by Karen Moltenbrey
A new type of "immersive" effect used in the opening title sequence for the latest James Bond movie, The World Is Not Enough, is anything but crude. Using a special "liquid" environment created by Smoke & Mirrors, an effects boutique in London, the animators digitally created flowing pools of what appears to be crude oil and, through a unique digital processing technique, transformed the liquid into lush oil-covered ladies.
"A Bond title sequence-which follows the first roll of exciting action-traditionally involves beautiful women and special effects," explains Daniel Kleinman, director of the sequence. The three-minute film footage for The World Is Not Enough, the 19th Bond film, is no different. It contains the trademark Bond psychedelic-style imagery of guns and women, but in this case, the composites of their various shapes and movements evolve into oily liquids. "The major theme in The World Is Not Enough is oil," he adds, "so the sequence entails submerging the women in and dripping onto them what appears to be crude oil."
The unpleasant prospect of enveloping models in thick, sticky oil to achieve the effects was eliminated through the first use of a cutting-edge digital imaging technique created by Sean Broughton, digital effects supervisor at Smoke & Mirrors. This process entails simulating the movement of liquid by using a dry environment and turning figures shot against a green screen into liquid people.
 |
| Inferno artist Sean Broughton devised special digital imaging techniques, which were used in the opening credits of the latest James Bond movie, to create the illusion of women submerged in crude oil. |
Broughton had just completed a few months of R&D to achieve these effects of when he learned of Kleinman's vision for the movie scene. "The initial concept for [Kleinman's] design married up well with the technique we had been working on that turned real people into liquid without requiring us to shoot a lot of actual elements," the artist says.
While the methodology of "capturing" the movement of the liquid and of the women varied somewhat, the special image-processing technique created using Discreet's (Montreal) Inferno-which produced strangely shadowed shapes and further gave them their oily look-was the same. And likewise, so were the results: making uniquely textured, flat 2D images appear three-dimensional.
 |
| Many of the oil elements in the sequence were computer-generated, created either through the digitizing technique in Inferno or in Maya. However, in some instances, actual liquid, such as olive oil, was used. |
To create the oil and its movements, Broughton devised a unique "dry-liquid" environment. This consisted of a large tank filled with tiny (approximately 3mm in diameter) white polystyrene balls into which the actual women were immersed. The theory behind this procedure, he says, was that whenever the women moved around in the polystyrene, the tiny balls would react and move in a manner similar to that of a thick liquid, such as oil.
Filming from atop the tank, Brough ton captured the rippling effect of the balls, which were lit from the sides of the tank. As the balls moved and passed through the light, they produced shadows. After the film was scanned, these shadows enabled the artists to automatically create luminescence and bump maps in Inferno, using illumination levels that varied depending on where the shadows from the balls were cast. Since the balls were moving, the maps changed from frame to frame. The group then fed the bump maps directly into GenArts' (Cambridge, MA) Sapphire software, a set of Inferno plug-ins, running on SGI (Mountain View, CA) Onyx workstations.
Next, the artists generated the desired textures (such as petroleum and var ious reflections) and applied them to the bump maps within the Inferno Batch processing option using Sapphire and various plug-ins from 5D (Surrey, England).
"The lighting and shadows provided the necessary depth for generating the unique texture, which gave the imagery an abstract quality," says Broughton.
 |
| To capture the fluid, graceful motion of the women, the artists filmed professional gymnasts and dancers against a green screen. From their shapes and shadows, the artists created bump maps in Inferno, onto which the petroleum-like textures were then appl |
According to Broughton, Inferno's Batch processing enables 2D artists to work in a way similar to 3D artists. In stead of "wrapping" the textures around the images, the artists generated them directly onto the 2D images. As a result, the textures took on the moving, changing qualities of the filmed images. "This is what makes this process new and different," says Judy Roberts, a Smoke & Mirrors artist.
This entire digital process works especially well when emulating thick, slow-moving liquids, such as oil, Roberts notes. Furthermore, "The movement of these tiny balls when they rolled off the women mimicked remarkably well the characteristics of liquid as it is poured over a person," says Brough ton.
