Spring 2019

Cover Story

The Innovators

James Cameron and Jon Favreau, who have made it their business to push the cinematic medium beyond what was previously possible, exchange views on technology, their roles as directors and achieving their creative visions

Directors James Cameron and Jon Favreau (Photographed by Kevin Scanlon)

At Manhattan Beach Studios, in the heart of L.A. County's South Bay, directors James Cameron and Jon Favreau have been long immersed in projects shrouded in secrecy: Cameron working simultaneously on the four planned sequels to his groundbreaking 2009 blockbuster Avatar, and Favreau on a TV series based on the ever-expanding Star Wars universe, The Mandalorian.

On the surface, the two filmmakers couldn't be more unalike. Cameron broke into the business via visual effects before rapidly mastering all phases of the production process, while Favreau started out as an actor-writer who closely studied the methodology of the directors he worked with. With Terminator and Avatar, Cameron has built his own franchise empire; with projects like the Iron Man movies, The Jungle Book, the upcoming The Lion King and The Mandalorian, Favreau is bringing wondrous dimension to existing Marvel, Disney and Lucasfilm properties.

What they have in common is a whole-hearted embrace of bleeding-edge technology, the ability to keep pushing the envelope and raising the bar on their craft, and the endless curiosity characteristic of the medium's pioneers past, present and future. They have visited each other's sets and learned a thing or two in the process. And Favreau has benefited from the proprietary technology that Cameron and his Digital Domain company have honed to help bring the characters and the world of Avatar's Pandora to vivid life, while wielding a few tricks of his own.

Recently, the two sat down to talk tech, directorial process and the primacy of storytelling, character and emotion when paired with eye-popping VFX. The format was freewheeling, while the intent was to simply be a fly on the wall and not get in the way. The following is an edited version of their conversation.

(Photo: Kevin Scanlon)

F: The first time we really talked was when I interviewed you on stage at the DGA (a Guild 75th anniversary event in 2011). I studied up on all of your work and watched it all back-to-back. It was very insightful for me, not just the technology, but how a career evolves. And that you can't really look at individual films, but you really have to look at the whole body of work to see how things start in one movie and then maybe two movies later, they come to fruition.

C: I've actually never done that—watched a person's entire body of work, like a binge watch. But that could be really instructional.

F: In your work, especially. Because there are archetypes and themes and story points that you keep being drawn to, and even characters, and certainly technologies. You have to look at The Abyss to understand T2. And you have to look at The Abyss to understand Titanic and then you have to look at Aliens to understand Avatar.

C: I have often said I couldn't have made Titanic a good film if I hadn't done The Abyss and I hadn't made the mistake of losing sight of where the human drama was eclipsed by the necessity for spectacle. I don't think you get spanked by an audience not liking a movie. I think you get spanked by you not liking a movie—the feeling of frustration that you didn't communicate what you wanted to.
It's getting the tech in balance with the drama and the emotion. I think I was a little more in love with the visuals on The Abyss, and by the time we got to Titanic, I thought I can't do that. I've got to pick my battles.

F: Well, you had so much to do on top of trying to break through in the CG space. Just the tank work and the diving work—I don't know how that can't take energy away from other aspects of filmmaking when you're dealing with so many logistical challenges.

C: Water almost psychologically does what it does to the body, which is to suck the heat out of the body, the energy out of your brain. You're just numb after a while.

F: I try to keep as much energy focused on the task at hand. So I find I get a lot of energy and recharge from being in my environment, being around my family, living at home, having reasonable hours—because these movies take so many years. The Lion King now has taken me three years. And that one starts off as an animated movie with storyboards and every way you break that out. Then, in sort of an extension of what I learned in looking at your work in Avatar, I understood through making The Jungle Book, we're working with those technologies.
In [The Lion King's] case, VR as opposed to motion capture. But essentially the same thing, virtual camera…

C: So you're standing all the creatures up with key frame?

F: Yes.

C: And then are you going in and doing all your cameras using the VR interface?

F: Yes. Because I kind of see each movie as a puzzle, and I try to apply whatever technology's relevant at the time to that puzzle. In Jungle Book, we had a real kid and parts of real sets, and so it was grounded through that. The characters that were CG and the set extended upon something organic. [On Lion King], there was nothing organic at all. There were no sets, there were no characters, there's no motion capture. There's no photography at all. So we eliminated lights and cameras and sets, and we created a volume, which was not unlike when I visited you as you were doing capture for Avatar.
So in Lion King, we set up the animation using the game engine Unity. We built all the sets first, and then we would go in VR to the environment. So we could actually walk around and would do scouts together, all in VR, in the real environment.

C: Using headsets?

F: Yeah. HMDs [head-mounted displays].

C: And how are you creating shots?

F: First we would go in there with (DP) Caleb Deschanel and (VFX supervisor) Rob Legato and (production designer) James Chinlund and just scout. It's like six people, in headsets, together, and we created basically a multi-player filmmaking game. And instead of just using a handheld on splines, with a screen, we used the hand controllers so we could scout within it. And then when it actually came time to film, we had a full crew in the volume.
So we had a dolly grip, a camera with wheels and cranes. And so the human touch and the analog feeling of the camera movements gave the key frame animation more of an organic look. And so by creating a really robust interface that mimicked the set, even with an AD [and] a full crew. So if you walked on our set, it would look like a movie set except you were in the middle of a black box.

