CGSociety talked exclusively to Jeff Unay, the Facial Lead at Weta Digital about his experiences in production with James Cameron on AVATAR. Now with nine Oscar nominations under their
belt including one for Visual Effects, the crew on AVATAR have even more reason to feel good.
Transforming the actor's performances from the sound stage actions through into 100% Na'vi-laden animation. During the capture of each and every scene with Na'vi, the actors were mocapped from head
to toe. Jeff Unay takes up the story.
"New technology was developed on set to capture close up video of the actors' faces through the use of a head rig," Unay describes, "essentially a helmet with a small camera pointed a few inches
away from the actors' faces. So for any frame of any shot, we had a record of what their faces were doing at all times. During shooting, the actors performed with little painted dots on their
faces."
Weta developed new 2D tracking and solving software that allowed these dots on their faces to then drive the facial rigs via
a FACS-based (Facial Action Coding System), muscle mapping approach.
Andy Jones, Weta's Animation Director on AVATAR, oversaw the character and creature performances from all stages of production to ensure continuity between motion capture, motion edit and keyframe
animation.
The Weta facial team is the group that designs and constructs the various facial rigs that are driven by the motion capture and animation departments.
They use this face cam video primarily as frame-by-frame reference to learn how each actor 'uses' their face during their performances by breaking down the expression into what facial muscles they're
activating. This video reference is invaluable to us when creating/editing individual or combination FACS poses in the rig.
Generally speaking, the hardest area to manage was the sheer quantity of facial rigs needing to be built for AVATAR.
"In one of the first conversations we had with Jim when I began on AVATAR back in 2006 was a numbers question," says Unay. "'How many Na'vi are we talking about here?' we asked. And the rough estimate
of unique Na'vi on screen was 300. That was all we needed to hear in order to begin constructing a facial pipeline that would allow us to create a high number of rigs, with a small number of people,
within a relatively short amount of time."
"We already understood the quality that was expected on such a film. All you'd have to know is this is a James Cameron film and everyone on it knows he won't accept anything but the highest quality, a
new industry standard for facial rigging.
"This was motivating for us. We were all up for the challenge of creating such a robust facial pipeline."
Before AVATAR, Weta had created memorable characters like Gollum for
'The Lord of the Rings' trilogy and Kong for 'King Kong,' but that was only one main digital character per film.
"In the beginning, we asked ourselves 'How do you create 300 Gollum-level
facial rigs?'"
"Somewhere in the middle of production on AVATAR, we built a pipeline that would allow us to build a new facial rig for a new character with one button push. No exaggeration. So if we can create an
initial facial rig with all rig components transferred, we can have an artist refine the rig to the nth degree. The artist won't have to worry about all the technical steps necessary to begin a
character and also to publish a facial rig."
Early in the project, Weta delivered mocap facial rigs for the principle characters in AVATAR. These low resolution rigs were designed to be used onstage and reviewed in Motion Builder.
This gave Cameron the ability to see the Na'vi or AVATAR character's complete performance from head to toe on a monitor while directing the actors at the same time.
