The scene wasn’t working. It was a moment from the Pixar film Coco, still in production at the time—the part when the family of Miguel, the main character, finds out he’s been hiding a guitar. It takes place at twilight or just after, a pink-and-purple-tinged time of day everywhere, but even more so in fictional Pixarian Mexico. And Danielle Feinberg, the photography director in charge of lighting the movie, didn’t like it. She pressed Pause with a frown.
Lighting a computer-rendered Pixar movie isn’t like lighting a film with real actors and real sets. The software Pixar uses creates virtual sets and virtual illumination, just 1s and 0s, constrained only by the physics they’re programmed with. Lights, pixels, action. Real-world cameras and lenses have chromatic aberration, sensitivities or insensitivities to specific wavelengths of light, and ultimately limits to the colors they can sense and convey—their gamut. But at Pixar the virtual cameras can see an infinitude of light and color. The only real limit is the screen that will display the final product. And it probably won’t surprise you to hear that the Pixarians are pushing those limits too.
Of course the people at Pixar still have to make all the choices that’ll produce the final outcome. To prepare, Feinberg had gone on multiple trips with the team to Mexico, taking lots of pictures and notes on the lighting and colors she saw there. And even though this critical moment in Miguel’s house looked lovely, it didn’t look right. But it was awfully late to realize it. “We had finished lighting. We were at the point where we were going to show it to the director,” Feinberg says. “And I asked the lighter to put a green fluorescent light in the kitchen.”
It was an unusual request. In the conventional chromatic grammar of today’s motion pictures, greenish-tinged fluorescence usually means a movie is about to turn eerie, even ominous. But Feinberg wanted to see the kinds of lights she remembered from the warm, homey kitchens they’d seen in Mexico. “I wasn’t sure the director was going to be happy with me putting green fluorescent light in the background,” Feinberg says. “It was a little bit of a risk.”
But after seeing the light, the director, Lee Unkrich, agreed. It looked like Mexico, he said. He remembered those lights and the resulting mood from their travels too. The green glow, which usually had one narrative meaning, assumed another.
In a way, every filmmaker is really just playing with moving light and color on surfaces. That’s the whole ball game, a filmic given. But Pixar takes it further, or perhaps just does it more self-consciously and systematically. Its emotionally weighty, computer-generated animated films deploy precisely calibrated color and light to convey narrative and emotion—from the near-total absence of green in WALL-E (until postapocalyptic robots find the last plant on Earth) to the luminous orange marigolds that symbolize Miguel’s trip to the magical Land of the Dead in Coco through the contrast between the cool blue luminosity of the afterlife with the warm, snuggly sepia of New York City in last year’s Soul.
In fact, almost every Pixar movie works within a specific color palette, a story-specific gamut that filmmakers like Feinberg pull from and use to plan the look of each scene, a road map known as the color script. But Coco complicated that process. When its story moves to the Land of the Dead, it cranks up all the dials, colorwise. Those scenes look made out of neon, like a bio-organic version of Tokyo’s Shinjuku District at night. “When it came time to do the color script, it was like, ‘The Land of the Dead has every color. All of it takes place at night, so we can’t use time of day to elicit emotion. There is no weather in the Land of the Dead, so we can’t use weather to elicit emotion.’ Those are three pretty typical things we use to support the story,” Feinberg says.
Using color to express emotion is a hallmark of life. (Humans aren’t even the only animals to send signals with a bit of sexy red or dangerous green.) But the mechanical production of color has defined and changed human cultures since before recorded history. The technology for making colored things and the science of how those colors work in the world and in our minds changes and evolves, transforming culture along with it. Right now, that technology is evolving again.
If talking about music is, as someone once said, like dancing about architecture, then talking about color is like doing a trapeze act in zero-g on a space station. But here goes: First of all, you have to forget the dorm-room philosophizing about whether you see the same red that I do even though we both call it “red,” man. If we both agree—and let’s agree to agree—that “red” is light with a wavelength of somewhere above 620 nanometers, well, waves of what, exactly? (It’s fluctuations in electrical and magnetic fields, as if that helps.) Or we could agree that “red” light is made of subatomic particles called photons, the irreducible quanta of energy—1.8 electron volts, to be more or less exact.
