After scouting and rejecting the existing studios that offered water-tank facilities (a process Cameron has previously gone though on The Abyss), the filmmakers decided that they would have to custom-build their own facility to meet the project's various needs. A 40-acre portion of oceanside property in Rosarito Beach, Mexico was acquired, and plans were drawn up for what would become Fox Studios Baja.
Constructed of steel pipe scaffolding and wood, the 1,300,000-pound Titanic set was built in the six-acre Tank One just a few yards from the Pacific Ocean, which offered a natural horizon line when the ship was seen from the starboard side. To accommodate multiple sinkings, the faux ocean liner was built on an elaborate gimbal system that could repeatedly submerge and raise the set (approximately 40 times over the last six months of shooting) into a 40'-deep well at one end of tank. At full tilt, a 250' stern section was hoisted into the air. Most of the 15.4 million gallon sea-water pool was only about 4' deep.
Great lengths were taken to make the ship set and props feel authentic. Production designer Peter Lamont (Aliens, True Lies) began by obtaining the actual Titanic blueprints from the original shipbuilders, Harland & Wolff, who had constructed the liner and her nearly identical sister ship, the Olympic, in Belfast, Ireland. In addition, Lamont discovered that the manufacturer of the original carpeting for the ship was still in business; he had the firm re-create the exact patterns and colors used throughout the ship. Thousands of props, all based on original designs, were built were for the film, including china plates, silverware, a stunning crystal chandelier, and a brass-detailed Marconi system identical to the model that had transmitted the Titanic's last desperate S.O.S. signals.
At the south end of Tank One, a duplicate stern section of the Titanic set was built on a gimbal capable of inverting the structure to a 90-degree angle replicating the position of the ship just before it fully slipped below the surface.
Additional interior portions of the vessel (such as the grand staircase, ballroom and dining room) were built in Tank Two, a 30'-deep, 4.5-million-gallon pool fully enclosed in a 32,000-square-foot soundstage. The sets were constructed upon a hydraulic platform at the bottom of the tank that could lower and raise them for repeated sinkings. Dry sets were constructed in two adjacent, more traditional stages.
Baja Studios was built over a 100-day period, and as Carpenter recalls, "The facility was still under construction when we arrived. Even as we were shooting you could hear the noise of buildings going up."
The need for a reliable power supply was a major concern, as it had been on such films as Dune and Total Recall, both of which were shot primarily at Churubusco Studios in Mexico City. "My predecessor had come up with some distribution designs for the studio, but no one had really tested the power," says Buckley. "The first thing we did was to put a testing device on the current coming off the Mexican grid. We found that it had fluctuations of 40 volts and 30 cycles, which is completely unacceptable for HMI and fluorescent lighting. We could have gotten around the power problem if we'd shot exclusively with tungsten sources, because we'd have been immediately aware of the output fluctuations. But you can't shoot with just tungsten anymore; it simply isn't practical. Tungsten fixtures don't put out enough light especially these days, when everybody wants soft light and deep stops. We ended up bringing in generators from the States. To give you an idea of the power we were consuming, the lighting for the ship riser alone took 40,000 amps."
Importing equipment quickly became a major headache that would continually trouble the production. "Logistically, we ran into problems," Carpenter concedes. "Due to the delays in customs, we often had to decide what we needed weeks in advance. We couldn't just call up a rental house and have something there the next day." With a knowing smile, the cinematographer adds, "That isn't to say that a piece of gear sometimes wasn't just thrown in the back of a car so that it could mysteriously 'show up' on the set!"
The most effective solution to this problem was to keep an arsenal of production equipment on hand. As Carpenter notes, "By the time we got certain pieces of equipment packed up and back across the border, it would be just about time for it to come back. There were days when things sat around without being used, but there were also times when we had the two Akelas, a Wescam, and the Technocrane all in action on the same day."
About half of the Baja Studios crew was from Mexico. "I respected them, and we got along tremendously," Buckley says. "They knew what they were doing, but in the beginning it was hard because some of the [American] crew seemed elitist. Part of that was due to the language barrier. The Mexicans would bring the equipment and the L.A. guys would start taking it out of their hands. I said, 'That stops now. Whoever brings the light in gets to set it.' After the first month, it worked great. I don't see movies just as making a movie. There's a spiritual side to it all, a feeling of brotherhood, and and I try to make sure that that part of the project gets worked out also."
As one would expect, safety was a primary concern for the Titanic electrical department especially since the cast and crew would be perpetually wet during some nine months of shooting in and around sea water. Buckley ran everything on AC current through ground fault interrupters (GFIs), which would immediately trip the power if a circuit was compromised. Individual 20-amp GFIs were assigned to certain fixtures, while custom-built 1,200-amp versions were used for series of lights or large fixtures such as 10Ks and 20Ks. However, the 1,200-amp GFIs played havoc with the gaffer's dimmer system, as they "somehow created an electrical loop inside the dimmers, causing everything to shut down and taking down a whole section of lights. You could set the lights on individual GFIs, but we couldn't dim them all the way out, or they'd trip. It was a compromise, but we went with it for safety reasons.