This complicated process enabled the artists to achieve the "slick effects" of the oil and the oil's movements in the shots where the women appear to plunge into the sticky liquid.
 |
| The specially devised "dry-liquid" environment was particularly effective for emulating thick liquid, such as crude oil, the focus of the film. |
For creating the lush liquid ladies who sashay and gently bound across the set in the sequence, Broughton filmed brightly lit gymnasts and dancers against a green screen as they performed graceful leaps and other movements on a trampoline. The 2D film images of the performers were then scanned and placed against a black background so that the Sapphire program could determine the luminescence values of their body surfaces-creating gray scale-like contour models similar to those used to replicate the pools of oil. The artists then created petroleum-like textures and, using Sapphire and other plug-ins, applied them to the bump maps of the women.
The group chose this method of creating the movements of the liquid-like ladies for several reasons. "Although it was tempting to emulate in oil the seductive 3D rendering of the digital 'circuit girl' featured in the previous Bond film, Tomorrow Never Dies, the director this time wanted the grace and movement that only actual dancers or gymnasts could provide," says Broughton.
For The World is Not Enough sequence, Kleinman preferred using the new 2D "liquidizing" technique instead of a 3D computer-generated model whose movements were motion-captured. This is because Brough ton's approach uses the actual motion that is filmed, rather than a computerized interpretation of the movement, as is the case with motion capture.
 |
| One of the key advantages to using Broughton's technique was that it enabled the artists to work in Inferno, from start to finish, for the entire project. This was especially important because many of the shots were designed by the director and artists on |
The end result, says Roberts, is more fluid and organic motion. Be sides, using motion capture to animate all the women appearing in the cuts would have been extremely time consuming and difficult, she notes. And Broughton's meth od offers the ad vantage of en abling the art ists to perform all the necessary work on just one system, Inferno, from start to finish.
In essence, explains Broughton, this entire digital process captured the gymnasts' shapes and enabled the artists to turn them into oil by compositing multiple layers of the images into the scene. "And the beauty of our technique was that the gymnasts could submerge themselves again and again in a clean environment, without any discomfort, allowing for multiple takes [during filming] and a faster turnaround time," he adds.
 |
| This solar system scene-with 3D planets, actual fire elements, and transforming women-was one of the most complex in the sequence, taking days to render. |
To augment the liquid ladies, the artists used Alias|Wavefront's (Toronto) Maya to create various 3D elements for the sequence-such as huge oil splashes that reverberate and assume the shape of the women, and a digital bullet that plunges through one liquid lady, turning her back into a pool of oil. The group also infused footage of actual water and various types of oil into the scenes.
In one of the more difficult shots within the opening-credit sequence, the artists added a solar system comprising textured spheres and moons, the most complex of which was a blazing sun-like globe, complete with solar flares that erupted in time to the sound track. To create most of the planets, the artists used Maya; for the texturing, they added various oily textures using the Inferno software. For the sun, the group used clips of actual fire for the flares, but the requisite scale and level of detail prevented Brough ton from simply attaching selected bursts of flames to a burning computer-generated 3D ball. "I don't think it would have been convincing enough," he maintains.
Instead, Broughton, who performed the majority of the graphics work for the sequence, generated a series of textured layers in Inferno (similar to the process of generating the textures for the women) to represent the sun's crust and magma-like surface, with constantly evolving chan nels that would emit streams of light. The final output resulted in one of the most complex shots of the entire sequence, he notes, requiring three days to render.
In all, the group spent two weeks visualizing and designing the shots, followed by four weeks of film-resolution compositing. Because they had so little time to plan the shots, Broughton and Kleinman instead designed many of them on the spot, which is why they chose the flexibility provided by Inferno for creating the digital effects, rather than using typical 3D modeling tools.
Although only three minutes in length, the sequence contains more than 2.2tb of scanned elements (32,000 frames), ranging from green-screen shots of the women to simpler digital oil drips. One scene-involving a woman lying on her side fashioned to represent the horizon line and dozens of flying women that metamorphose into flowing oil-alone required several days to design and execute, and four days to render.
Kleinman notes that with a small budget, which he declined to specify, the group was able to accomplish a breathtaking three-minute sequence. "The reason everyone put so much time and effort into this project, far beyond the monetary reward, is because it's James Bond," he says. "It's exciting to be part of the phenomenon."
 |
| Artists added a 3D model of a bullet, created in Maya, that plunges into one of the liquid ladies, turning her back into oil. |
Broughton agrees, emphasizing that the project was a large undertaking, especially to achieve just three minutes of footage. "But I think everyone was blown away," he says. "I don't think anyone has seen anything quite like this before."
Karen Moltenbrey is an associate editor for Computer Graphics World.