C: There's no lights.

F: Right. And then when you popped on the headset, you would see the lights. Or the skydome. And we would move that. And so Caleb would prelight for a long time, and he's never been involved with any technology…

C: But lighting's lighting. I think it's healthy to have somebody who understands real-world lighting coming into a virtual world and approaching it as they would photographically.

James Cameron conferring with Sigourney Weaver on the set of Aliens. (Photo: Photofest)

Old-School Tropes in a New-School Paradigm

F: And so you had all the tech people who didn't necessarily understand filmmaking that much, and you had the film people who didn't understand tech. And by joining them together, what we ended up doing was inheriting a hundred years of cinema tradition through the people who it's been handed down to through apprenticeship or school.
We were just building on what you did with Avatar, and [applying that to] Jungle Book. And a lot of the same people you had trained were working with me.

C: Bob Legato and (animation supervisor) Andy Jones—yeah.

F: That whole crew. But then you have to explain to them, "You can't just walk over and move something" [on the set]. I find with a lot of new technologies, it doesn't just disrupt the technology, it also disrupts the culture. And I tried to maintain that culture on set.

C: It actually creates a new culture. The thing is we're in a merged state between the cinematic paradigm of cameras and cranes and Steadicams and all that sort of thing, and the VR paradigm, which is just the CG camera, which can go through the eye of a needle and rocket up to the top of the Empire State Building.
And what I find is that the Steadicam and the dolly and the crane that you are physically using on set—see, I rejected all that 12 years ago—and said, "I'm going to keep it in my mind." Because I've worked with all those tools and I know them. But where I think that can be super-important is if somebody's coming into it that never worked with all those tools, and if they just start flying the camera around, it becomes weightless and massless, and you get these impossible angles.
What I have is a top A-camera operator—a camera operator who's also a Steadicam operator, and he works with me. So I do the gestural camera. He says, "What was that?" And I say, "Well, that was a technocrane, and I was pushing out with the stick and then I rose up and I came over here." And then he'll refine the movement I did—smooth out the base-move but free up the rotation, so it can still have a little bit of a wheels feel.

F: Well, we took it a step further in this case, with limitations, because we felt we had to make a case for (The Lion King) feeling live action, which was the goal. Because if it felt animated, it would have been redundant with the (Disney animated feature) from '94. So with a title that still holds up like that, you have to create something that feels like a completely new medium. And so part of the way we accomplish that is not just saying, "Apply the limitations of the crane," but actually having a dolly grip working, or a grip working the crane arm, and somebody else working the pickle and somebody else working the wheels.

C: So many liberties are taken with CG lighting. But to really get it right, you need a first-unit, top DP to come in. Maybe they don't know anything about the damn CG lighting tools, but lighting is lighting. A backlight's a backlight. And where it gets interesting is you're going to backlight a person, you put a 10K Xenon behind them and they block it. Every tiny little hair on the side of their neck is going to light up and become an outline.
But if your CG model doesn't have those hairs—they're not going to behave the same way. That lighting gag is not going to work.

F: What's helpful now is that game engines that are built for gaming and drive VR are getting to the point where there's enough code in there to emulate all of these analog aesthetics that you would see when you were really moving a light around. So to some extent, we had it on Lion King—that was Unity. On The Mandalorian, the stage set you visited me on—we ended up using an Epic Game engine (Unreal Engine, Epic's realtime rendering engine) with these new Nvidia video cards for gaming. There's enough processing power to get all of these lighting effects so that you're actually seeing [it] in camera. You wouldn't have to explain to a DP, you could actually have him see it, so when I move the light, it reacts like they do in the real world.

C: But you still have to light the CG environment that you're going to project on your LED display…

F: Well, so this is the next step, which is like, "Can we do it actually in camera and not have to comp it with green?" Because when we first built the walls, I thought I was just going to solve problems that we had on Jungle Book, which is it takes so long to…

C: To move the blue screen.

F: You've got to move the green screen or blue screen every time you move the camera. And then you've got to relight the blue screen and then you've got to hide the lights. And then you've got to hide the lights that are lighting the blue screen.

C: Yeah. What's smart about what you were doing with the big LED array is you were doing like a garbage matte around the character. That's where your green was. And I can't remember how you were generating that, but it was almost like a big soft shadow of the character. And that's only where the green was.

F: Yeah, so you've got a halo. So you'd have no green…

C: …and it moves with the character.

F: It moves with the camera. And then you're able to create a skydome. And if you have the environment ahead of time, you get all the interactive light that was so convincing with what Alfonso (Cuarón) did with Gravity with those little video panels. By having a whole video environment, often we never even had to hang a light. It was like shooting outside.
But in the beginning, we weren't really going to get in-camera shots that required no rendering at all. And no comping. I think up to 70 percent of the shots [on The Mandalorian] we were able to get in camera because we were dealing with hard surfaces. It was sci-fi, it wasn't a jungle, which would have been much harder. And we shot anamorphic, which has got more falloff.

C: Right. But you also had separation of the characters from the background. They weren't interacting. They weren't running through the bush on a full shot.

