In the summer of 2019, audiences packed movie theaters to see the African savanna—and not a single frame of what they saw existed in the physical world.
In March 2017, Favreau sent a small group of key leads from the future The Lion King—including production designer James Chinlund and VFX supervisor Adam Valdez—on a two-week trip to Kenya. They scanned rocks, shot panoramas, collected textures of bark, earth, grass. Mount Kenya and Kilimanjaro stood before them, and the task was purely an engineering one: return home with enough data to later recreate 58 square miles of digital landscape on a computer. Every rock, every acacia tree, every gradient in the sky of the finished film was the result of that expedition, filtered through months of work by artists, modelers, and programmers. The irony was that real Africa was needed precisely in order to make an unreal one.
The film’s production stage was in Playa Vista, a Los Angeles neighborhood, and it looked less like a movie studio than a virtual-reality lab. The soundstage was outfitted with an OptiTrack tracking system—dozens of infrared cameras tracked people’s positions in space. Participants put on HTC Vive headsets and found themselves inside a digital model of the African landscape. Favreau could walk across virtual Pride Rock, look down, assess the light—and do all of it while standing in an empty soundstage the size of a tennis court. The result of this process was a film released on July 19, 2019, which critics, marketers, and even Disney itself often called “live-action.” Not one living animal, not one actor in the frame, not one built set—and yet “live-action.”
The film’s director of photography was Caleb Deschanel—a man nominated for an Oscar six times for work with physical cameras, with film, with light reflecting off real surfaces. He was invited to shoot a film in which the camera did not exist as a physical object. It was like asking a violinist to perform a concerto on a synthesizer with no body, no strings—but an infinite number of settings.
In conventional animation, the camera is an abstraction. The animator specifies a viewpoint, an angle, a movement path, and the software calculates the image. No limitations: the camera can fly through a wall, hover unsupported in midair, instantly jump a kilometer away. Favreau and his team deliberately gave up that freedom. They wanted the virtual camera to behave like a physical one—with all its imperfections. If the camera is on a crane, it moves in an arc, and you feel that arc. If the operator holds the camera on their shoulder, it sways slightly, and the viewer unconsciously reads that as the presence of a human body. If the lens is telephoto, the background blurs and compresses; if it is wide-angle, space stretches and perspective lines become more aggressive.
One thing about this setup really got me: the team reproduced even the limitations of exposure. In real filming, if the camera is pointed at the sun, the rest of the frame falls into shadow—the sensor cannot capture the entire brightness range. In animation, it is easy to “fix” that, to even out the light. But The Lion King’s creators kept the blown highlights and crushed shadows, because without them the image lost that elusive texture of authenticity the viewer’s eye recognizes instantly.
The virtual production was handled by several teams working in tandem. The main VFX and animation vendor was MPC (Moving Picture Company), which employed 130 animators from 30 countries. The production’s VFX supervisor was Rob Legato—a man whose career has been built at the intersection of practical and digital filmmaking. Animation supervisor Andy Jones coordinated the movement of every character. Set supervisor Audrey Ferrara was responsible for the digital landscapes. And the pipeline—that conveyor belt along which data moved between stages of production—was built around the Unity game engine.
Choosing Unity was no obvious decision. A game engine can render a scene in real time, albeit with simplified graphics. That allowed Favreau and Deschanel to walk through the virtual savanna and see it immediately, without waiting for a render. Roughly speaking, they got a draft of the shot instantly, while final quality came later, once the scene went through the heavy artillery—Maya for modeling and animation, Houdini for particle simulation (dust, water, fire), Katana for lighting setup, Nuke for compositing, and RenderMan for final rendering. Each of these packages handled its own task, and coordinating them required engineering discipline comparable to managing a complex factory operation.
MPC’s in-house tools are a story of their own. An internal muscle-simulation system moved the flesh beneath the animals’ skin: when a lion turned its head, the neck muscles contracted, the skin tightened, the folds redistributed themselves. Without that layer, even a perfectly animated character would have looked like a rubber toy. And the procedural grooming tool Furtility generated fur—millions of individual hairs, each responding to wind, moisture, and contact with other surfaces. The name seems to be a blend of “fur” and “fertility,” and it is an exact metaphor: the tool literally “grew” fur on digital bodies.
This brings us to the question that haunted the film from its very first trailer. Viewers saw photorealistic animals and asked: “Are they real?” Then, once they learned they were not, they asked: “Then why?”
The paradox is that the closer a digital image gets to reality, the more harshly it is judged. The 1994 animated film could afford exaggerated emotions on its characters’ faces—Simba frowned, Scar smirked, Pumbaa stared wide-eyed in fear. Real lions do not do that. And when Favreau’s team created a lion indistinguishable from a real one, that lion inherited the limitations of the real thing: its face could not express subtle human emotion, because a lion’s face was not built for it. The animators worked in a narrow corridor between anatomical truth and emotional legibility, and many critics felt that corridor was too narrow.
