After sharing an editing credit on “Titanic” with two other people, helmer James Cameron planned on cutting “Avatar” by himself. But even the indefatigable Cameron soon found there was simply more sheer work than he could handle. Eventually he again found himself with two collaborators on his billion-dollar blockbuster: John Refoua and Stephen Rivkin.
“Our original projections about how the film would be (edited) were predicated on an early test with a simple scene of two people talking — but pretty much everything else in the film is not two people talking,” says Cameron, acknowledging that “the amount of labor and hours required … was probably our biggest underestimation. It quickly became clear that I wasn’t going be able to do everything.”
At first Cameron resisted the idea of hiring other editors, thinking, “How do you bring someone up to speed when there’s no one in the world who knows how to do this other than within our group? We’re starting from scratch with a new process. The answer: get smart guys and have them figure it out with you.”
Cameron hired Refoua first, but then “we realized that John would not be able to handle it all, so we brought in Steve.”
Editing “Avatar” was complex largely because most of the film’s footage consists of performance capture, a technology that offers many more editorial options than straight live action. It also let Cameron exercise his perfectionism and explore a process that goes way beyond standard live-action editing by giving him the ability to isolate and combine the best individual actor performances from each take.
In fact, the team had to invent new terms or meanings for what they were doing, such as a “stitch,” a “combo” and a “load.”
As Cameron explains: “You might like Zoe (Saldana) from take four and Sam (Worthington) from take eight. Now you’re doing a ‘combo.’ Or you like Zoe from take four until a certain line of dialogue, then from take seven. Then you’re doing a stitch — stitching her motion together.”
Rivkin elaborates: “Normally, when you shoot a live-action movie, and there’s more than one actor in a shot, you’re limited to the performances in that take. In ‘Avatar,’ after the performances were captured, Jim often wanted to combine performances from different takes into a camera ‘load’ — or section of a scene — and build a composite from different performances.
“We were able to create a ‘master’ that had the best performances from every actor — combining actors in the same load and ‘stitching’ them together if necessary,” Rivkin adds.
“We could also use the performance capture from two different takes of the same actor, which could be made into a continuous take, so you aren’t limited by one particular performance. The first line can be from one take, the next line from another.”
The goal was “to have every option available,” Rivkin emphasizes. “Some actors peak at take one, some at take 10. By combining different performances in the same load, you have the flexibility to create the best possible master.”
This “building of performance loads was one of the most complex workflows you can imagine,” Rivkin observes.
The three editors worked together closely. “Steve and I were right next to each other, Jim was down the hall,” says Refoua, who got a longer gig than he bargained for when he joined the project. “Originally I came on for just six weeks to help out. I was still there 120 weeks later.”
Rivkin, who also worked for 2 years on “Avatar,” was anticipating a break after working on two “Pirates of the Caribbean” films back to back when he got the call from Cameron.
“Jim was looking at a set on his viewfinder. Nothing was there on the stage, but in his camera there was a whole virtual world. … He showed me some of the stuff they were working on and I realized that day that this was an extraordinary project.”