"Meanwhile, I didn't want to burn any HMI or fluorescent fixtures in the water at all," Buckley says. "We did one GFI test by throwing a Kino Flo in the water, and it continued to burn. With the current going through the ballast, the GFI doesn't necessarily read the voltage fluctuation and won't shut down. So unless we had waterproof HMI Pars which are only 1,200 watts we used tungstens. At the end of the shoot, when we were sinking the front section of the ship, we had every underwater Par available, from both HydroFlex in L.A. and Pace Technology up near San Francisco."
Additionally, the Titanic set's gimbal could pitch the structure to a 12-degree angle. "That doesn't sound like a lot of tilt, but it is when you're actually experiencing it," Buckley attests. "Everything had be to chained or tied down, which slowed us down a lot; right after we'd set everything, Jim might find a different shot he preferred, and we'd have to take everything apart."
Buckley details, "We also built 70 miles of 4/O feeder lines so we knew they'd be clean and that whatever we put in water would be perfect with no cuts. We had 8'-wide carpets of 4/O leading out of the ship and along the edges of Tank One. People coming out of the water, dripping wet, would come out of the tank and walk on these lines. There were two instances where people got shocked, but because of the GFI system all they got was voltage. There was no amperage, which is what will kill you. Nobody was really hurt, although the members of the electrical crew got shocked more than a couple of times!"
As a testament to the complexity of Titanic's lighting scheme, Buckley offers a collection of over a dozen laminated rigging diagrams (see foldout section) detailing each and every fixture that was used on the ship set in Tank One; the corridors and dining room set built in Tank Two; the interior sets depicting the forward hold and boiler room; and other areas as well. The numbered fixtures ran into the hundreds. While simply keeping track of all these lamps was difficult, "the hard part was that we were trying to cheat the laws of physics because we were trying to primarily light the picture with practicals," Buckley says. "Jim and Russ never wanted to shoot under a T2.8. That took us right out of the realm of reality, because we needed such high-wattage globes, and the distances were so great that we couldn't do it. But we could get close. Most every lamp we had on this show was either a 300-, 500- or 1000-watt household bulb. They were all encased in watertight housings, so you can imagine the heat that was generated."
The fact that these hot-burning practicals would be repeatedly dunked in 60°F sea water was a major stumbling block, as the rapid temperature shifts would burst bulbs, crack housings and cause seals to fail. Dozens of lamps were lost on some takes, requiring them to be arduously replaced before shooting could resume. Buckley remembers the frustration of the situation: "Jim often said, 'I've been down at the bottom of the ocean with the Titanic we built lights to go down there.' But the reality is that a sub goes underwater and stays there for prolonged periods. You don't rapidly bring it up and down."
Despite testing and improvements, the dilemma persisted. "We still had to replace practicals between takes," Carpenter says, "but we developed a replacement system that was reasonably successful."
Creating an overall lighting plan for the ship set was a monumental task. For Carpenter, a potentially problematic Titanic fact was that the moon had been in total phase on the evening of April 15, 1912, dispensing with any cinematographer's primary nighttime lighting motivation. Starlight would have to do. The system used to create this soft overall illumination was in constant evolution during the Rosarito Beach shoot. Recalls Carpenter, "When I first came on the picture, we had an immediate meeting over at Digital Domain, where there was as big model of the ship as it was situated in Tank One, with Mylar used to replicate the effect of the lights reflecting on the water. Looking it over, John Buckley, [key grip] Lloyd Moriarity, Jim and I figured that we could primarily use a Musco and a Night Sun from the eastern edge of the tank, as well as lighting balloons on the ship itself."
Fortuitously enough, Buckley, who'd previously used balloons while working with Ron Garcia, ASC, had found some new floating fixtures that seemed right for the job. The gaffer explains, "I'd gone to the ShowBiz Expo in June of '96 and seen a new 16K HMI balloon made by the French company Solaire and distributed by Industrial Wholesale Electrics. When I got down to Rosarito, I ordered two of them."
Caught in a tangle of red tape at the U.S./Mexico border, the first two Solaire balloons never arrived. "There was some question about what they were going to be used for," Buckley recalls, "but we reordered and ended up with four of them. They're great lights, and Russell and Jim loved them, but they have to be used in a very specific way."
In other words, the balloons could only be utilized in relatively windless weather which soon became a rarity at the seaside Baja Studios as the shoot continued through the winter months. "A lot of thought had been put into where the studio would be built," Carpenter explains. "The tank obviously had to be near the ocean to sell the desired effect, but the surrounding geography was important too. We needed a clear horizon for about 180 degrees, with no mountains or structures in the way. But what wasn't considered were the winds that could come up. By the time we realized how strong they could get up to about 60 m.p.h. a couple of the balloons had been blown out to sea."
"The wind was pretty intense," Buckley confirms. "After we had ripped a few of the balloons, we began changing our rigging design. Normally, the balloons are just tethered to the ground, but I wanted another tether on the top so it would be held at two points and couldn't really go anywhere. We ended up using a construction crane for the upper tether, and that system worked really well once we got it finished."
After some refining of the attachment points, Buckley and his crew ordered more balloons, cannibalized parts from the damaged fixtures, and prepared to deploy four of the floating lights. "By that point, we'd done about $120,000 in damage," the gaffer admits. "Everybody got paranoid about the wind coming up, so we didn't use the balloons as often which was too bad, because we'd finally worked the kinks out of them."
[ continued on page 4 ]