When the Academy’s visual-effects peer group meets today to vote on this year’s short list, among the films they’ll be examining is Warner Bros. The Hobbit: An Unexpected Journey. Industry audiences and ardent fans will be pleased to see the familiar goblins and orcs, but visual-effects supervisor Eric Saindon says much of the technology underneath the characters is virtually unrecognizable from the first Lord of the Rings trilogy. Saindon recently spoke with AwardsLine about how much things have changed.
AWARDSLINE: What’s the biggest difference between the technology used on The Lord of the Rings and The Hobbit: An Unexpected Journey?
ERIC SAINDON: Back on Rings, we motion-captured Andy (Serkis), but on the first two Rings we motion-captured him on a stage. We got very rough motion—it was not bad, and it gave us his general performance. Then we always had lots of video cameras on him so that the animators could go in and then hand key-frame all of his facial poses. On this film, we actually capture all of Andy’s performance when he’s acting with Martin (Freeman), and we capture all his facial performance. We have a small camera attached in front of his face that actually captures his exact facial performance. Instead of an animator going in and putting that animation—or Andy’s performance—back onto the puppet, we sit with Andy and we go through his (performance) as a separate thing: “OK, this is your happy face,” “This is your sad face.” Our computer analyzes what pose he’s in when he’s on set making all these faces and puts it back into the pose Gollum would be in if he was making the same pose. Rather than an animator going in and doing it frame-by-frame, the computer analyzes Andy’s performance and then fires Gollum’s muscles to do the exact same thing. Really, the big difference is on Rings, everything was captured post. Frodo (Elijah Wood) did his thing on set—Andy was there most of the time acting also—but then Andy would have to do it again. Three weeks later, Andy would have to do the same performance, and Elijah’s performance couldn’t change. (This time) Martin and Andy actually just acted together, they acted off one another, and that performance—that exact performance—went back onto Gollum.
AWARDSLINE: It’s amazing that Andy was able to achieve the performance that he did in those first films.
SAINDON: Well, he’s not afraid of the technology. A lot of people get the suit on, and then they freeze up. Outside of the suits they did great, but then you get them in the suit and it takes them two days to get used to (it). It’s definitely an art that not everyone can do.
AWARDSLINE: What kind of challenges did the 3D aspect of the film pose for you?
SAINDON: One thing we did a lot of on this movie is 3D scans of every single set, and with this 3D scan we could bring it to the computer and have a 3D representation of the entire set with textures, model detail—everything that was actually shot. Because everything was in 3D in this movie, we needed a proper depth. On Rings, we could easily just cheat something. A foot contact that didn’t work or something that didn’t look like it was the right depth or space, we could just scale it down a little bit. With this movie, because everything was 3D, we had to do it at the proper scale and the proper depth, so when you’re watching it, it doesn’t hurt your head, and it doesn’t pop out as (not looking) quite right.
AWARDSLINE: As the technology constantly changes and improves, does it ever get easier or faster to put together a movie like this? It seems like the answer is probably no.
SAINDON: I would love to say that it’s easier, but it’s never easier. Our computers are 50 times faster than they were on Rings, and it still takes all night to render a shot properly. (Plus), the audience is expecting more. The general viewer going to see the film can spot a digital double, where back in the day, they really didn’t think about it as much. So you could get away with a little more. But nowadays the digital doubles have to look just like the actors. On Rings, we did very simple digital doubles (with) little bit of cloth and a few strands of hair. On this movie, we’ve done full-cloth simulations, skin dynamics, fat dynamics. It’s taken to a much more extreme level. Obviously, you can do things with a digital double that you never could do before, but it also requires a lot more computer power—a lot more time—to get it right. On Rings, we had about five terabytes of disk space for the whole movie. Then on this one, we’re in multiple petabytes of information, which is insane.