F: But if you remember from Jungle Book when you visited, using the virtual system, we would plan what the shots were and figure out what we were going to interact with.

C: Like the Techvis.

F: Like the Techvis, much like you were doing capture on Avatar, which is kind of where I learned it. And so we would do full motion capture, first on Jungle Book, and then we would build little cookie-cutter sets that would integrate into the green-screen extension of the set. And so we learned how to provide for what you need to interact with, which isn't that different from what they did in The Wizard of Oz. They would figure out where the yellow brick road was and how much they needed to actually build.
And so even on what's a "practical film shoot," like a Marvel movie, there's still a green screen back there. So you have to plan ahead of time what you're going to interact with and where the light's coming from. Since I wrote most of [The Mandalorian], I wrote it to fit within our volume, and in trying to keep the scale of it like the first Star Wars film, which was a relatively low-budget affair, even though the effects were spectacular. And so we tried to bring that aesthetic into this because it actually paired up quite well with these techniques.

C: That's fantastic. It's been a while for me being on a live-action set. I'm going to be doing live action in May in New Zealand. And it's a lot of hard-surface sets, but we have some jungle stuff as well. And we're mixing capture, or captured performances, with live action. And so the biggest thing I've found was problematic when we did it before was eyelines. If you're capping CG against CG, you can always adjust the eyelines slightly in animation later. But if your eyeline gets baked in when you have a photographic element, you've got to be hitting that right point…

F: What was interesting (on The Jungle Book), especially working with a young actor, we ended up with puppeteers, because a kid would look at a puppet and not at the face of the person reading. And then the puppeteer could also improvise and change things and adjust.

C: Puppets are great; I love puppets.

F: So for that application with one character, it's good. But if you have a group of characters looking, you have to have everybody look at the same place. And I know they've been trying to mess around that with Simulcam or having two different sets.

C: We did a lot of capture. We've got another problem too, which is we've got characters in different scales that are played by humans that are the same size. So I've got humans and now the Avatars. So we've got two adjacent sets in the same volume. On one I'll have a kid who's supposed to be a human kid, but he's acting to his Na'vi friend, a Na'vi teenager.

F: Simultaneously?

C: Simultaneously. So we've got a puppeteer who's got the Na'vi form, and he does the arms and the hand touches and all that, and so we're capping him against the other guy over there. They've gotten so good that we actually were able to have it sync perfectly. And we used little people.

F: For the other set.

C: For the human scale relative to the other performers. And it worked beautifully. And we have a stock troupe of actors that worked with us for about 16 months.
I think the eyelines and giving something to the actors is still the most important thing. You know, the tech can solve any problem, but it can't solve an emotional problem for the actor.

Jon Favreau, right, with his Iron Man, Robert Downey Jr. (Photo: Photofest)

Creating a Comfort Zone for Actors

C: I think the more technical the film is, the more important it is for the director to get all the tech out of the way; to sit down on an apple box with the actors and talk to them about the emotional throughline of the scene, where they are in the story and where they are in the world. And every actor is different. So figuring out how to key into each one of them psychologically, especially when you're going into a big scene—somebody dies, somebody's born, somebody falls in love—there's always a lot of trepidation on the part of the actors. And if the tech starts to get in the way and slow things down and become cumbersome, there's a moment where everything needs to just go away and you just need to be with them.

F: As an actor, what I learned is most useful to me—because a lot of times [shooting is] out of sequence or certainly it's not in real time because you're filming it over a month, two months or whatever it is—I always found that the best direction for me, at least as a starting point, is giving me my frame of reference, reminding me what happened in the scene before. If we didn't shoot it yet, reminding me what we discussed over dinner or in rehearsal. To calibrate me so that I know what energy to bring in so that I'm fitting into the context. And so often my direction, when I talk to an actor, will be, "OK, remember this scene?" It's like tuning their guitar.

C: It's the emotional continuity. And we just spent the last year and a half capturing Avatar 2 and Avatar 3, two close-to-three-hour movies, at the same time. So on Tuesday, we do a scene from Avatar 3. And on Wednesday, we do a scene from Avatar 2. And there were similar scenes in both movies. They weren't straight repeats but they were touchstones in movie 3 to movie 2, and there were some similar moments. And, man, I'd get confused, I'd have to do my homework ahead of time. So all right, this is where we're coming out, this is what we did, this is what it resonates with before. This is what we're setting up for after.
But you can't act what you haven't experienced yet, so blank that out, you know. And I just sit with the cast and just not be in a hurry no matter how much pressure there is. Let them kind of stumble into it, start to feel it.
I've found my cast particularly really loves what I call a two-fer or even a three-fer, where it's just like, "All right, let's just reset, do the whole scene again." And they just love that because once you get to that emotional peak, you may be too primed for the early part of the scene. But man, you're right there in the sweet spot when you get to the back part of the scene. To me, that's where all the really fun discoveries happen. And that's true in live-action work or mo-cap. To me, that doesn't change. That's just a human processing thing between actors. But I love that part. To me, that's as much fun as the visual splendor.

F: I agree. It does require a certain amount of compartmentalization to be able to deconstruct the filmmaking process.

