Navigating 'The Golden Compass' -- Part 1

Since the publication of the first volume in 1995, Philip Pullman's His Dark Materials has become one of the most beloved and celebrated book series of the past decade. Although the trilogy is not as illustrious as Lord of the Rings or Chronicles of Narnia, New Line Studios certainly hopes that The Golden Compass, the big screen adaptation of the initial novel (opening Dec. 7), will meet the same enthusiastic response from audiences worldwide.

To supervise the all-important visual effects effort, director Chris Weitz enlisted Michael Fink, (Batman Returns, X-Men and Constantine) as senior visual effects supervisor. "We started preparing the movie in January 2006, and began shooting in late September," Fink says. "Principal photography then wrapped in February 2007. We later regrouped in April and shot more until July. Post went from then until November 16, 2007."

For Fink and Visual Effects Producer Susan MacLeod, there were, of course, many technical challenges, most notably figuring out how to manage procedural fur grooming for interaction between animals and humans, and creating a complete 3D-animated sequence (the bear fight) in the middle of a live-action movie. "Chris Weitz was kind and trusting enough to let us do our job with only gentle, but quite specific, guidance from him," Fink notes. "He really focused on those things most important to him and to the film, and let us concentrate on completing visual effects shots in what was a very tight schedule. Without this approach, I doubt we would have finished in time for the release of the movie, given how little time we had for going back and forth to figure out a shot."

The shots were spread among nine different vendors:

Rhythm & Hues & Rhythm & Hues India 755Cinesite -- 475Framestore CFC -- 322Digital Domain -- 150Rainmaker UK -- 114Peerless Camera -- 74Tippett Studio -- 28Digital Back Lot -- 13Matte World Digital -- 7

The final cut includes about 1,100-1,200 shots, although total shot count is closer to 1,500 as there were numerous omits and changes (more on that later).

Meet Your Inner Daemon

The story takes place in an alternate world, very similar to ours in many ways, in which humans all have a personal daemon: that is, the physical representation -- in animal form -- of their souls. Featured in more than 500 shots, the daemons were created via 3D animation at Rhythm & Hues (in collaboration with Rhythm & Hues India, which contributed animation lighting and composting on about a third of the shots). The team was led by VFX Supervisor Bill Westenhofer, Co-VFX Supervisor Raymond Chen and VFX Producer Gary Nolin. This represents the studio's most ambitious project in its 20-year history. "With 490 shots in the film and over 250 more that we worked to about 50% or better before they were cut, it was the largest workload for a single job we've ever had," Westenhofer confirms. "The main thing I'd have to comment on was how smoothly it went -- a testament to the solid pipeline we have developed in the studio."

Meanwhile, there were a lot of similarities in the nature of the work they did between Narnia and Golden Compass, including the creation of a large number of highly detailed character models, a tremendous amount of keyframe animation and large battle scenes requiring the use of Massive for crowd simulation. "The biggest difference was the amount of interaction between the animated characters and the actors. Daemons are often climbing on the person they belong to, and that turned out to be one of our toughest challenges."

During plate photography, Fink and Westenhofer used various props to represent the daemons on set, from green beanbag "footballs" to mouse sized objects on sticks. They also had an assortment of "stuffies" for the main animals that were used for lighting reference and to help the actors focus their performance. The green bean bag assortments were used to take space and give the actors something to hold on to.

Another early challenge dealt with concepts of how the daemons were actually going to look. The initial plan was to create a special treatment, dubbed the "spirit look," which started as both a creative way of distinguishing them from real animals and as a budgetary strategy to find a way to render the daemons in a more cost effective way. "The early R&D was focused on this look," Westenhofer observes. "In practice, however, it proved difficult to find something that both looked 'cool' and satisfied the intent to cost less than a normal render. In the end, we settled back on a realistic render, but retained a very slight chromatic sheen that appeared in the highlight areas -- a subtle touch to emphasize the special natures of these creatures."

The CG daemons effort was divided among six teams led by Sequence Supervisors Pascal Chappuis, Jimmy Jewell, John Paszkiewicz, Jonathon Robinson, Josh Saeta and Edwin Rivera. Modelers started with maquettes created by Neal Scanlan Studios for all of the hero daemons. For the other characters, the team made good use of Rhythm & Hues' extensive model library and adjusted the models as needed. Using Maya, they built more than 40 base models for the daemon menagerie, with many additional versions to add variety. On top of model adjustments, there were also several variations of fur types and colors to make the total number seem even larger.

