James Cameron goes deeper into VFX and the significance of Avatar.
On Feb. 28, the Visual Effects Society will honor James Cameron at the Century Plaza Hotel with its Lifetime Achievement Award as part of the 8th annual VES Awards. So, I invited Cameron the opportunity to put Avatarinto a much larger VFX context, reminiscing about The Terminator, Aliens, The Abyss, Terminator 2: Judgment Day and Titanic as well.
So, we're gonna talk about effects, huh?
Bill Desowitz: Yeah, we're gonna look back at your previous films and talk a little bit more about Avatar.
JC: What hasn't been said? Is there anything that hasn't been said?
BD: I don't know. Did you have the retreat?
JC: Yeah, we did the retreat and it worked out very nicely.
BD: What did you discuss?
JC: Well, what I was looking for were themes that were common to everyone's comments, and really just the fact that we explored a few blind alleys and it took that to come up with the best possible version of the process, and we didn't start firing on all eight cylinders until the last few months of the virtual production process, and I think we now believe we know what that should look like from the inception on another project, so we think we can do it a lot more efficiently.
BD: Any specifics?
JC: No, it's all proprietary stuff at this stage. The stuff we were talking about was generating a competitive advantage over everyone else. I don't exactly want to blather on about it.
BD: What about more generally?
JC: Well, the consensus was that, actually, we had done it right. We made a lot of mistakes, but in refining it we eventually wound up at what we had imagined to be the best possible approach using current technology. As newer, faster render engines become available, we will ultimately convert to something that probably produces a result that looks more like a contemporary game engine with realtime shadows and lots more moving effects like wind, but we're not gonna do that in the next generation.
BD: And you were obviously pleased with the performance capture.
JC: Yeah, I mean, we looked at it from all perspectives, including getting Weta's perspective and Fox's perspective. And the consensus was it took three-and-a-half years to get to a point where we were really running optimally, but we got there. And if it looks like that from day one on a new project, then it's going to run a lot more efficiently and take less time and money.
BD: I know it would've been groundbreaking to get Zoe Saldana an Oscar nomination. What's it going to take to turn around the perception that it's more performance than animation?
JC: It's a learning curve. And the acting community needs to understand it before they can embrace it. And right now they don't understand it. They don't know if it's animation; they think it's like you do a voice part and then a bunch of animators create your character and they're not that interested in that. You know, if it's three-day's work and they can get a lot of money to do the voice for a Pixar film, then they're fine with it. But let's not start confusing that with Academy Award quality acting. But when they realize that it's not three-day's work but a year's work, including months of preparation and months of work and it's as exacting and as disciplined and as pure an acting form at least as what you'd have acting for a camera. It's just there's no lens. Then they'll not only understand what it takes but also what the possibilities are. And how it doesn't erode their jobs: it actually creates new jobs and new possibilities. But that's not going to happen overnight.
BD: No, but, on the other hand, there's also a greater appreciation of the animation feats that you accomplished.
JC: Right, the thing is, there were two forms of animation on Avatar: there was classic keyframe creature animation, vehicles and the plants. And bringing these creatures to life we had the best animators in the world and they looked alive: the way they breathed, the way they moved, the mass, the agility, even down to the high-frequency fluttering of the trailing edge of the wing membranes of the Banshees. I mean, there was an incredible observational attention to detail. But when you say animation in connection with the human performed characters, you have to use the term in a very narrow sense because even though they were the best animators in the world doing it and it took all of their skill, they weren't animating in the classic sense of creating a character. The character and the performance were created by the actor. And it was up to the animator to preserve that performance without embellishment and without diminishment. And that's a different category of animation from a Pixar-type character animation or the kind of creature animation that they were doing on the same movie. The same animation team was doing Banshees and Hammerheads and Thanators at the same time they were doing Neytiri and Jake. But they were very different specializations.
BD: Have you given more thought to a sequel?
JC: No, I haven't had a chance to think about anything except running around and doing interviews and red carpet and award ceremonies.
BD: OK, so let's look back at your VFX experience and how this is really a culmination of sorts.
I did an NPR interview and they were talking about the [Roger] Corman days and Terminator was an extension of techniques from when we were doing the Corman films. It was a snapshot of that time and how you did stuff in camera and a little bit of opticals and a lot of practical gags. And I realized that pretty much everything that we did in visual effects in those days has been replaced. Everything! We don't use film -- we don't even use lenses anymore. We certainly don't do bi-pack and optical printing: everything's digital composite. But the essence is the same: you still have to design it and previsualize it and execute it -- and you still have to be telling a story with characters and you've still got to do lighting even though it might be virtual lighting as opposed to actual lighting.
