Including the time an Evian spray bottle was used for outer-space particles.
Ron Howard’s Apollo 13 turns 25 this week. The VFX Oscar-nominated film was one of Digital Domain’s landmark early calling cards, full of model and miniature effects and a wealth of nascent CG imagery and digital compositing created by the then-new studio.
Much has been written about the impressive effects work, especially for the dramatic Cape Canaveral launch sequence and scenes set in space. So on the quarter-century anniversary I thought I’d ask the film’s visual effects supervisor Rob Legato about some of the perhaps lesser-known aspects of the VFX production.
Here, Legato shares how he came up with a ‘pan-and-tile’ method on Apollo 13, how Ron Howard suggested he shoot the launch sequence like Martin Scorsese would shoot it, how some rocket effects were re-shot from computer monitors, how an Evian water spray battle came in handy for particle effects, and how the splashdown scene employed some parachutes that may not have opened at all.
1. Panoramas for backgrounds: imagining a pan-and-tile approach
Rob Legato: I had come up from production and commercials. When you did a location scout, you would get these ‘One Hour Photos’ done where you’d take a bunch of snapshots going all the way over a landscape or a location and tape them together to make up one panorama. It’s just like a panorama that you can do now on an iPhone, but back then the One Hour Photo method was the only way you could do it quickly.
I was doing some work on Interview with the Vampire and I figured I could use Photoshop to do a better job of making this kind of panoramic tile so that I didn’t have to physically tape One Hour photographs together. I could scan them in and tile them together.
Back then, we had a computer called an Abekas, it was like a digital frame store. What we could do is have these frames, these tiles, chopped up in multiple frames. And then you had a little track ball and with that you could determine what speed you would go. So it was basically like panning, ie. going from frame to frame to frame. Because I could control it going back and forth, I thought, well, this is motion control. At that time, too, we could start tracking things on the Flame, and then apply those numbers to this big tile of photographs.
I thought this was also something we could use on Apollo 13. I had to go to Cape Canaveral to shoot scenes, but we hadn’t really determined what all the shots we’d be doing were and what the camera angles would be. Normally to shoot plates, we’d have to know that we’d be shooting with say a 20mm lens, tilted up 20 degrees or whatever it needed to be. But, if I shot a ‘pan-and-tile’ by taking all these photographs – which is today so commonplace – if I did that, I could then pick and choose, when I decided what the framing of the shot was on stage (where we shot miniatures), what the background would be.
My suggestion was that we go to Cape Canaveral where I’d shoot with a film camera to do pans and make these high resolution plates. I called it ‘hosing down the background’. A tile might have 30 frames at best, and then we would transfer 30 frames on film. But there was some resistance against the idea because they thought it would cost a lot of money. I thought it was the opposite. Instead of me shooting 30 plates, I’m shooting 30 frames, putting them together and now I have any number of shots I can do. If I want to change the shot or make it better, I can do whatever I want, and I’ve only transferred 30 frames of material. I figured it saved money. It also meant you didn’t need to bring motion control out to Cape Canaveral.
So, I decided to do it anyway. I knew it was going to work and save me money and help me make a better movie. So that’s happened, and we could use these panoramas for backgrounds to do pan or tilt shots. Then, ultimately, a couple of people who were working on Nuke at Digital Domain back then wrote a pan-and-tile interface and workflow, which is still in Nuke now.
2. ‘Shoot it like Martin Scorsese would shoot it’
The best piece of advice I ever got was from Ron Howard when we were coming up with the launch sequence. I did not want to use stock footage for it, which was the original plan. I wanted to shoot specifically to match into the rest of the movie. It always used to bug me on movies where you watched the normal film and then you’d suddenly see this grainy stock footage – it pulled me out of the movie. Even if it’s an authentic shot, it just looks like you cut to stock footage that was filmed at some other time.
Ron Howard said, ‘Well, if you’re going to do that, can you shoot it like Martin Scorsese would shoot it?’ I wasn’t sure exactly what he meant, but I looked at all of Marty’s movies. I had an Immix Video Cube – a non-linear editing system – at the time so I could analyze Scorsese footage. One of his films, The Color of Money, was really interesting, especially the pool-playing scenes. They were really cool and explosive; it was hard to put your finger on why that was but there was something that just made them particularly interesting.
