In the first Star Wars, when people watched the X-wing flying down the trench, did audiences think that was really happening? Did they believe it was a real X-wing flying? A model of an X-wing flying? What was their understanding of how real that was? Was it more real because the model was part of our world? Is that what people mean by real, a tactile thing? These are philosophical questions that you have to consider when you’re doing a visual effects movie.
I wasn’t interested in making a retro movie. That was not what I thought we should be doing and it would have had a limited shelf life. When you think of those movies, people were excited to see them, to go along for a great ride with amazing characters. Part of that world achieved a certain charm and authenticity because the filmmakers built sets and went to real locations. That part is fantastic. We wanted to hold onto the charm the original stories had but use the technology available now in a way that would make a more contemporary film.
How do you get the audience invested in the characters and willing to go along for the ride? Part of their brain says this is not a familiar world and what they’re seeing is slightly odd. How do you get past that? You want to convince the audience everything is happening for real by hopefully creating a world that’s a natural extension of the real physical world. My theory is that this is a bizarre extension of primitive storytelling. Think of how invested children become in simple puppetry. This is a crazy extension of that. The whole thing is a trick of some kind.
Clearly you were successful in making the digital effects believable. How did you achieve that?
Part of it was trying to photograph as much in camera as possible to help the actors understand the world they’re immersed in and as a better foundation for our work. We went to places and tried to photograph as much as we could. We tried to make each moment as real as possible and blend the line between real and imaginary as much as we could.
f you’re tasked to do something that by its nature undermines the concept, it’s hard to pull off. There’s a certain point, if you overexpose your hand … a film like this is so ambitious, so fantastical that no one could imagine it could have happened. The nature of filmmaking in a real space is that you see the place. You react to it. You design shots around that place. If you remove the idea that you’re really there, it becomes a different thing. We tried to show a real level of restraint to keep it in the charming category without losing a level of excitement, without ever being ludicrous.
In old-fashioned filmmaking, you didn’t go to Tunisia and then change it. You’d go there and that was in the movie. Sometimes, on other films, we might go to a location and then the director would want another version — taller buildings, a different sidewalk. They want their desert a different color. They don’t embrace what they have. This is the VFX era. Every blockbuster has big visual effects. People shoot and figure it out later. Maybe that’s why so many people shoot on green screen. We didn’t want to do that. We had a director who clearly wanted to go to the desert, and he was happy with the desert he went to.
When you have a ship that travels 600 to 700 miles per hour, it’s difficult to shoot plates, especially with the shot choreography we wanted, unless maybe you hire a jet fighter. So, in the Falcon chase, pretty much everything is digital. We could have just based it on the location and made up the rest, but we would have been in a place of guesswork. So we shot a lot of reference footage from a helicopter. It wasn’t directly usable, but it was incredible reference. You’d notice things about the way the desert behaved — dust coming off surfaces as we travelled. More importantly, the quality of the light on the sand, the geography of the environment. We ended up with 18 hours of aerial footage alone from one camera, and terabytes of photographs. We could see the nuances of colors of the sand dunes from photography based on the time of day and conditions.
The other thing we did was to survey the crap out of everything and everywhere we went. We researched and recorded environments as accurately as we could. We had a whole team on this. We had size and scale. We had stereographic views. This isn’t outside the bounds of what people do for visual effects movies, but we really wanted to be accurate.
When you see the little speeder moving across the landscape, it’s a completely digital shot. But because we’d been there, when we recreated that moment, we did it from a physically-based approach. I don’t know if J.J. [Abrams] knows how we constructed the images, but he was at the location and he knew how it looked.
The other thing is that we photographed these events in a way that I think did not cross the line of the impossible with complicated camera moves. We made sure if you had a Millennium Falcon, this is how you’d shoot it. We tried to photograph some kind of version of all these events to give us a point of reference even if we threw it away. One advantage I had in putting shots together in the final movie is that I shot second unit. If a shot involved visual effects, I knew what we’d need to achieve.
And our modelers had so much affection for these ships. Dave Fogler [asset build supervisor], who came to ILM from the model shop, was interested in rebuilding the Falcon in the way it should be built, with integrity and a referential quality. He had an inherent knowledge of the processes the model shop guys used. It was a very loving process in many respects.
Probably about a third of the time BB8 is digital. This is a perfect example of riffing off something that exists. It’s another version of my desert analogy. If you don’t build BB8, where does the personality come from? The personality came from being there with the actors and J.J. directing BB8 just like another actor. When we created the CG character, we had absolute reference. Each animator had a template. People could comment about the droid’s performance just like they would with an actor because his personality was defined.
We did motion capture tests at Andy Serkis’s [Imaginarium in London], and then did motion capture on set. Ben Morris, who set up the ILM studio in London, was on set for the shoot, and Mike Mulholland supervised the London work. I’ve always believed that if you cast the right person, the rest takes care of itself. You have to believe in the actor; motion capture won’t save a bad performance. Maybe it might, but chances are it won’t. Casting is what it’s all about.
Andy’s character didn’t move around very much, but we needed a great face performance. We shot Andy and then re-mocapped his scenes once we got the edit. We went back and refined the shots in his studio. He can play anything — small, large, in-between. Lupita is way more restrained. She does a softer, nuanced kind of performance. For her, we used image-based capture and some keyframe work. Andy’s studio helped us out with that, but it was really about the face, and we used systems from Disney research. The face of the character she plays is very different, so we had to make sure that translation was successful. But again, like with BB8, she defined the character, the nuanced performance.
I think I’ve learned that maybe showing restraint is something we should do more of in the future. Restraint sounds negative, and that’s not the connotation I mean. Just trying to make sure the shots are focused. More isn’t necessarily better.
It’s interesting how much the work has advanced, even in the last couple of years, and certainly from Episodes III to VII. It’s a completely different lay of the land, and the people are capable of so much more. You don’t want to do things you’ve done before. You want to do something more challenging and exciting in different ways. Sometimes, by defining what the box is, you can have a more interesting problem to solve.
- See more at: http://www.studiodaily.com/2016/01/ilm-vfx-supervisor-roger-guyett-on-star-wars-the-force-awakens/#sthash.4KvYrE0k.dpuf