"They (the animators) may tweak the camera to best play to the animation of a character or a shot," Wilkinson says. Animators have the ability to check or view a shot in stereo at any point, but generally concentrate on creating their performances in 2-D.

Once approved, the shot is software rendered, then heads back to the mono pipeline for master lighting, rendering, paint and composition. The rendered elements are then returned and the additional pieces required to render the second eye are created so it can be assembled into a stereo film.

"We're not relighting it for stereo, we're not making new materials for the story, so we're pretty intelligent and efficient about how we go back and forth between the two pipelines," he says.

There will, however, be some adjustments between the 2-D and 3-D, such as depth of field or blurs that need some tweaking before the movie is finished, Wilkinson says. The main motivation in fashioning this pipeline was to keep the processes that had worked for previous Blue Sky projects while avoiding redundancies. That has posed technical challenges, such as synching up the work schedule for both versions in the pipeline, as well as finding animators with the right skill sets.

"I need people who are good camera people and good compositors -- so there are two roles on their head," Wilkinson says. Animators also, again, can't use the 2-D cheats, and while it's possible to fix many such shots, a few do have to go back to animation to be fixed up. Education and experience will minimize such problems in the future, he says.

Depth blur and optical effects like lens flares and halos have proven a particular challenge in stereo. Blurs tend to flatten out and optical effects, which typically happen in the eye, are difficult to place in 3-D space to look believable or natural.

Bolt's creators used a multirigging technique to dial in appropriate depths for individual elements in a single shot © Disney Enterprises.T

Some of the ways Wilkinson thinks 3-D can add to storytelling is in shots where a character can lean into and enter the audience's space, as well as the overall feeling of depth in a scene.

To date, most animated 3-D films have been converted from 2-D versions. Over at Disney, producer Don Hahn, who has worked on converting both older films (The Nightmare Before Christmas) and recent ones (Meet the Robinsons and Chicken Little), suggests that the process is more complicated than it sounds.

"On CG films, for example, there are many patches and paint fixes on the original 2-D image that reveal themselves in the conversion process. A lot of the work on those first films was deciding if it could be done, and, if so, how much patching and painting work had to be done to fill in spaces behind characters," he says.

Early films also posed major aesthetics issues, as it had to be decided how far to take the 3-D effect. Hahn says they tended toward a conservative approach where the 3-D effect was more subtle and only popped out occasionally for a big effect.

Hahn is currently four months into converting Disney's Beauty and the Beast, though, with some improvements. The production began by going back into the original CAPS files for the film, which were converted into a more contemporary software package with all the original elements preserved in layers.

The 3-D models for the characters are created with a number of new techniques, including a proprietary system that automates the process.

On the aesthetic side, Hahn says he's working to create an immersive experience for the film, so that the audience feels like it's in the ballroom and the hallways of the castle appear miles long.

Meanwhile, Walt Disney Animation Studios took a similar approach to DreamWorks with Bolt, which is up for an Oscar for best animated feature. Robert Neuman, stereoscopic supervisor on the film, says this was the first feature at the studio where 3-D was part of the making of the movie instead of being a mostly post-production process.

While 3-D was integrated into the film's design early on, Neuman says it was still the 2-D engine pulling the train. "The challenge was to deliver an uncompromising 3-D experience, but still be able to make the best 2-D film that the filmmakers wanted to make at the same time."

The early creative development was done in 2-D: storyboarding, animatics and editorial, Neuman says. Once sequences with shot breaks were determined, shots went to layout, where the final 2-D and 3-D cameras were built. While many creative decisions had been made, Neuman says this was where the ways in which 3-D can add to the narrative were considered.

"We had the sense that we were going to be using 3-D as a storytelling device, something more than just throwing on an added dimension to it," he says. "We wanted to use that as something that would actually support the narrative to the film instead of something that's just tacked on."

Supporting that was a depth script, in which Neuman went through the entire film and used the major beats of the story to guide the depth of each shot. "We'd wind up at the end of the day with a number that would reflect the emotional intensity or the conflict level of the shot, and that would be what we'd use to guide the application of depth."

Depth was especially useful in amplifying the emotional impact of each shot by controlling the relationship of the subject to the screen. "If the audience is supposed to feel closer to a character, then we'd be literally taking the character and putting them on the audience's side of the screen," he says. "And if we were trying to create emotional distance, we'd do the opposite and push them behind the screen."

The amount of depth used was governed by a system of supply and demand, with the demand limited by what was comfortable for the audience.

One tool used on the film to ensure a progression in the 3-D effect is what Neuman calls a "floating window." The technique is essentially a mask introduced to each eye with a stereoscopic offset that allows the filmmakers to control the perceived location and orientation of the screen.

"Say (that) for narrative purposes we wanted to keep a character behind the screen. I was able to buy extra stereo depth bandwidth by moving everything out into the theater," he says. "That was a huge win for us, being able to use it in a narrative sense and still maintain a comfortable viewing experience while giving plenty of volume and internal depth to the screen."

Bolt also used the multirigging technique to dial in appropriate depths for individual elements in a single shot. Neuman says this reduced the effect of individual elements looking like cardboard cutouts, a particular problem when using a long lens. Similarly, the film also dealt with depth of field as an effect, added in compositing so it could be adjusted for the 2-D and 3-D versions of the movie. The production went for efficiency in using the left eye image as the 2-D version for most of the movie.