Iloura’s VFX supervisor tackles the effects for George Miller’s iconic post-apocalyptic adrenaline-filled feature, including a toxic sandstorm and evil warlord Immortan Joe’s stronghold, The Citadel.
Chills, spills, mutant vehicular contraptions Boris Karloff could only dream of, a flame throwing double-necked speed metal guitar player, and a one-armed Charlize Theron packing more heat than an NRA board meeting as Imperator Furiosa – George Miller’s Mad Max: Fury Road has something for everyone.
Helping bring all the mayhem and madness to the screen was Iloura’s Sydney studio, led by VFX supervisor Tom Wood. With two major sequences – a toxic sandstorm and evil warlord Immortan Joe’s stronghold, The Citadel – plus hundreds of stunts and vehicle crashes to contend with, along with an ever-expanding project schedule, Wood and Iloura traveled a long, difficult road to help bring this film to completion.
AWN recently had a chance to talk to the VFX supervisor about his role on the film, how Iloura ramped up to take on such a huge project and how his team handled integrating their visual effects into such a stunt-heavy live action shoot.
Dan Sarto: So tell me how you came onto the project? What were your duties on the film?
Tom Wood: I was working in London in 2012. I had just finished Mirror Mirror and was approached to help out as the vendor supervisor on Mad Max: Fury Road. Around August I traveled out to Namibia from London for a couple weeks to visit the production onset for a meet and greet. From there I moved with my family to Australia [to work on the film for Iloura].
The thing that struck me [initially] was that they [George Miller, the film’s director and his development team] couldn't give me any concept work, or any kind of broad brushstroke sketches of the toxic storm, which was the big thing they wanted us to start working on. There was scripted action, but nothing drawn. I got the concept guys together at Method Studios’ London office and we brainstormed and put together a whole book of concepts for how the storm could look. How the action could unfold inside. I showed that visual model to George, which he loved. He apparently doesn't usually like concept work. The storm images were pretty much the same in the movie as the ones that we took along on that day, on that trip.
It was a really, really good start for me to get into it with George. Hitting the right notes from day one. That really kept going through the whole movie, and we've had a great relationship, a great friendship, since then. I had a very easy ride. It's been a very, very long ride, but an easy one because we had a good relationship. I was always confident that we were hitting designs notes for him very, very quickly and easily. It's been 30 months, but it's been very, very smooth.
The other side of it is working for Iloura. The project [initially] was going to run for 10 months, and then it extends, and extends, and extends. That a hugely difficult thing to manage financially and operationally at a company which was pretty much a startup in Sydney. There was a very small office there before. We've taken in one of the biggest shows of the year to run it through that company. That was an intense management process to make the money work, to keep the team small and run longer. We had anticipated going very, very big but for a short time. That was also easy, but you were still always balancing George’s tastes against trying to make sure that the company still keeps people employed and makes some money.
DS: Even in the best of times, that’s no small feat. So you started as an independent working with Method, and then moved on to joining Iloura for the bulk of the production?
TW: I was independent. I was approached by Dan Glass, the chief creative officer at Method. My wife is Australian, and we wanted to come to Australia for a long time to work on a show. The Method office in Sydney needed help to do something as big as this. Someone a bit more heavyweight to help it through. I was approached and said, "Yeah, I'd love to." Because I'm also a collector of directors and George Miller was on my list...
DS: …You needed to put him up on your mantelpiece?
TW: Yeah, exactly.
DS: Quite the iconic film to put up there as your first George Miller experience.
TW: I should do that. I should put up little stuffed heads.
DS: My understanding is that while a tremendous amount of this film was shot practically, there was also quite a bit of visual effects work. How did you integrate the visual effects with such a large amount of stunt work and live-action shots?
TW: Colin Gibson is the production designer. He's a wonderful guy, and worked very, very closely with George and some vehicle designers for years developing things. George worked with his story guys. For the design work, Collin’s input was absolutely invaluable throughout, to create this complete dystopian world. I guess one of the big things is the detail within the back story to almost everything in the movie. Everything that you see. When you’re in the truck cab, the hipbone used for the gear shift, which turned out to be a huge knife, that hipbone has a back story. That's there for a reason, because it connects with something that happened earlier in a previous part of the story.
That kind of thing, when you understand that everything you see in the film, it run backwards somewhere to some significant previous story point. Then you understand that the whole world is very, very complete. In some movies that comes across, and in some, it doesn’t. I remember when Star Wars came out, I was completely sucked in because I completely believed everything in terms of the world they were living in.
In terms of then pulling visual effects into the film, the visual effects quota was really divided between things that just could not be done onset, like the toxic cloud or the front of the Citadel. Places where we just couldn't actually shoot what was designed and what was required. Those were obviously visual effects requirements. Those were things where in prep, they didn't really address anything to do with how that would look, or how that needed to be shot. They just shot from the hip.