Go ahead and map those electron volts and nanometers for red, plus the ones for all the other colors you can name, into a straight line, or even wrap them into a circle as the physicist Isaac Newton did. You still won’t be capturing everything that comes together to mean a color. The real map needs more dimensions than that. It needs the amount of color, from pastel to saturated. It needs the amount of light you’re talking about. That’s “luminance,” or sometimes “intensity.” Color that’s made of light is different from color that’s light bouncing off a surface, changed not only by how that light reflects or refracts but also by whether the surface is colored itself, maybe by a pigment. Map all those values together, usually in three dimensions, and try to match the objective numbers to the vagaries of the way human color vision works—we see yellow as brighter than other colors, even if the actual brightness is equal, and that’s just the beginning of the headaches—and you have what’s called a color space.
At the movies? Whoof. Even more complicated. The pictures you see on a screen are made of light shining through a colored strip or generated by a digital device, projected outward onto a reflective surface and then bouncing into your eyeballs. (And what happens once it gets in there, where biochemical photoreceptors transduce photons into neuroelectrical signals, is a whole other thing.)
The point is, “color” means a lot of different things, depending on how you’re using it. And using it has been a defining trait of humanity since we all first started thinking. We see colors in the world, in nature, and we use what we see and learn to make newly colored things. It’s a hallmark of human activity, practice, and culture. We started by collecting objects with colors, turned to grinding rocks into powders and pastes and smearing them onto cave walls and on our bodies—and have arguably reached an evolutionary acme with the ability to control and create light with the precision and fidelity of a Pixar.
None of that highfalutin philosophy would help Danielle Feinberg, though. Her team had a job to do. With too many colors in play and too wide a gamut to narrow, she couldn’t use specific colors to code for emotions. So Feinberg’s team did it with varying amounts of light—with luminance.
Take the scene where the old ghost Chicharrón dies unremembered in the Land of the Dead. It’s a tear-jerking sequence, but the color palette is still just as wide (though it does lean hard into moonlit blue for this moment). Instead of taking away color, the scene is actually just less bright, lit not by the virtual neon or glowing-orange cempasúchil flowers but by just a couple of lanterns. “That was the way we had to do it on Coco,” Feinberg says, “just because it was a colorful, lively world, but we still needed to elicit that emotion.”
Control the lighting, control the colors, control the feelings. That’s filmmaking. As of this writing, Pixar’s last 23 movies—going back to 1995’s Toy Story—have made a combined $14 billion globally, and that’s not even adjusting for inflation. Kids like them; adults like them. Even in a locked-down, movie-theater-free world, the latest Pixar movie, Soul, grossed $117 million worldwide.
But I’ll tell you a secret: When it comes to wringing emotion from color, Pixar cheats.
In a very special screening room at Pixar’s Emeryville, California, headquarters is a very special screen. It’s not huge, perhaps just 10 feet across, and it’s at the front of a room dominated by a huge control panel studded with five smaller computer monitors and at least two keyboards. The ceiling is covered in felt, and the carpet squares are black instead of the gray that’s standard at Pixar, to keep light contamination to a minimum.
Explaining what comes next requires me to deliver some bad news. Remember the primary colors you learned in elementary school? Red, blue, and yellow, right? So, yeah, that’s wrong. You were supposed to be able to mix those into all the other colors, but that never worked, did it? Blue and yellow were supposed to be green, but you got brown. Red and blue were supposed to make purple, but you got … brown.
That’s partially because subtractive colors reflect some wavelengths of light and absorb others. Mix them together and you absorb more and reflect less. Things get dark. Unless you carefully manage the pigments and the mixing, and you start with the primaries cyan, magenta, yellow, and black—the CMYK beloved of magazine designers.