But at the level of the environment—landscapes, light, atmosphere—photorealism worked without qualification. The artists studied how African morning light differs from midday light: the angle, the color temperature, the length of shadows, the amount of diffuse light in the air. The dust kicked up by antelope hooves was not drawn by hand—it was physically simulated, particle by particle, in Houdini. Water at the watering hole refracted light. Grass bent in the wind not uniformly, but according to the height and stiffness of each blade. Together, all of it created the sense of a place that existed independently of the camera—and that sense was the film’s chief achievement, even if the characters’ faces remained contentious.
The 2019 The Lion King did not use motion capture. The actors voiced the characters in a recording studio, but they did not put on sensor suits and pretend to be lions. Donald Glover (Simba), Beyoncé (Nala), James Earl Jones (Mufasa) worked with their voices. The animals’ movements were created from scratch by animators, drawing on video references of real animals and their own understanding of anatomy. That is a fundamental difference from, say, Planet of the Apes, where Andy Serkis performed Caesar in a motion-capture suit.
The film was not the first project to use VR in filmmaking either—similar experiments had been conducted before, including by Favreau himself on The Jungle Book (2016). But the scale of virtual production in The Lion King was different: if The Jungle Book combined a live actor—the boy Neel Sethi—with a digital environment, here the physical world was excluded entirely. The only shot in the film filmed in reality is the opening image of the rising sun, and even that remains a point of debate among VFX specialists.
What surprises me is that a film with a budget comparable to the biggest blockbusters was shot on a soundstage where almost nothing physically happened. People in headsets walked through empty space, pointed at invisible objects, and discussed the composition of a shot that existed nowhere but on the screens inside their helmets. From the outside, it probably looked absurd. From within, it was a film set—simply turned inside out.
The question that made this text worth writing: if the camera is a digital object and light is a set of parameters, what remains of the cinematographer’s profession?
Caleb Deschanel spent decades learning to feel light. He knew how the sun behaves at sunset, how it reflects off sand, how contrast changes on an overcast day. None of that knowledge lost its value in a virtual environment—it became the language in which he spoke to the engineers. When Deschanel said, “I need backlight with a slight halo around the mane,” the technical specialists translated that into rendering parameters. His eye remained the instrument of judgment—does the shot look right, is it convincing, does it evoke the needed emotion? But the hands that once turned the focus ring were now free.
Rob Legato, the production’s VFX supervisor, served as a bridge between cinematographic thinking and digital execution. His task was to make sure the virtual camera did not do things a physical one could not. It is a counterintuitive restriction: why forbid the camera from doing what is technically possible? Because the viewer, even without realizing it, feels the difference. When the camera glides smoothly through a herd of antelope half a meter above the ground—it is beautiful, but the brain whispers: “That does not happen.” When the camera is mounted on a virtual jeep and jolts over virtual bumps—the brain relaxes, because it knows that angle from BBC documentaries.
In essence, Favreau’s team created a system of voluntary constraints, imitating the imperfections of the real world. Virtual lenses had focal length and depth of field. A virtual crane moved in an arc, like a real one. A virtual Steadicam system produced that distinctive floating motion familiar to anyone who has seen Kubrick’s The Shining. This was not work with technology, but with the viewer’s memory—with their unconscious catalog of “this is what real footage looks like.”
The film grossed more than a billion and a half dollars worldwide. The Academy nominated it for an Oscar for Best Visual Effects—the statuette went that year to the war drama 1917. The arguments over whether it is “animated” or “live-action” have not died down to this day—Disney entered it in the live-action category, and the Academy did not object, though the film was created entirely on computers.
But the technological question is more interesting than the commercial one. The Lion King demonstrated that cinematography can exist without a physical camera. Deschanel did not become an animator—he remained a cinematographer, because his work lay in choosing: where to look, from what distance, from what height, in what light. Those decisions do not depend on whether the lens is made of glass or code. But the tool changed radically, and the boundary between “shooting” and “building” in this film has been erased so thoroughly that perhaps we need a third verb.
It seems to me that the most accurate description of what happened on the Playa Vista stage is neither “filming” nor “animation,” but something like navigation. People moved through a world that simultaneously existed and did not exist, made decisions based on decades of experience working with physical light, and recorded the result. After finishing the film, Favreau carried the principles they had developed into The Mandalorian, where virtual production took on another form—with LED screens and the Unreal Engine instead of VR headsets. The technology did not stop at Pride Rock; it moved on, and the soundstage in Playa Vista turned out to be only a way station on a route whose final stop no one can yet see.