BD: Linwood Dunn told me the same thing about storytelling remaining basically unchanged despite changing technology. And that was around the time you and I first met to discuss Terminator.
Yep, as you well now, Linwood basically invented the optical printer. And they used to call it The Trick Department, and I strongly debated when we were founding Digital Domain whether to call it Digital Domain or The Trick Department. I always loved that and the idea that there were three guys back in a room someplace who were the only ones on the studio lot who knew how this stuff was done. And it was like Linwood and two assistants.
BD: So you came along at just the right time.
JC: We anticipated that [transition]. When we founded Digital Domain, Scott Ross, who was one of the three founders, was coming from the ILM culture and he said he had some great ideas for who to staff the optical department with. And I said, "Scott, there's not going to be an optical department -- it's called Digital Domain." We are going to be the first company that starts from scratch, all digital. And he said it won't work. But obviously he got talked into it eventually, because that's what we did.
And what was terrifying one year was obvious the next -- that's how rapidly it was changing. But when we set the place up, we were still on these SGI Onyx engines and it was a pretty obsolete paradigm by present standards, but when we set it all up, the first thing we had to figure out was, what's our all-digital composite color space? We brought in Price Pethel and he set up the whole pipeline and for how composite would work and we had code writers that wrote Nuke and Flame and all that stuff and it was really an amazing and fertile time. And it was a turning point because all the other houses went, "Oh, shit! They don't have an optical department!" And there was about a two-year period where we were the best out there at 2D composite. But, meanwhile, ILM were leapfrogging forward on CG character creation doing the Jurassic Park films and they were just blowing us out of the water in that area. So we actually got into these specialized ruts for a while and Avatar was actually a direct response to us being in a 2D composite rut. We founded this company to be like Disney and creating characters and figuring out ways for CG to become the new form of character creation because that's what Stan Winston did for a living, and when Stan and I founded the company, that's what we wanted to be doing. So, I said, "All right, I'm going to write the ultimate, kick-ass, CG character and creature movie and then we'll have no choice but to be the best in the business or die trying." And that's what Avatar was.
BD: What do you think of in looking back at your other films?
JC: There's an overarching narrative, and it goes like this: Terminator and Aliens were just an outgrowth of what I had been doing at Corman in terms of technique: front projection, rear projection, in camera stuff, optical printing and motion control. And maybe a few new tricks that Stan Winston and I had come up with for the prosthetics. You get to The Abyss and there was one new idea, which was a soft surface, computer-generated character. And, in fact, that character, because it could transform, actually incorporated some of the principal characters, briefly. This was a big idea and it was riveting to people when they saw the movie. When people think of The Abyss, they either remember the scene where Ed Harris brings Mary Elizabeth Mastrantonio back to life as a piece of drama, or, from an effects standpoint, they remember the pseudopod, and we had no idea how good it was going to turn out when we started: we just took a leap off the cliff and it worked.
But at that point, I started to realize the power of dream imagery: something that couldn't be explained. That was cinema magic. And people use that term a lot but I'm using it in a very specific sense: the way Clark defines magic. Any sufficiently advanced technology is indistinguishable from magic. And, for me, when I was seven-years-old, it was Mysterious Island and Seventh Voyage of Sinbad. I had no idea how that stuff was done -- it was magic. So we were showing people discernible magic.
So I thought, all right, with Terminator 2, I had already thought of the liquid metal guy seven or eight years earlier, when I was writing the original Terminator, but I took that character out because we didn't know how to do it. Now we knew how to do it, so The Abyss was a sort of wet run for doing Terminator 2. So now we leaped off the cliff again, it was your main villain sot it had to work throughout the film. And everyone remembers the liquid metal guy from Terminator 2, but not too many people remember that there were only 42 CG shots in the whole movie. And it seems to almost go up by an order of magnitude every time I do one of these things because then we came to Titanic with 420 effects shots of all kinds but mostly CG shots. Even some of the model shots had CG crowds, to Avatar, released 12 years later, which had 2,600 CG shots -- basically, every shot in the film. I think there's less than a minute that's not an effect of some kind.
Bill Desowitz is senior editor of AWN & VFXWorld.