It’s also wrong because oftentimes people confuse light streaming from a source like a TV or a star with the color that happens when light hits a surface. Those primary-school primaries aren’t the only possible primaries. But even Newton was a little confused about this. His primaries are the specific basic colors he identified in the spectrum he projected from a window onto a wall in 1665, holed up at his mom’s house while a pandemic raged back at his university. You can relate, right? Newton broke whitish sunlight into a rainbow’s worth of colors and chose to draw the borders at seven: red, orange, yellow, green, blue, indigo, and violet. He called that a spectrum, but of course that categorization leaves out a lot—the “extraspectral” colors like pink or purple or, yes, brown. (Brown is just dark yellow. Shh.)
If you’re reading this on a screen instead of on paper, you’re seeing a concatenation of light generated by red, green, and blue pixels—a whole other set of primaries, not coincidentally at similar wavelengths to those the color receptors in your eyes are tuned to. A little more or a little less of each, and just as with CMYK pigments (and white light or white paper), you can make just about every color that the human eye can discern. Point is, the colors we see aren’t actually mixed from a list of available ones, like buying from a paint store. It’s a continuum of light and reflection, interpolated by the biological sensors of our eyes and the not-totally-understood think-meat just behind them.
That big screen at Pixar isn’t lit by a typical projector. Instead, mounted in the wall behind us is a custom-built Dolby Cinema projector head. If you’ve been to a theater with a Dolby setup, you were looking at images cast by a projector that was actually a pair of triple-barreled laser guns—red, green, and blue beams capable of combining to produce a range of colors closer to what human vision can perceive than anything else out there. The two guns had wavelengths slightly offset from one another so that special 3D glasses can distinguish them, one lens for each, and your brain can combine them to create the illusion of dimensionality.
But at Pixar, all six beams come from one source, which means this projector has six primary colors. Also, the Dolby rig has a span of brightness, from dark-dark to bright-bright—in screen terms that’s called dynamic range—and the one at Pixar is more than 10 times brighter than one in a civilian-class Dolby Cinema.
Part of how we see color is how much light is behind it, how much overall energy is pumping toward us. So most modern color spaces have an axis that measures this, with black (no light) at one end and white (all the light) at the top.
The standard unit for measuring what’s called luminous intensity, the amount of light coming from a source over a given angle of view, is the candela—as in one candle’s worth. But if you’re talking about “luminosity,” the amount of light emitted by something like a TV screen, what you want is candelas per square meter, also known as a nit. Dolby Cinema output is 108 nits, but Pixar amps it up even more. Sitting at the control panel of the Pixar system, senior scientist Dominic Glynn practically glows with praise. “We’ve goosed this projector with an extra 600 percent laser power. We can get well above a thousand nits on this screen,” he says. “It’s one of the most linear, perfect reference color-grading displays you can conceive of.”
So this projection room is where wide-color-gamut, high-dynamic-range colorcasting abilities merge with Pixar’s creation of virtual sets, each with their own virtual physics of light. People like Glynn can actually generate a world of color wholly unlike the one you and I usually live in. “We could light the whole set with a green laser,” Glynn says. “That’s kind of hard to do in the real world.”
You saw it in Coco, but the movie where it might have made the most difference was Inside Out. That’s the one about personified emotions living in the brain of an 11-year-old girl. When Inside Out was in production, Dolby was working on its in-house version of new standards for high dynamic range.
The range of colors it could convey was bigger. The “gray scale ramp” between darkest black and brightest white would allow a theater equipped with these lasers—only a half dozen initially—to turn its light output so low that the screen becomes a black indistinguishable from the walls (“exit signs notwithstanding,” Glynn says). It was an entirely new standard of color, but Pixar’s directors of photography were already working to expand even that envelope.
The colors a projection system can reproduce are bounded by a triangle-shaped color space—red, green, and blue at the corners, and everything else a mixture of those inside the lines. But that color triangle is invariably smaller than the possible colors of the universe, or even those that the human eye and mind can distinguish. Which leaves a little wiggle room for Pixar. “The specific hues at the red, green, and blue corners of that triangle are not really what you’d experience under, say, ultraviolet illumination,” Glynn says. “We said, ‘Hey, what would happen if we tickled all the portions outside a traditional cinema gamut?’”
