By Bill Desowitz
In the summer of 2001, Titanic Director James Cameron organized a new expedition to the site of the legendary 1912 shipwreck to shoot an unprecedented 3D documentary titled Ghosts of the Abyss, which opened April 11 through Walt Disney Pictures in IMAX and other specialty theaters. Using new advanced 3D technology, Cameron and his crew of scientists, historians, and actor friend Bill Paxton (Titanic) traveled deep below the surface of the ocean and throughout the ship to discover images and artifacts hidden from previous explorers, including exquisite leaded glass windows in the dining room and an upright drinking glass in a stateroom. This technology included the large-format Reality Camera System, which Cameron invented in collaboration with Sony and cinematographer Vince Pace (The Abyss, Titanic), repackaging Sony’s CineAlta imaging capabilities into a custom 3D system; a special underwater camera housing invented by Cameron’s brother, Mike Cameron; and two revolutionary Remotely Operated Vehicles (ROVs), also invented by Mike and his Dark Matter team. Made in collaboration with Pace, Producer and Visual Effects Supervisor Chuck Comisky (Terminator 2 3D) and Editors Ed Marsh (The Abyss), Sven Pape, and John Refoua, Ghosts of the Abyss presented Cameron with many challenges. Here’s what he told Below the Line.
Cameron: This is an interesting film because I saw my role primarily as expedition leader and secondarily as director. So I turned to [Cinematographer] Vince Pace and told him whatever happens, just shoot it. Don’t screw up. I’m going to be over here working on the next dive. So they said, “Aren’t you supposed to be directing this?” And I said, “No, just shoot something interesting. Like if they’re putting the sub in the water.” Actually I was usually in the sub, so all those shots you see of the sub in the water—those were all done by the crew after I left the ship. So directing maybe is not an accurate term…
Below the Line: But you developed the camera system…
Cameron: Yeah, we obviously developed the camera system. Vince and I did that together. He was on The Abyss back in ’88. We have done a lot of work together. We had a pretty good shorthand. The trick on this was figuring out how to deploy the camera system, trying to anticipate what would happen next. We couldn’t stage anything; we couldn’t do anything twice. If you missed it, you missed it. And the camera system had never been used before. Vince is a tough guy, but I saw him come that close [to losing it].
BTL: How long did it take you to develop the camera system?
Cameron: It was a couple of years of working with Sony. I don’t want to say that was the easy part, but that was the calm part. But there was this curve of greater and greater difficulty as we approached the shoot, because once you commit to a project, you’ve gotta go. Stuff was flying at us the last second, and Vince was under a lot of pressure to get the optics ready for the underwater camera—because we were going out there to not just shoot in 3D but to shoot underwater in 3D. So the whole camera system had to be integrated into a housing that could withstand the pressure. The housing had to be built for the camera. And my brother Mike designed the housing but he didn’t do the optical integration of the 3D system for the housing. And the way it works is—it sounds a little strange at first—we had to build the housing as if we were going to get inside it. It had to be man-rated, which sounds dumb because you’re obviously not going to get inside the housing. But the implodable volume was such that if it had failed, catastrophically, it would’ve sympathetically failed the submersible and killed us. The man-rated specification for our view port became what we had to build to. Now they didn’t build ports for stereoscopic viewing. The view port is a dome with a conical seat—and it was built to what they call the PDHO spec, which is a Navy spec for building spherical concentric dome ports. Now we can put the camera behind the port, we can put it in the water, but the image you get is crap. One lens is not looking through the optical center of the port but to the right of it, the other lens is looking to the left of the optical center, and they’re creating two separate images with different distortions. We had to correct this by adapting the optical distortion back to zero for each of the two lenses.
BTL: In other words, for the first time you get more natural stereoscopic viewing.
Cameron: Yes, because the lenses are separated by the same distance as between our eyes: 70 millimeters. And Vince had to figure all that out in the last second. And it was insane because he just turned his shop into an optical test bench and built a pool with a test port…At the same time he hired a very high-end optical design company and Vince and I would call up each other every day and say try this, try that. So Vince was literally grinding lenses to test and on the day that he cracked it, with 3D live feed and everything working, we got the results from the computer-assisted optical design company that said you can’t do it.
