealing with weather. Gore was out every day working with all the department heads, not just visual effects. On set, our task became how to disseminate information about the previs to everyone. Then, in post-production, we put the shots together trying to achieve the previs. So it hit on a lot of tasks that were different from Rango. But working with Gore was similar. Gore doesn’t go in circles. He always aims at one point. He’s very specific, very precise.
This is an unusual Gore Verbinski film. There’s no monster, no Davy Jones. The one big effect is the train. There’s no showcase piece.
So, you’re carrying the banner for invisible visual effects this year?
I’m overall excited that the invisible work is getting some recognition. So many people put their hearts into it. I’m really happy they’re getting recognition.