The team used the same overall rigging system they have been utilizing since Narnia. It combines a muscle/skin system with traditional deforms. "We have used full muscle systems, but they tend to get unwieldy very quickly," Westenhofer remarks. "We now use a hybrid that employs traditional deforms for most of the work and augments that with muscles and even blend shapes where they provide the most benefit. For areas where skin slide is required, we add a special deformer that allows a mesh to slide within the hull of the already deformed shape. When we combine this with relaxation controls, it acts just like skin sliding over a hard surface. You then blend that with tension driven displacement maps to produce wrinkling skin, and the result is very convincing.

"For hair animation, we used dynamics simulation. The golden monkey daemon [belonging to Nicole Kidman's Marisa Coulter] was particularly challenging in this respect, perhaps more so than Aslan from Narnia. With a small animal, like a monkey, the relatively lightweight of the hair tends not to produce a lot of flowing, overlapping motion in the way that long hair on larger animals does. Flowing motion is what dynamic simulations are most well geared to produce, unfortunately. For the monkey, the hair needed to compress and to pull apart from collisions with the body and other hairs, and needed to show a degree of friction as it dragged over other areas. But it couldn't flow all that much and needed, if anything, smaller vibrations as a result of the monkey's movements. Since dynamics was essentially out as a catchall solution, we needed to find lots of ways to 'cheat.' Technical Animation Supervisor Jen Bahan and Lead Jay Gambell developed something they called the 'hair bag,' which was a cloth simulation that stretched between the skull and upper back. This 'bag' is never seen, but its motion informs the guide hairs how to move. It was a really neat solution to the problems we were having. Altogether, the golden monkey took close to a year to perfect."

From One Daemon to the Next

A tricky part of the daemon assignment, meanwhile, was generating realistic transitions between daemon forms. As a reflection of the changing nature of a child, the characters are able to adopt the shape of another animal effortlessly. This ability disappears when the child reaches adulthood. For lead human character Lyra's daemon, Pantalaimon, no less than nine models had to be built, of which seven appear in the finished film. These included an ermine, wildcat, hawk, sparrow, seagull, moth and mouse.

"The transitions between daemons were really fun to do," Westenhofer suggests. "We were able to take advantage of the fact that the transitions were always intended by the director to be quick and to feel natural (i.e., without a lot of flashy effects). That meant we could essentially do the equivalent of a 2D morph in 3D. An animator would see a rig for the animals on both sides of the transition (e.g., the cat and the bird). He would then dial in the deform to mush the starting character into the shape of the latter, while animating out the same control on the second character to bring it from a mushed shape of the first animal into the proper form of the second. We would render both animals for the entire range of the transformation, and a 2D cross dissolve would control how much of each made it to the final image.

"What really sells the transitions though, has nothing to do with technology of the morph. When the forms were animated, we paid attention to overlapping action that resulted from the shape change. If a mouse changes into a cat, the head is thrust forward from the momentum of the size change, and we give a slight recovery at the end to the cat that really makes it feel authentic. Senior Animation Supervisor Erik-Jan De Boer would often have the implied movement of the shape change be a part of the action the animal was taking."

An unexpected challenge arose when production asked for many daemons to be integrated in plates that had not been filmed to include them. When the crew knew that a daemon would be in a shot, they would leave space on set and frame appropriately. However, since all the characters have a daemon, it meant that, inevitably, editors needed daemons to be added in places where no space was left, or where framing required "creative solutions" to make it feel natural...While animation would often solve these problems by posing the animals to look comfortably in shot, it often required putting them in places where the lighting team would have to deal with additional integration issues with hair, clothes, etc. CG grass, moving snow and gravel were then created in Houdini, while 2D artists did wonders to fit the CG characters in and around hair and furry garments.