BTL: Now what about actually shooting in 3D underwater?
Cameron: …When we got there, Vince and his guys had to figure out how to use this HD camera system—shoot in stereo, which means there’s two of everything: two cameras, two electronic processing units which we call “bodies,” two record decks—the stream has to record in sync and how do you do that all over the ship? It was a huge production problem. I mean what do you do when on the first day you didn’t shoot anything, or the second, or the third? Meanwhile, we’re already on the expedition, but we’re not shooting the expedition because we haven’t figured it out yet. So there was a tremendous amount of pressure on the camera team to solve all this stuff.
BTL: In terms of filmmaking what was the big 3D breakthrough here?
Cameron: The breakthrough [was] that we’re working in digital and in HD, so you’ve got all the speed and ease of HD, and you can electronically sync very easily. And the way we wound up working with it was we’d take the HD tapes and we’d load them into a two-channel server, and we’d run them back out of the server, so there were never any sync issues. And you can post manipulate the HD images so much easier than you can film images. And come right back into projection immediately. In other words, you take the digital files, you go off, you perform a manipulation, you bring them back, you stick them on the server, and you don’t like it, you do it again. You reduce a 36-hour cycle down into an hour. I don’t want to go back to working on film anymore, certainly for stereo use, and maybe even for general use. You’re painting and moving in real-time in a projection suite environment that looks indistinguishable from a final projected image in a theater.
BTL: I’ve heard that you want to do your next feature in 3D.
Cameron: Vince and I are working on that right now. We’re in meetings figuring out what it’s going to take—how many cameras we’re going to need. We’re doing it over the next year-and-a-half. The time for theory is over. Do you remember John F. Kennedy’s speech about the space program? He said we do these things not because they’re easy but because they’re hard. Vince is now working on Spy Kids 3, part of which is in 3D. But you know that’s just one part of it—building the big robot vehicle that was used to light the wreck—that’s a whole other story. How do you light an object that big? Even having that other vehicle with 10/1200-watt HMIs wasn’t enough to light up Titanic. It’s like trying to light up a football stadium at night with a flashlight underwater. There’s no ambient light whatsoever. So we had to build this vehicle that was essentially a big ROV, figure out how to have ballast for these big HMI lights and how to control them and fly them from two-and-a-half miles up from the surface. Now the subs are diving untethered, so they’re diving in a whole other way. Then to explore the inside—that was the hardest part.
BTL: The ROVs…
Cameron: Jake and Elwood [from The Blues Brothers]. My brother’s the engineer of the family and he and really three other guys built them over a period of three years. A lot of headaches, only because they’re revolutionary vehicles based on principles diametrically opposed to ROVs. These were designed to work differently from all previous generations of ROVs. They aren’t made to go inside the wreck and explore all the different rooms. The whole idea of this ROV is that it’s not trying to drag its tether and feed the tether in behind it; it’s paying the tether out from the vehicle. The tether comes around a corner, it kind of hangs up—no problem. It just keeps going. It gets hung up somewhere else, it keeps going.
BTL: Like a fishing rod.
Cameron: It’s like running around with a fishing reel in your hand through a building. And we put almost 3,000 feet of tether in that vehicle. And meanwhile through that 3,000 feet whether CHK it’s spooled out or whether it’s still wound up inside that vehicle—that’s an open fiber-optic light. So those light pulses are shooting around through those 3,000 feet at 186,000 miles-per-second, and that’s how you control the vehicle. Pretty cool idea and that’s how it worked. But there were no off-the-shelf components. We had to build the machine from scratch. It was basically like saying we’re going to build a Ferrari but we’re not going to start with a frame and an engine—we’re going to make the pistons and make the fan and make the radiator; then we’ll have a Ferrari that’s really expensive, but don’t break it.