I went in frame by frame and I realized what they were doing was double cutting it in a three or four frame ‘mis-edit’, which would double up the action, but not enough for your eye to quite see that it doubled up the action. If it went one or two frames more, you would see the repeat. But the way it was edited there was this fantastic snap to it.
I also noticed that Marty did everything in rhythm. There was a rhythm to the shots. It wasn’t arbitrary. You felt the beat. And you were almost anticipating the next thing. It changed the way you watched the movie. It was almost like music, you could almost predict when the next beat was going to happen.
For Apollo 13, we shot the underside of the engines going off in miniature with pyrotechnics. We’d have one angle, then cut to another angle and I would repeat it two or three frames off. So instead of the next piece of footage starting in continuity, I’d back it in just a few frames. It was really delicate. If I went two more frames, you would have seen what I was doing.
I cut the launch sequence myself, because it was very specific to this music and this editorial rhythm. I gave it to the production. What was really lovely is that the editors left my edit intact and just added in the live-action parts. Then James Horner sensed the music that I was cutting into. I would play heroic music in dailies and then I’d go and play it against the the shot and was influenced by it. That would be magically in sync. James would find the flow and the crescendo of it and he made his music and I don’t think it was even conscious. He was creating the beat of it from the edit and the edit was created from music originally. When I saw it the first time, I just got goosebumps. I’ve never had an experience yet where something was even better than I’d hoped.
3. Extra bloom, from a monitor
One of my favorite shots, partly because of how cheeseball it is, was where we used a $25 Revell model of the Saturn V rocket. It was one of the long lens shots where it’s traveling up as if it’s being shot with a 500 mm lens and the camera’s shaking a little bit. It was spouting out all the fire out of the back of the rocket.
Using stock footage reference, we replicated all the nuance of that, but it didn’t look right, it looked like a video game. I had our artist convert the shot into digital frames and we set up a high resolution monitor and then we photographed it frame-by-frame off the monitor with a Black Pro-Mist. It was shot maybe five stops over-exposed. That got rid of the video game look.
Everybody thought that was crazy until they saw the result. It gave a more realistic bloom of the light. Before it looked like cartoon artwork; there was no sense of light or shadow because everything was the same level. Now it had the proper blown-out exposure look.
4. Particles, from a bottle
For some of the shots we needed to show the effect of these little thrusters which would spray something out of them to push the module one way or another to ease into what the course would be. Then there was also the part where the astronauts take a leak and open up a hatch and it would go out in space and crystallize.
We spent about a month doing something in CG, but it didn’t look right. There really weren’t enough particles. Back then, too, it really taxed the system to do these simulations.
Luckily we had cameras there in the back room at DD, and we had lights. I sent somebody out to go get an Evian spray bottle. I put it on a C-stand and we built a little thing to hit it without shaking the stand. We backlit it and we shot it. It couldn’t have been more than a 15 minute shoot. We transferred the footage, and what we had were tons of particles coming out and that’s what’s in the movie for some of the shots.
Originally they wanted the splashdown with the capsule and the parachute to be done in CG. I said, ‘No, we’re not doing that.’ Instead, I said, ‘We’re going to throw it out of a helicopter and photograph it. It’s going to float in the air and then hit the water and look right.’ I knew I could shoot all day with two helicopters and a bunch of ground cameras and have 20 hours of footage if we wanted.
They showed me a shot from another movie that had a CG parachute shot and it did not look good, I thought. I said, ‘Are you showing me to convince me that I’m correct?’ I was pretty forceful about it; I wanted to make sure we were shooting realistic things.
We got guys who make parachutes and who are real skydivers, and I said to the main guy, ‘So, this will work, right?’ And he said, ‘Parachutes want to open.’ I said, ‘Sold.’
It was still a big engineering project to build a mechanism to release it and make sure it didn’t go up into the rotor blade. We had some really beautiful helicopter shots looking straight down as it went away, and then the hero shot of it hitting the water was long lenses and it really looked like it was real footage.
Anyway, this guy who told me about the parachutes was almost in tears when he saw the footage and I thought maybe he was being emotional and patriotic, but he said, ‘The truth is I didn’t know those chutes would actually work. There was probably a one in 10 chance they would actually open.’ I said, ‘Well, you didn’t tell me that!’
For more information on Digital Domain’s impressive VFX for Apollo 13, check out Cinefex #63. Berton Pierce also has two video pieces on the miniatures work. Rob Legato’s TED Talk, ‘The art of creating awe’, features discussion of the film, too.