C: When actors are coming in to perform capture for the first time, they have a certain trepidation that it's a highly technical experience that will work against them. And I say, "Look, if I'm on a live-action set, I'm thinking about you. But I'm also thinking about the light, I'm thinking about the lighting cue that's got to happen for that, I'm thinking about the extras in the background, that waiter is coming in late, he's going to ruin your best take…"

F: Yeah.

C: And I'm looking at 50 other things. What if I can take all of that out of my consciousness and just focus on what you're doing? And we did it.
You know, Sam (Worthington) and Zoe (Saldana), they'd come over; they'd be glued to that reference shot. They want to make sure the emotion is there. They want to make sure that they stuck the landing in the scene. And everybody all grouped around the reference camera monitor.

(Photo: Kevin Scanlon)

"A lot of people think there's humanist filmmaking and then there's technological filmmaking. I maintain it's all one thing. It should always be one thing."
—James Cameron

Directors' Teams in the Tech Age

C: I put a lot more responsibility on my ADs now. I used to just kind of charge in and do everything and make it all happen. I don't do that anymore. I really rely on my AD team to precook everything and get it all set up and do drive-arounds.
Maria Battle-Campbell has worked with me now so long she'll even do rehearsals, if it's more of a straightforward thing. I don't mean performance, but if we're doing some kind of straightforward action thing, she'll actually rehearse. And I'll say, "Call me when you've got it working," because there are sometimes glitches or alignment issues and all that. So I let her kind of do that and it really requires such a deep knowledge. I don't think anybody knows more about this specific type of thing we're doing on Avatar, with the underwater stuff and all the motion-capture stuff, than Maria.

F: I came into this learning from being on sets [as] an actor mostly. From the beginning, the AD was an important position for me, and UPM, because they're there to help guide you in areas you don't have experience in. And so you seek out people who are much more experienced than you to help you look like you know what you're doing, whether it's little pointers like change your socks at lunch or the big things like understanding the interface between each department and making sure everybody's getting to do their job properly, and everybody's communicating well and keeping the set safe and in a good spirit.
And right now, my 1st AD, Kim Richards, was a trainee on Jungle Book who then got into the Guild and, because she had knowledge of the workflow of motion-capture volume, had been on our whatever you want to call what we did on Jungle Book, hybrid live action, but on a volume so you're still calling the cadence out on motion capture, which is different because now you're not just interfacing with a crew but you're also interfacing with a whole technical staff.

C: The brain bar…

F: The brain bar's there.
And to emulate a real live set with Lion King, it wasn't just the people on the stage but it's also the communication. And often, I don't know if you've noticed it, but in tech situations, it tends to be a little bit more democratized.
You usually have a lot of smart people who are problem-solving and bringing solutions to the table. But when you're in the rhythm of a film, there is an efficiency that is associated with the way that people have been making films for a long time. And you want to fold that creativity and that tech culture into the film culture.
And the AD becomes the pivotal part. And so we had a full AD team even though there were no actors on set and there was no background.
But you need people who are comfortable with both worlds.
I think a big part of why we were able to preserve all of that stuff was because you had the AD team in there being the glue between the different departments and those different worlds. Just like I drew upon the talent pool that you helped develop through Avatar when I was doing Jungle Book, now you have people who can go out in the world and share this knowledge…

C: … And then it morphs, because every film is different, and every production environment is its own culture. It morphs and then you look over the fence and see how somebody else is doing it, and you say, "Oh, that's similar, but that's very different. I can use that."

F: They can communicate what they're learning from other sets. And it's not that we don't want to share that information, because you've been very generous with helping me understand and learn this process.
Whereas, if I'm working with somebody, a producer who's worked with you, a UPM, an AD, that person can help communicate and steer [me] through and avoid the traps.

C: Yeah, I think that's how the knowledge propagates, right? And I think it's not critical for the director to know every nuance of how the system works as long as they still know, "This is how my camera works and this is how I communicate with the actors and this is how we get to see the display." You don't need to know the 10,000 things that are happening in the backfield, but the ADs do.

F: And you can still keep that organizational structure even though the job description changes. Because technology always offers [both] opportunity and a threat to what came before it. And if you could assimilate the new information into the old culture, then you're not disposing of the richness of experience that can inform.

Jon Favreau and James Cameron (Photo: Kevin Scanlon)

The Learning Curve and the Spark of Provocation

C: The thing that amazes me about your work, your career, is that you started as an actor and a writer. And you just absorbed all this stuff as part of your craft. But the tech will never seduce you to the dark side. [laughter]

F: Well, I was more of a Luddite in the beginning. All the early ones I did: With Elf (2003) it was in stop motion. It was forced perspective. There's a little bit of CG in the end for the reindeer. When I did a movie called Zathura (2005), I didn't want to do that with CG…

C: With Stan…

F: We did it with Stan Winston, that's right. We had animatronics, we had motion-control spaceships and miniatures and models, but I always felt everybody was doing CG then. First of all with Iron Man, also with Stan, some shots were practical suits, some shots were CG. And it was hard surfaces, reflective surfaces. And when I was starting to give notes on the practical shots, I knew it crossed the line where I could no longer tell the difference. And so I was convinced. But I still hung on to wanting to keep as much practical as I could.

