By Steven Kent
Special to the Tribune
Published November 30, 2002
In “The Fellowship of the Ring,” the first movie in “The Lord of the Rings” trilogy, audiences had a brief glimpse of a shadowy figure climbing a distant wall. In “The Two Towers,” the trilogy’s second installment, that distant shadow becomes a vital presence.
Gollum, a degenerate creature corrupted by the ring, is a tangle of deviousness and emotion. He is shorter than a man, yet has disproportionately long limbs and an elongated neck.
“It was important that Gollum be as real as the other characters,” said Remington Scott, motion capture supervisor for Weta Ltd., the New Zealand studio where much of the special effects work for the trilogy is created. “He had to be real emotionally, psychologically and physically because Frodo [the ring bearer] is supposed to see himself reflected in Gollum. Motion capture is the only way to do that.”
Motion capture is a technology used by video game designers and filmmakers that assigns real movements to computer animated characters. Electronic Arts, for instance, uses motion capture of real professional football players to create movements for the virtual players in its “John Madden NFL” series.
To bring quarterback Kurt Warner to life, game technicians capture his steps, his timing and the way he whips his arms. But bringing Gollum to the screen required something more subtle and exacting: Scott had to capture conversations.
To do this, he worked with actor Andy Serkis, who played both the voice and body of Gollum. “[Director Peter Jackson] hired Andy to come down and do two or three weeks of audio recording for Gollum because they were originally going to animate him,” said Scott. But “Pete liked his physicality enough that they kept him on set during production.”
The technique captures actors, dressed in black lycra suits studded with reflective white balls, performing scenes in the capture zone. The cameras do not record the actors; rather, they record the movements of the light-reflecting balls. This information is translated into computer data, which are then used to create skeletal models.
Graphic artists then design skins, placed over the skeletons, to create animated characters with realistic movements. Gollum, however, with his unique physique, posed an additional challenge.
“Andy’s proportions are very different than Gollum’s,” said Scott. “Over the years–I think he has been alive for 500 years–Gollum has changed. He has gotten long limbs, a long neck, and grown very thin.
“If we took Gollum and mapped him over Andy, he would move very strangely. We had to retarget Andy’s bones into the bone structure of Gollum.”
Scott’s team did this retargeting on the computer, stretching Serkis’ limbs and making him proportionally smaller overall.
For the first film, a 16-camera motion-capture studio was used. But that proved inadequate to create Gollum, so Scott increased it to 24 cameras to create a larger zone.
“When I first came out, we had 16 optical cameras,” said Scott, who joined the filming for the second installment, which opens in theaters Dec. 18. “This gave us 5 meters width by 5 meters length with 4 meters height. It was not a large zone. It was very limiting.”
Even with the extra cameras, creating motion capture stage versions of the real settings used in the movie was demanding. “There was a scene in which some Hobbits walk down a mountain. Gollum crawls down.
“They actually shot the Hobbits on some hill somewhere in New Zealand, I have no idea where. We had to track their positions in space and the plane of the ground. That way we could put Andy onto a plane that we had to create physically [in the studio].”
Still, Scott’s zone was far smaller than a mountain. “We had to build a mountain in four or five pieces,” said Scott. “We would have Andy crawl down one portion of the mountain, roll the next one in, and have him crawl on that one.
“Gollum runs around a lot, so we had to create a stage on wheels so that we could reposition the stage within the zone area. That way we could capture Andy, then blend it together later in edit.”
Real-time direction
And there was one other requirement. Because Gollum’s part was so integral to the success of the movie, Jackson insisted on directing Serkis’ motion-capture sessions himself. He could not, however, properly direct scenes in which Gollum speaks with Frodo and others simply by watching Serkis alone in a capture zone.
“It was essential for Peter to see Andy’s performance retargeted as Gollum in real time,” Scott said. To do this, Jackson watched a monitor instead of Serkis and the zone.
Movements synced with actors
“Peter would look at a visualization monitor the entire time,” said Scott. “He would be sitting right next to me, and the stage is right in front of him; but
he’s not looking at Andy, he’s looking at a monitor.
“What he saw was a fully rendered and retargeted Gollum in real time, but that wasn’t enough. He had to see Gollum grounded in the shots, so he could see him in relation to all the other actors and what they were doing.”
To do this, the computer superimposed Serkis’ data over still plates of scenes from the movie. The image of Gollum was placed in the actual scenes, and Jackson was able to see where he would appear and how his movements synced with the placement of the other actors.
Steven Kent is a freelance writer based in Issaquah, Wash.