Furious FX took on their largest project to date with Race to Witch Mountain, with an eventual shot count of 860. All images © 2008 Disney Enterprises Inc. All rights reserved.

In 1975, Walt Disney Studios used clever in-camera tricks, wire effects and optical composites to achieve the magic of Escape to Witch Mountain, a family film in which two teenage extraterrestrials desperately try to go back to their spaceship. The movie was so popular that it spawned two sequels and a remake. So it comes as no surprise that Disney has rebooted the franchise with Race to Witch Mountain, taking full advantage of CG.

That task was awarded to Furious FX -- it was to become the largest project ever undertaken by the Burbank-based company. "This project was greenlit at 353 shots, with a majority of the visual effects elements shot on second unit," Executive Visual Effects Supervisor David Lingenfelser says. "During filming, we all had a sense of it growing, mainly due to the amount of camera coverage, and so many successful takes between the stunt and action sequences. Once the first edit was assembled, and everyone loved the excitement level in these scenes, we found ourselves nearing 960 shots, though these eventually tightened to around 860."

Lingenfelser adds that the biggest challenge was adapting to the growth of the project, and ramping up their smaller facility with the highly skilled artists, hardware and software needed to create all of the vfx. "We had prepared for the initial shot count, though my business partner/Executive Producer Scott Dougherty and I soon found ourselves posting job listings, then setting up phone interviews and negotiations while working on location! In the end, we had six months of post-production, which was a good amount of time, though quite solidly packed and fast-paced for our 40-person crew. In addition to the work itself, we also had to accommodate a variety of studio needs such as temp versions for trailers, marketing presentations, and audience preview screenings. These are all key ingredients for a successful film release, and we have grown accustomed to allowing for these deliveries in our schedule."

Overall VFX Producer David Yrisarri brought Proof on the lot during pre-production to previse the three major sequences that had the most complex effects. These approved animatics provided the reference for the plates that were shot on location. "We are fans of capturing everything that can be got in camera, as is [director] Andy Fickman," Lingenfelser notes. "When the need arises to look beyond live-action solutions, that's when we discuss how a CG approach can help. In the past, we have worked on movies where there seemed to be a desire to use CG for no particular reason, though it is always better to complement with a digital approach due to limitations of the set, location, actor safety, time constraints, etc. The two work much better hand-in-hand."

Witch Mountain was entirely created on Mac platforms. The primary 2D pipeline consisted of Nuke for compositing and Silhouette for rotoscoping. 3D work was based on Maya and mental ray.

Frontal Impact
Furious FX faced one of its more challenging shots with what later became the most memorable image from the film's marketing campaign: a slow motion shot where alien teen Seth (played by Alexander Ludwig) takes on the full impact of a speeding SUV. "The success of that image depended greatly on the photorealism of the vehicle and its animation. In addition, the actor was shot fairly close-up using a wrap-around camera move, and the decision was later made to slow the action down to a crawl at the moment of impact... So, there was no relying on a quick cut or motion blur to make things easier. Every piece of debris required the same level of detail, right down to the last lug nut!"

This particular shot, and the rest of the project, was entirely created on Mac platforms. The primary 2D pipeline consisted of Nuke for compositing and Silhouette for rotoscoping. 3D work was based on Maya and mental ray, with some tasks being completed in LightWave.

The first step was the creation of a 3D mesh of the SUV, which utilized the on-set cyber-scanning services provided by Eyetronics. The practical SUV was photographed using a high-resolution camera and two flash units. One of these was used to project a grid on top of the SUV from multiple angles, while the other captured texture information. The data was then used to generate a highly detailed and accurate computer-based model of the vehicle.

"The next step involved setup and rigging of the model to allow for animation," explains CG Supervisor Mark Shoaf. "A combination of high resolution blend shapes and IK rigging was used in order to properly affect the SUV in relation to its collision with Seth. The entire crash was then hand animated by a character artist utilizing this setup. This approach turned out to be more efficient and less problematic than trying to simulate the collision using rigid body dynamics. The final step was setting up the full-detail render. High resolution texture maps were painted and applied to the individual surfaces of the SUV, an HDRI environment was built using photos taken on location, and the final images were then rendered using mental ray."