C: A lot of people think there's humanist filmmaking and then there's technological filmmaking. I maintain it's all one thing. It should always be one thing.

F: I was very much drafting behind you on Jungle Book because that Pace (Pace/Cameron Fusion Camera System) system, all the gear, and the MotionBuilder, all that stuff…

C: Well, I shot native 3D…

F: Native 3D, using your camera system, MotionBuilder. We were using Photon also, but we were in a volume, and we were essentially using the path that you had created. And we were refining it for our particular application because we were doing photography and marrying it much like you're doing now with that technology. But you look at George Lucas, and few people appreciate how many things he came up with that we now couldn't be making movies without, [such as] CGI, EditDroid…

C: Multi-axis motion control. All of it—it was revolutionary.

F: And so we need those people to help humanize technology, because film has always been a marriage between technology and old mythic storytelling. And both of you, I felt, took that seriously in your own different styles.

C: Well, I was inspired by George. He had a 10-year start on me, something like that. And Steven (Spielberg). They were the guys that I was admiring and studying.

F: What's interesting about Steven is he'll sometimes go completely old school and then sometimes push the envelope with Jurassic Park or Ready Player One. He loves tech.

C: I think it's less that he loves it as much as he likes to create a provocation to the tech people to rise to some new challenge. I think he is less interested in actually figuring it out than saying, "This is what I want—here's what I want you guys to do," and pushing them to a new level.

F: He's definitely like the old tradition of a classic director [who asks], "How do I tell the story?" Whereas in looking at your body of work, often you're mulling over a technology and then the stories are inspired by them, like with Titanic.

C: Literally. Avatar resulted from me being one of the founders of Digital Domain, and seeing where CG was going, and wanting to not only ride the wave but actually help build the wave. And this was in '95, we were a year and a half into Digital Domain, and we had come out of the blocks hot with digital comp, because everybody else was still doing opticals. There was about an 18-month window, or a two-year window, where we were doing digital comps and everybody else was still looking around going, "Hey, I don't need my optical printer anymore," which was an interesting moment. But I thought, "We've got to stay ahead, and how do we do that?"
So I wrote Avatar to push Digital Domain. And I literally said, "All right, what can we do?" I was trying to create the grand provocation to get them to the next level. And everybody said, "No, that's out of reach, we can't do that yet." But we always knew we wanted to do human emotional capture, photoreal. And it started there. And the thing with Stan (Winston) was, that was our dream. And he saw that we were going to make this transition from prosthetic makeup to essentially…

F: …Digital prosthetics. Stan was very important to me because, again, he was another one who was passing that knowledge he had developed with you and Steven. And so when working with him, especially on Iron Man, [there was] always a child-like enthusiasm. And the thing is, I think the reason why Jurassic Park still holds up is because there really aren't as many [CG] shots in there as you think. I think there's like 60 shots.

C: Yeah, that sounds about right. We only had 42 on T2, and that was like a year and a half earlier.

F: Well, but you would, as did Steven, marry a practical thing that kept the CG honest. I assumed all the liquid [metal on the T-1000 villain]… I didn't realize you were shooting real mercury to get the liquefied metal.

C: You wouldn't do that now.

F: And even at the time, you just assumed some of that stuff was CG. But what it does is it grounds the CG, and I think that a virtual camera does that as well—having something for reference in the frame. But now you've evolved to the point where you can create full worlds. I think it helps, honestly, that Pandora has a look to it, a stylized look when you're doing photoreal without any reference. You'd better go out and shoot a lot of photography.

C: We shoot tons and tons of reference throughout the rainforest.
Like in the first film, when Zoe Saldana's character is drinking from the flower, that was me with a rubber plant leaf in Hawaii, pouring a bottle of Evian down the leaf and she was drinking it. And we got the lighting the way we wanted.

F: Well, that's what also has been very helpful with the key frame animation, because with a movie like Lion King, to transpose the human performance onto a photoreal animal, there's so much interpretation with the animators that we just went completely key frame. But when we do key frame, instead of doing capture, because we don't need all that data, we create a black box theater and have the performer…

C: You're filming…

F: With a video camera—no tracking…

C: You film the gesture…

F: It's to give the key frame artist the reference for performance, which I think is a lot of what's valuable about motion capture: you're allowing the actor to make decisions about eyeline and status, when they hesitate, when they perform. If you just have a voice track, a lot of that is determined by an animator, and often animators come from a different tradition.

C: Sometimes it's the mistakes or the hesitations or the pause between the lines that's important. If all you're doing is just taking those audio clips out, you're missing what's happening in the pause.

F: And I can have them interacting. So in our volume, we swept out all the tech and we just had video cameras, and we had people in their street clothes. We were on long lenses, so there was no camera in their space. They would overlap each other, they would improvise, and so we could give it again that human [dimension], just like with the cameras…

C: The messiness… That was the first thing you responded to when I was having Jake [Sam Worthington] running through the jungle, and you saw a rough cut of the scene, and you said you liked the messy camera. Because I was intentionally trying to overshoot and catching him and all that sort of thing. And just mussing his hair as much as possible.