For daemons and human interactions, the props that were used on set were efficient enough in the majority of cases. "When the interaction was not adequate," Westenhofer continues, "our tech animation team would add motion to the clothing by projecting the original photography onto a tracked piece of geometry (tracking supervised by Wilmer Lin), and then deforming it appropriately. This worked great for collars, shirtsleeves, and even a comforter. In one shot, Lyra hugs her daemon (in cat form), which essentially required us to bury him into her hair, an action that was not covered by anything on set! In order to make it work, we actually built CG hair for that side of Lyra's head, animated it, and blended it back into her real hair. We even ran a cloth simulation of a sheet covering part of the daemon's body to further integrate him into the shot."

To maintain continuity, no animatronics were ever used to represent the daemons on set. However, real animals do appear as daemons in the film, once in the Magisterium scene, and then in the Bolvanger scene. Fink recalls, "There was no practical reason why the daemons in the Magisterium scene had to be real animals, but the scene was shot very late in the schedule, and the production wanted to take pressure off of us by reducing the need for CG animation. Personally, I think this was not a good idea, but it was a production decision. The animals in Bolvanger [dogs] were used because Chris Weitz had the idea, early in production, that some people would have been successfully severed from their daemons, and these daemons would be more like pets. It was a good idea, but as the film was cut, the story point made by the severed daemons was lost, and they appear just incidentally in the film."

A Massive Daemon War

Rhythm & Hues' daemon assignment included contributing many character animation elements for the final battle sequence. Initial concepts for the action were laid out by Fink and a team of previs artists. R&H was asked to insert hero animation in the foreground in the available ground space, and to augment the action in the background using Massive. Individual background actions were animated early on using scaled down versions of the hero rigs. The Massive team led by Mark Welser and Jason Quintana then created the "brains" to drive the characters, and added them into shots. "We had to keep the CG crowd in the back a bit sparser than in the Narnia battle because our creatures would be sharing space with the stunt actors," Westenhofer observes. "Believe it or not, we didn't really do anything special to get our Massive agents to avoid the humans. It was more of a 'shot gun' approach where we put more or less the right density of battling agents into the shot, and simply plucked any 'offenders' who clearly passed through an actor. Generally speaking, however, the layering caused by our rotoscopers -- overseen by 2D Supervisor Craig Seitz -- resolved the vast majority of such cases."

Rhythm & Hues employed its proprietary renderer Wren for all of the daemon shots. To optimize render times, most Massive agents were furless models with textures that simulated the fur. They were also built at lower resolution than hero characters. Additionally, Wren was able to break frames up into tiles to render them in parallel across many processors. The rendered elements were composited using proprietary package Icy.

Dust particles in this universe are a key source of intelligence, so Rhythm & Hues created fluid simulations to visualize them. Most intriguing is the way the company pulled off the death of daemons during the battle through the fiery disintegration of dust particles using this same fluid simulation.

A unique aspect of the battle sequence was that every shot was handled by two or three different vendors. "The shared shot work went amazingly well," Westenhofer marvels. "The reason for this was that we abandoned the notion of a 'final compositing' house that took all elements and assembled everything. Instead, we used pre-composites, with each facility finishing everything on their end and passing the digital files on to the next house. Cinesite would finish the environments, and then hand it to us to lay in the background daemons. We would then send it to Framestore CFC to put in CG polar bear Iorek, and they would give it back to us to lay in the final foreground daemons."

The project ended up being the first all-digital show within R&H's pipeline. "This was the first time that we did no Film Outs to check either works in progress or finals. We employed Truelight to measure the DI facility in London and give us LUTs such that what we saw on our screen matched theirs as closely as possible. With digital, we were able to have several film resolution dailies per day, and the turnaround time for feedback was faster than it had ever been."

Interestingly, Rhythm & Hues did a large amount of R&D on the final sequence of the film (a surprising confrontation) that ended up being moved to the start... of the proposed second film. "For a number of reasons, the sequence that ends the book couldn't be integrated contextually in the film version, so the choice was made by the director and New Line to save it for the start of the second film. Perhaps some of the concepts we developed will find their way into the production of the next film..."

Alain Bielik is the founder and editor of renowned effects magazine S.F.X, published in France since 1991. He also contributes to various French publications, both print and online, and occasionally to Cinefex. In 2004, he organized a major special effects exhibition at the Musée International de la Miniature in Lyon, France.