Glynn taps on the control panel keyboard and calls up a scene from Inside Out where Joy and Sadness walk into the Realm of the Subconscious. Glynn hits Play; Joy and Sadness enter a dark room and see a forest of giant broccoli, lit from the side so it seems outlined in a bright green. They move to a red staircase headed down into infinity and then meet another character, the clownish imaginary friend Bing Bong, imprisoned in a cage of candy-colored balloons. “These are all basically as saturated a color as one can achieve in digital cinema today,” Glynn says.
Then he cues it up again, in super-high-end digital cinema fireworks, using everything the screen can give us. “They go through the doors, and you see the little long shot of them in the distance, then all of a sudden we kind of have everything.” The shot widens, and the camera heads toward the broccoli forest, but now the broccoli is laser-pointer green, glowing against the blackness.
The red archway around the staircase is the most vivid red I have ever seen, and when Joy and Sadness start walking down the stairs, the edges of the screen disappear. The room, the world, is nothing but black except for the stairs. The balloons of Bing Bong’s prison look unearthly, like a Jeff Koons dog with eldritch powers. “I want to say 60 percent of this frame is outside the gamut of traditional digital cinema,” Glynn says. “We have a version of this film that has been creatively approved and built for exhibition on televisions that don’t exist yet.” You can see them only if you saw Inside Out in a fancy-pants Dolby-equipped theater.
You can’t buy these colors for your house. But Pixar does have a prototype of what that TV might be like. It’s in a room next to the screening room. I convince Glynn to show it to me in action, and when he fires it up to maximum brightness, it’s actually painful to look at. The light leaves an afterimage like one caused by staring at the sun.
Once these technologies are in every movie theater and every living room, maybe even on every phone, things are going to get really weird. They will test the limits of human color perception and maybe even extend them. Poppy Crum, the neuroscientist who runs research at Dolby, has been working on all the ways that, for example, seeing images in really high dynamic ranges can trigger not just autonomic responses—like flushing from heat exposure after just seeing video of flames—but psychological ones too. Crum says her research shows that these tricks of light heighten the entire emotional experience of moviegoing.
The Dolby screen has given Glynn some pretty out-there notions too. He asks if I know how color receptors in the human eye can “bleach,” which is to say that they essentially use up the molecules that absorb specific wavelength-ranges of light and transmit color signals from retina to brain.
I tell him yes. “You’re talking about contrast effects and afterimages,” I say.
“For sure,” Glynn answers.
This quirk of human color vision has vexed scientists since before anyone knew about the color photoreceptors in the eye. Color-thinkers in the 19th century recognized that the same colors—or rather, objects of the same color—might look different depending on context, on what colors they were adjacent to.
They also recognized the obverse—different spectra can appear the same in different contexts. This was one of the tricks that the color-seeing brain could play. Varying levels of brightness change the colors people see. Look away from a bright light, like a candle, and the afterimage you’ll see is the color of that light’s complement on a color wheel. In all those cases the brain seems to be generating colors that aren’t there.
Now, Glynn says, it might be possible to take control of those illusory effects. Blast the middle-wavelength greenish receptor in the eye with light at its peak sensitivity and “you can actually heighten the sensitivity or perceived sensitivity to other colors in complement to that.” It’d be like a laser-powered version of Jasper Johns’ famous painting Flags, where you only see the “correct” colors of the United States flag when you look away, as an afterimage.
So what if, Glynn proposes, a scene in a movie added, subtly, light in a very specific wavelength of green? Then just kept ramping up, more and more green—and, at a key moment, the screen dropped all the green out at once. The movie would induce the complementary color as an afterimage. You’d imagine you were seeing a specific red, not projected on the screen but as a neurophysiological response to stimulus. And if you pick the precise wavelength, “you could actually cause someone to perceive a color that they could never otherwise see. Like, there’s no natural way for you to have the perception of that color.”
That color wouldn’t be onscreen. It wouldn’t be anything a projector could cast or a computer could generate. It’d be a function of pure cognition, different for every viewer, existing only in the mind, then fading to nothingness. Which is true for all colors anyway, when you think about it.
Excerpt from Full Spectrum: How the Science of Color Made Us Modern by Adam Rogers. Copyright © 2021 by Adam Rogers. Available May 18, 2021, from HMH Books & Media.