James Cameron with actor Sam Worthington in a motion capture suit during the making of Avatar. (Photo: Photofest)

Motion Capture, 3D and Assembling the Proper Rig

F: I don't know if they're up to the quality of full capture, but there are suits now [with which] you're getting great capture without all that expensive technology. And on Jungle Book, we used all consumer-facing tech. We used Oculus in addition to HTC HMDs. We used all the stuff, all the VR gear you can buy off the shelf. Because I was noticing that the latency on the motion capture setup, the camera, it was dropping more frames than VR was.
But if you take a real camera operator who comes out of the brick-and-mortar world of filmmaking, and then you stick them in there, no matter how much you tell them to look at the Simulcam monitor, they always look at the green-screen monitor because it's not dropping frames. It's always a little rubbery. And even though the Simulcam's wonderful for framing, camera operators aren't deciding things with their prefrontal cortex. They're reacting instinctively to an image that they're looking at through the viewfinder. And so if you take that away from them, they're not really operating the way that they normally would on a set.

C: Simulcam is an amazing, amazing tool. And it's evolved a lot. I operate everything myself—I'm talking about on live action—except for Steadicam. Because I just know there are people who do it better.

F: You go single camera? You're always operating?

C: For 3D, I think two cameras is about the max. And I almost never use the second camera.

F: And what are you using, the 65…

C: A Venice. It's the new high-level Sony HDR camera. They're claiming 17 stops of dynamic range, it's pretty stunning.

F: How big of a chip?

C: Sony did something special for us where they unmounted the sensor block, so that we can mount the sensor to the lens essentially. Check this out: a 29-pound system, in handheld mode, it's tethered, 11-axis motion control system, built into a 29-pound camera with two HDR Venice cameras on it.
DPs really torpedoed native 3D, I think. And they did it because they said, "No, we can't use that cheesy ENG Fujinon lens that weighs 2½ pounds. We've got to use the big 18-pound Zeiss lens. Or a Panavision zoom lens. And so you put two 18-pound lenses on there, you're at 36 pounds before you've included the rig or the cameras or anything else. Now, how are you going to do a handheld shot?
So we did Avatar and we won an Academy Award for cinematography with two-thirds-inch sensors. And that was 10 years ago. But the point was the lenses were small. So two 3-pound lenses, that's not breaking the bank. You start putting in two 10-pound lenses or 15-pound lenses and you're screwed. So I've just been waiting for the pixel count to get high enough and the sensors to get small enough to get back down to those small lenses.

F: Now you're a bit of a purist—do you ever just go single camera and then convert for a challenging shot?

C: Yeah, you have to. I'll tell you where I did my greatest compromise to date on 3D: We set out to shoot the whole exhibition when I went down to the Challenger Deep in native 3D. We wound up shooting about half of it. And converting half. And it was fine.
I'm still a firm believer in native 3D photography. I would never convert a movie that I was making now. I converted Titanic; I converted T2. We had a year to do the conversion; it was done with loving care, and there was good volume in the faces, and it was really nicely done. But typically when you have a finished cut of a movie…two months before delivery, there's no time. It's always a rush job.

F: I found also, because we were trying to figure out on Jungle Book if we should, and the first argument, of course, was it's always easier to convert [because], well, all the backgrounds are digital anyway. It's just the kid… the hardest thing to get right is a close-up of a child with hair.

(Photo: Kevin Scanlon)

"With a title that still holds up like (The Lion King), you have to create something that feels like a completely new medium."
—Jon Favreau

Placing Human Emotion Front and Center

F: That's all that really matters. Anything in the background you can fake. But it's a human form—that's always been [where I] prioritize my energy. If your brain has a lot of real estate dedicated to something, that's going to be the hard part to fool. So the brain is really good at physics and facial recognition.

C: And facial interpretation…

F: Of emotion.

C: So is it somebody that's going to kill me? Is it somebody that's going to give me a present or feed me? Or…

F: Your survival when you're an infant is based on all of that. So it's very hard to fool those things, which is why it's easier to do a lion than the human face.

C: Sure. Creatures are relatively straightforward.

F: And so that made me not scared of Jungle Book, because on Jungle Book I knew that I had the kid there to cut to. And in Lion King, I can't cut to the kid. And so even though you may have had whatever it was, 1,600 shots in Jungle Book, where you are looking on that screen was generally something that was real.

C: Yeah. It was tangible, and the lighting set the lighting for the scene.

F: Now there's nowhere to hide, so I'm in your world again. Which is hard; there was a learning curve. And I felt like with Jungle Book, when I was done with that, I finally knew how to do it. [But] I didn't know how to do it when I started. You learn on the job. And then at that point, that's when I sought out Lion King because I was ready…

C: I shadowed you that day. And you went into a big session with the animators and the creature designers, and you were so meticulous and specific about the different faces of the different animals. There was the panther, the tiger, and it was all approached from character. I thought it was just really interesting to watch how you did it. And it really reminded me a lot of what I go through on a daily basis, dealing with "What does everybody look like, and what does it mean?" Every change you make means something. And I think a lot of people look at directors who do imaginative stuff, like you're sitting on some high-res version of a movie that already exists in your mind, and you're just stumbling to try to explain it.
But that's not what it is. There's an out-of-focus picture, and when you see something that looks like the picture in your head, but it brings it more into focus, you say, "Do that." That to me is selection.