So, in a way, we were coping with shots where we had to do a lot of work to make them work. Like within the storm or Citadel situation that hadn’t been shot perfectly before. Because onset was very much driven by the practical effects, the practical shooting techniques.
DS: Did you have any previs to work from?
TW: There was some previs, so we knew what the vehicles would be doing. Like for the storm, when the car gets lifted up and all the guys get torn off the War Rig. That wasn’t a hard thing to get around, the fact that they couldn't actually shoot it. The Citadel was planned out shot by shot, in terms of angles, as prevised. There were still frames [the previs], and they pretty much stuck to that. But at the same time there was no allowance for the huge rock edifices casting shadows across the crowds, where the lights were coming from to light that scene. That was the slightly short-sighted view of that sequence, though I think we managed to produce a fantastic looking result. I think that's partly due to...I've always had an approach to shooting where you allow things to shoot because shoots will have a momentum. They have a talent base onset that’s far more suited to filmmaking than the previs guys.
I was looking at some previs from another show a few days ago and it was just awful how they constructed the shots and story points. I just find that you have to have absolutely the right people on previs to make it work in a way where you can film it.
So those shots did get shot and we were then struggling to put in the proper shadows and light. I'm sure George wouldn't mind me saying that. The shots of the elevator lifting the trucks up, or lifting the people up where they're falling, those were a couple of beautiful backlit shots, with people falling, with dust and light rays streaming through. All that dust was put in really to cover the fact that the sun isn’t in the back of the shot. The actual back of the shot is a huge cliff. There was no way the sun could get there. When we were done everybody was surprised how different they turned out, how we managed to make these shots work.
DS: Were you onset during these shoots?
TW: I was on set briefly in Namibia for two weeks, and then I visited the Sydney shoot for a couple of days only. Andrew Jackson was the show supervisor and he had it all covered.
DS: Besides the two main segments of the toxic storm and the Citadel, where was the CG used?
TW: Apart from the storm and the Citadel, the rest of the movie is littered with visual effects shots. From the simplest, which would be vegetation removal from the background of the shot, where we've got a driving shot, through to complete environment replacement and digi-doubles. We only did...I think we only did 20 shots where we added a stuntman. There were stunts where we'd be removing harnesses. The guy being launched onto the buzzard car, killing himself, he required a really involved harness to throw himself over that. That was a massive paint-out.
When the Jaguar gets rammed by the excavator, and the two guys jump out and roll on the ground, they were shot static. The ramming of the Jaguar was shot on the move. Then they tilted up the Jaguar with two stuntmen on a static plate, and then as soon as they hit the ground they become CG to roll on the moving ground. There are a few huge combinations predominantly where we’re trying to go with a traditional 2D, split screen, or background element, foreground element, type of look. When Max is fighting the polecats, when they're surrounding the truck later on, he's on the top of the War Rig, and gets shot in the head. That was all a stationary vehicle, but with background plates around him. The wide shots of stuntmen, that’s all live on a moving truck. Trucks primarily were traveling at about 40 kilometers an hour, which is about 25 miles an hour. George wanted them to look like they were going twice as fast. A lot of the time we’d paint out ground and put in faster moving ground. And faster moving wheels.
I’m very sort of two minds about George’s press coverage, saying you know, there's no visual effects in the film, no CGI. We did 1,695 shots in the cut. I think we worked on 1,797 altogether. I kind of agree with him though, because you don't want to overshadow that stunt work. There are 300 stunts. There's one every 4 shots or something, it's insane. You don’t want people thinking in any way that they are not proper, amazing stunts. There were extraordinary things the guys were doing.
DS: In the final film, the quantity and quality of the stunt work is extraordinary.
TW: Yeah, I'm glad you feel that way. I think it's the sort of sad and dark side of the VFX industry that to promote a film you have to say there isn't any. I think there's some appalling VFX out there, and I hope that producers take note that something needs to be done. Someone needs to be restrained when they ask for another huge fake explosion.
DS: What were the biggest challenges you faced on this film?
TW: The toxic storm. The storm was a huge thing to do as a startup company. We really had to break it down into component parts, try and minimize the load on us and maximize the final visual result, so that we had very, very reusable parts. We had three R&D guys who wrote a thing called Mincer, which is particle iteration software, and that meant we could run fairly complex particle simulations through Houdini. Then through Mincer we would render out something that was intensely full of particles. Just a crazy, crazy number of particles that were still guided by the artist. That gave us a huge advantage. That was the hardest thing, because it was working on a unique film, doing the effects that were required upfront like the toxic cloud and getting a company on its feet.
DS: That's no small task.
TW: No, I don't want to do that again.
DS: Tell me a little bit about the dynamic of working with George Miller.
TW: George was always incredibly relaxed but focused. He was always charming, a delight to be with. He's very, very generous with praise. He was really a ... It was a revelation for me, because I've always worked with…Danny Boyle [director of Slumdog Millionaire] was very similar. He was always very, very pleasant, relaxed, and comfortable with the process and how things were happening. Even if things weren’t right he’d be charming and helpful. George is the same.