F: Whereas you were very iterative when I've visited you on the set. That's the other good thing about acting—you get to be on other people's sets and really apprentice quietly, silently in the background, you pick up on things. And watching Marty (Scorsese on The Wolf of Wall Street), he much more has an instinct, tries it, will do lots of takes, lets it evolve.

C: There are a lot of downsides to performance-capture work because it's tedious, it's very detailed. But the upside is you get the performance and it can be take 4 or take 6 or whatever. Once you have that performance, now you start worrying about lighting, now you start worrying about photography. And you can blow your move five times, but you haven't hurt the performance.
I've found that, interestingly enough, actors, when they're free in a performance-capture space, tend to move around laterally a lot more.
And it's true behavior vs. camera-specific behavior.

F: But I contend that we are not trying to emulate real life. We're trying to emulate a real movie, at least for now. Because you remember we used to add grain all the time with old TV shows when they went from film to digital.

C: Yeah.

F: They went to the Genesis (digital camera), it would add film grain.

C: You know, I find the hardest aspect of directing, when you're finishing a film, is to always be able to react to the cut like you're seeing it for the first time. And I've found that the AVID actually worked against us as a community.

F: Because you see it… it's too easy to try things.

C: When you had to physically splice films, you didn't make tiny changes to the cut of a scene 100 times. You kind of just cut it, you know?

F: Well, audiences are very sophisticated now. I feel like the lighting has gotten better, the editing's gotten so precise. And people can metabolize a lot more cuts and a lot more information.

C: But it has an anxiety-producing effect sometimes, or a distancing effect. I admire when I see a filmmaker who knows how to slow down and smell the coffee, and then speed up and go like hell.
Close-up discipline is another thing. When everything's a close-up, nothing's a close-up. And I know—you do that well. You really get way out and then you come into a nice medium that's reading the whole group…

F: I think push-ins, also. You look at The Godfather and I think there's two push-ins in that whole film.

C: Yeah. And they're overused.

F: Because it's so easy to make something feel compelling. And then just like a close-up, you do it too much… it's like seasoning. If you hold back and save it when you finally go to it, it creates a visceral reaction. And I think what's important—and I know we both agree on this—is that all of this technology is purely a way to access the humanity in the audience.
And because, ultimately, we are looking for some sort of connection. Because we're such a communal species, that we're looking to either learn from one another or experience something vicariously through somebody doing something. So you can learn on a really subconscious level by seeing somebody go through some sort of trial and persevering and being rewarded for that.

C: I always think it's a simulation—a big emotional simulator. And we get to project ourselves in a kind of dreamlike way. I'm walking in the shoes of the characters, and it's like a dipstick to check my value system or my emotional plumbing. Would I do what that guy did?
I'll cry in a movie before I'll cry in a similar situation in real life. Which I think is interesting. In real life, my defenses are up. I'm supposed to be the strong guy. I'm supposed to be the guy that everybody relies on. In a movie theater, I can cry in a way that I can't in real life, which sounds backwards but it's very real. So these emotional simulations are very, very powerful for people.

F: I think it's a good way to put it. Part of the job of the filmmaker is to support that immersion. And often it relates to how you make choices that you believe the audience would make. So, if you're cutting a scene properly, it's almost like you don't need VR because you're looking where they want to look anyway.

C: It's that predictive algorithm that I think you have to create in your own mind as a filmmaker; I need to be able to predict where the audience is looking, and to really make the cuts flow like music.

F: It's interesting because then you get a guy like George Miller, who keeps everything dead center while he's doing his fast cuts. And so the effect is like you're on a roller coaster.

C: I never noticed that, but in Fury Road or Mad Max

F: …He keeps it in the cross hairs. That's why he can get away with all the things, all the manipulation that I couldn't get away with. He'll throw so much information at you and the effect is like you're strapped in the backseat of a car going 150 miles an hour. And you're getting overloaded, but to an effect that he wants you to have. It's not arbitrary and it's not the whole movie.

C: And he'll slow down, and he'll do static set pieces and then use the wide screen, the scope frame, which he loves—very Sergio Leone kind of compositions.
The act out for the end of the first act in Fury Road is probably one of the best peak dropoffs, stop-dead-cold moments I've ever seen in a movie. It's just like, "Can this get any crazier?" Then everything stops.

Jon Favreau with lead actor Neel Sethi while filming The Jungle Book. (Photo: Photofest)

Conspicuous vs. Inconscpicuous Directing Styles

F: I like when directors have a style, like I love the Coen brothers. I love that they're a character in the movie.
But for me, personally, I get embarrassed because I didn't come up as a director. And I feel like I have to be in service of the story and take myself out of it. I don't like when somebody says, "That's a great shot."

C: Yes, exactly.

F: That means they noticed the shot. However, when I'm watching a Scorsese film, I love a great shot. I appreciate it in others. I think it's like music or cooking. Everybody's got a different style.

C: It's like Bruce Lee called his way "the way of no way." It's kind of a style of no style. It's getting out of the way of the movie. It's the earnest narrative. I'm too earnest. I wear my heart on my sleeve as a filmmaker. It's just all out there, folks. I'm not trying to interpose, you know, some kind of pure camera [technique] or stylistic tropes. I just want you to feel like you're in the moment and not taken out.

F: Now one exception is for comedy. I come from comedy, so I will let you see what the camera's doing if it's funny.

C: When somebody falls out of frame, the camera doesn't go with them, and then they get back up.

F: Exactly.

C: Yeah, that's a stylized moment that aids comedy, and the only time I've ever done stylized stuff like that was on True Lies—the only time I ever took a swing at comedy with some success. There were some funny moments. It didn't always work, but the camera was more self-conscious.
But if you're creating another world in another time, using Avatar as an extreme example, it's a completely made-up world that we want you to utterly believe. So if you're doing a movie that takes place in an apartment or in a city or whatever, you can do a shot, you know, like (David) Fincher will do, you go through the keyhole and into the stove. Because you're taking the mundane and elevating it cinematically.

F: That's right.

C: But if you're starting with something extraordinary, you actually try to do the opposite, which is to have the mundane artifacts of photography come in and remind you that this is supposed to have been shot by a camera. With Avatar, I wanted people to feel like we went to Pandora. And we ran around in the jungle, using handheld cameras and we actually shot the stuff that we saw. And so lens flares or little pop zooms, almost borderline documentary.

F: The problem is most people notice bad CG. The best CG is CG you don't even know you saw. And that's happening more and more with filmmakers who are becoming more and more sophisticated. So as I go out to the theater now, I no longer have the aversion to CG in its execution.

James Cameron with DP Al Giddings shooting The Abyss in a 7.5 million-gallon tank. (Photo: Evrett)

The DGA's Support of New Technology

F: I like to see what people are coming up with because it's fascinating to me and for the aesthetic of the innovation. But to be relevant to a wider audience, you've got to pair it with filmmakers. And I think for new filmmakers, the sooner you can find a way to have an understanding of things that others might not, you give yourself some market position by familiarizing yourself with technology.
The DGA does Digital Day, which I think is great to introduce new filmmakers to new technologies. And the tech innovators are looking for people to take what they've innovated and pair it up with storytellers. Because unless you see somebody use it to tell a story, it's not going to be relevant to everybody.
And the Guild's been really helpful also for us because a lot of the workflow of these new productions means that, like, when I'm working on Mandalorian, I want to make those filmmakers partners with me early on. So as I'm trying to teach them about the technology, I'm also trying to find out what their vision is for each episode. And I want to partner with them early on and then have them come back later after we've built all the resources around their vision, and then make that available to them later. It's nice that the Guild is open to evolving and understanding that so they can accommodate the new techniques and still remain relevant going into the next phase of our industry.

C: First of all, the DGA does a great job. Digital Day is just one example of helping with this kind of knowledge dissemination and mentoring. We talk about passing on knowledge, but what's happening is the early adopters, the young directors coming in, they're cutting their teeth on stuff that we had to learn after we already knew how to direct a movie.

F: I was a native to AVID in my first time directing.

C: I still have the scars from the guillotine splicer. [laughter] People who come from photo chemistry and come from the physical world of cameras and lighting need to teach people who've never done that, and what the constraints are. Because those constraints, those kind of real-world artifacts [need to be created artificially]—because there's no lens and no lens flare unless you want to put one there; there's no camera vibration unless you decide to put it there; there's no raindrops on the lens; there's no air-water interface that you get from a splash housing if there's no splash housing.
If you've never been out in the world physically shooting and you're coming up in a CG environment as an animator or a shot developer or virtual lighter, I think we need to not just teach directors the effects. You need to teach the effects people real-world photography and those motifs so they can do their job better.

F: And the director is the one who has the opportunity to create that culture. And I think also, as technologies arise, it's always easier to try to create a new model. And I think it's important for us as directors, not only for the Guild, to figure out how to be relevant in that world. But we also have to impress upon the people that we're working with how important the Guild is.
Things like creative rights have to evolve over time. Just because the nature of filmmaking has changed—[and] the nature of our relationship creatively to that end product—the spirit of that has to remain. And we have to make sure that those we work with understand that creative rights are important.

Terminology Decoded

⇒ Fusion Camera System: Created by James Cameron and DP Vince Pace, a dual-camera 3D motion control system that allows for finely calibrated stereo lens movements that mimic human vision.

⇒ MotionBuilder: Real-time 3D character animation software that also allows for camera and light creation.

⇒ Multi-Axis Motion Control: Automated systems that allow for precise, repeated camera movements along two or more dimensions.

⇒ Native 3D: Refers to the process of shooting with 3D cameras during production, rather than converting a film into 3D during post.

⇒ Techvis: Technical previsualization, or a practical visual analysis of how an effects-heavy sequence could be filmed that accounts for issues like cameras and lenses, constructed sets and planned VFX.

Two prominent directors from shared experience discuss craft, approach, collaboration and the state of the business, with DGA Quarterly as a fly on the wall.
More from this issue
The latest DGA Quarterly includes an exclusive conversation between James Cameron and Jon Favreau, the DGA Interview featuring Anthony and Joe Russo, directors Patty Jenkins, Colin Trevorrow and Rupert Wyatt discussing VFX in tentpole films, and more!