On a day-to-day basis we did a review of shots with him maybe once or twice a week, these three or four hour sessions. He'd give his pointers, and during the sessions, the whole thing would be recorded on video, and we were to take those videos back to the artists. George wanted every artist to watch every review. That became impractical so after a while we had to filter that. I would watch it all, and filter what the important points were for the particular artist to watch.
That was a very, very useful thing. The artists would hear my opinion on what I took George to mean and also see George giving that opinion on the video. That was a very, very useful tool. That's one of the reasons we were hitting his notes very quickly.
Then beyond that he had input on everything. He wants to change everything in the shot [laughs]. The way we split screened in the cab, Fury and Max talking to each other, he’d gone in and ended up tanking half of it. That kind of thing. The intensity of each detail, the truth and accuracy in each shot is extreme.
Not having met George before, or worked with him before - Australia is so small, everyone had worked with him before, on Happy Feet or back on Babe and they all knew his MO - it was kind of a revelation for me. “Do we have to do that?...yeah, we have to do that.” When we took a CG supervisor or a compositor to one of our meetings, we’d have to give them quite a briefing and have them watch lots of videos to catchup in case he asked them a question on things that they weren’t even involved with.
DS: That’s funny…
TW: That was quite fun, and he knew that it was an unusual request. It’s not like he thought the whole world ran in this way. He just wanted something to happen that certain way. There were a lot of things where he said he just wanted things to go this way. And that was fine. I keep telling the story about the shot where the guy was getting ripped off the War Rig [in the toxic cloud]. That started with a previs shot from about five years ago, and George liked it. He like the way it looked in the previs, which showed a pretty basic pre-war rig truck, pre-final chase car and twister as just an inverted cone with some textures running around it, with lightning flashing up. The impact of the stream of bodies and this car running up the side of the would-be a twister was absolutely what he saw and wanted in that shot.
But, the camera wasn’t right. The previs was a bastardization of a number of cameras that had gone through a number of iterations. It had gotten to the point where the twister was just a flat plane with the guys rising across it. When we started building it with a real setup, we weren’t getting what he wanted. We had to keep changing the lens to get back to that flattened effect of those guys rising. They’re actually coming towards you. They’re at the edge of the twister. They’re at the edge of a cone – they should appear and come towards us. But [in what George wanted] they don’t – they just go sideways. So that was a problem.
He wanted these guys to have really rag doll kinetics, real world dynamics applied to them. He’d seen rag dolls in crashes and said yes, that’s what I want. He’d seen some really nasty Russian driving videos with car crashes where people were flung from cars. Clearly, they’d been killed. He wanted that look. So we did that. We got realistic simulations on our digi-doubles. When we went back to him, he said he didn’t like them. We brought him what real bodies do when flung or thrown hard or caught in a wind you can’t control. They become lifeless. It’s not like you see in films. We actually saw motorcycle crashes where the driver talked about the crash during the video. They became like a G.I. Joe doll if you threw it hard. They just stretched out and windmilled through the air.
What we realized was that George wanted real film world dynamics. In movies, the movement of people in that situation is completely controlled by stuntmen. What stuntmen have done in the past, when you see them performing in a crash, is a real over-performance. They’ll cycle their legs and swing their arms and turn around because they’re landing on a mat. There is a tremendous amount of over-performance in those stunts. That’s what we’re used to seeing as real motion in film. So we had to go back and add performance into the simulation to make them appear alive, to be struggling against this force that they’re being ripped away by.
DS: What will you bring to your next projects that you'll do a little better, a little smarter, enjoy a little more? What are you taking away from your work on this film?
TW: The way that George and Andy Jackson crafted the VFX. I think they actually got a really good balance trying to get things done more practically. There’s a huge benefit to going practical. I’ve always done that where possible. Shooting physical stuff, effects, performance, shooting as much as you can. In early meetings, Andy apparently was told, “This is going to be done in CG,” and he said, “Well, why? You can make that work practically and it would be a much better fit with George’s requirement.” I would certainly do that on my next show, and try to conceive how we could shoot as much as possible practically. Because you get better results.
For one thing you get the benefit of the guys that operate the cameras. That’s what they do for a living. Shooting it. So you get the better frame, and you get the energy of their performance as well if you actually shoot it. That's underestimated I think in visual effects.
DS: That’s the organic, risky and exciting nature of actually going out and shooting a film.
TW: Yeah, exactly. You’ve got the effect of the 600 people that day that were part of the shooting. Back in the VFX studio you've got five people and though they're having an effect, and you can't expect it to have the same outcome.
Dan Sarto is Publisher and Editor-in-Chief of Animation World Network.
Dan Sarto is Publisher and Editor-in-Chief of Animation World Network.