Clair Obscur Expedition 33

Look at a frame of Clair Obscur: Expedition 33 next to a frame of any recent photo-realistic AAA game (pick one: the latest Assassin's Creed, the latest Call of Duty, Black Myth: Wukong, Cyberpunk 2077 with all sliders maxed). Both frames are doing roughly the same thing, in technical terms: a real-time 3D scene running at thirty or sixty frames per second on consumer hardware. The frames look almost nothing alike. The photo-realistic frame is trying to convince the eye it could be a photograph of an actual place. The Clair Obscur frame is trying, deliberately, to look like something else: a late-nineteenth-century French oil painting that has agreed to be entered.

The standard cultural-criticism vocabulary for what is happening here is "painterly." This is the word reviewers use when a game looks deliberately like a painting. The word is correct as far as it goes and entirely useless for explaining why the Clair Obscur frame produces a different feeling in the body than the photo-realistic frame does. The feeling is not subtle. Players who have spent time in both registers can identify which is which without conscious effort. The difference registers below the level the casual vocabulary can name.

The frame this essay wants to give the reader: there is a measurable, neurologically specific reason why painted-looking games feel different from photo-realistic games. The reason has very little to do with subjective aesthetic preference. It has to do with which parts of the visual system the rendering choice activates and which parts it lets go quiet. Once the reader has this frame, almost every recent stylized indie (Cuphead, Hades, Pentiment, Disco Elysium, Children of the Sun) becomes legible as a specific perceptual proposition rather than as a stylistic decoration. The wave of stylized indie design over the past decade is not, as the casual reading often suggests, a regression to retro aesthetics under indie budget constraints. The wave is, in significant part, a small and largely unconscious recovery of perceptual work the medium had spent thirty years engineering away.

The work being recovered has a name. The visual neuroscientist Margaret Livingstone, in Vision and Art: The Biology of Seeing (2002, expanded 2014), spent most of her career documenting it. Livingstone's central observation, developed across decades of imaging studies at Harvard Medical School, was that the human visual system is not a single uniform processor of images. It is at least two distinct systems working in parallel: one that processes fine detail and color (the parvocellular pathway), and one that processes luminance contrast, motion, and overall spatial relationships (the magnocellular pathway). The two systems hand off information to each other constantly. They do not, however, respond to the same visual cues. The parvocellular pathway notices the small specific features; the magnocellular pathway notices the broad luminance structure and the motion. A face viewed clearly across a table activates mostly the parvocellular pathway. The same face glimpsed in peripheral vision in the rain activates mostly the magnocellular.

What Livingstone discovered over the course of many studies is that painters across the European tradition had, by patient empirical experimentation, learned to compose images that engage the two pathways in specific ratios that the photograph does not. The painter who organizes a composition with attention to the relative luminance of the depicted elements (rather than to their photographic accuracy) can produce visual effects the photograph cannot reproduce. The trembling quality of light in many Impressionist canvases, the specific atmospheric depth in Vermeer, the strange psychological flatness of late Vuillard interiors, are all functions of luminance-contrast choices the painters made that activate the magnocellular pathway in ways the photograph does not.

The photograph has its own visual virtues. It is a precise record of light reflecting off surfaces at a specific moment from a specific position. Photography activates the parvocellular pathway powerfully. The contemporary 3D rendering pipeline has been built, across roughly thirty years of consumer-hardware progress, to converge on the photograph's specific visual signature. Photo-realism in games is the result. The eye looks at a contemporary photo-realistic frame and the parvocellular pathway lights up: this is a detailed scene, the texture work is precise, the materials are differentiated, the lighting model is correctly handling the way real light moves across real surfaces. The magnocellular pathway, on the same frame, has substantially less to do. The luminance structure is, in most cases, accurate but not particularly meaningful. The composition is photographic, meaning the painter's specific luminance-organizing tricks have not been applied.

This is the part of the analysis that the casual conversation about game aesthetics has been almost entirely missing. The photo-realistic game frame is doing something to the visual system, and the something is not neutral. It is engaging one pathway hard and letting another go quiet. The audience that has spent decades playing photo-realistic games has been having a specific visual-cortex experience repeated many thousands of times, and the specific experience has been the parvocellular-heavy, magnocellular-light experience. The audience's visual system has, in a small measurable way, adapted to it. The adaptation is not a problem in itself. The adaptation is, however, what makes the eye respond as it does when a game like Clair Obscur arrives and starts asking the magnocellular pathway to do work it has not been asked to do in some time.

This is also, on a careful read, what most contemporary players are actually responding to when they describe a stylized game as "feeling different" without being able to explain why. The feeling is the magnocellular pathway noticing that it has work to do. The feeling is, in the strict neurological sense, the visual cortex waking up parts of itself that the surrounding visual diet has been letting atrophy. The phenomenology of the experience is roughly: the image feels alive in a way photo-realistic images mostly do not. The image rewards being looked at. The eye lingers. The body is, for the duration of the looking, present in a way the body has not been present at the photo-realistic frame, even when the photo-realistic frame is technically more impressive on every measurable axis.

Sandfall Interactive, the Montpellier studio that released Expedition 33 in April 2025, built a game whose every visual choice is in service of activating the magnocellular work. The composition of the game's settings is organized around luminance structure rather than around photographic accuracy. Surfaces have been textured with attention to the way a painter would handle the relationship between adjacent regions of the canvas, not the way a camera would handle the same relationship. Lighting is treated as compositional choice rather than as physical simulation. Shadows are placed where the composition wants them, not where the engine's lighting model would have put them. The result is a 3D real-time game that, viewed frame by frame, contains substantially more of the visual-information distribution found in late-nineteenth-century French painting than the surrounding category of contemporary 3D games contains.

The studio did not, on the public record, sit down with Livingstone's book before designing the game. The studio sat down with the paintings. Henri Le Sidaner, Édouard Vuillard, Pierre Bonnard, Odilon Redon, and the broader Belle Epoque French tradition are visible in the game's compositions to anyone who has spent time with the period. The studio is doing what the painters did: organizing images around luminance and atmosphere rather than around accurate light-transport. The neurological consequence is the consequence Livingstone documented. The studio arrived at the consequence by the route the painters themselves arrived at it, which is patient looking at how the eye responds to different compositional choices.

This is the proposition the broader stylized-indie wave has been quietly delivering for over a decade now, almost always without naming it. Cuphead is the same proposition in 1930s animation register. Hades is the same proposition in classical figurative-illustration register. Pentiment is the same proposition in fifteenth-century manuscript-illumination register. Disco Elysium is the same proposition in mid-century social-realist painting register. Children of the Sun is the same proposition pushed all the way toward graphic-design poster register. The choice of visual tradition varies; the underlying perceptual choice is the same. Each of these games has decided to engage the magnocellular pathway with specific compositional work the photo-real default does not require. The audiences that have loved these games have, on the available evidence, been responding to the same thing in each case, even when no one in the conversation has been able to name what the thing is.

The neuroaesthetics field, founded loosely by the British neurologist Semir Zeki at University College London in the mid-1990s, has been documenting this proposition with increasing precision over the past three decades. Zeki's work, gathered most accessibly in Inner Vision (1999), made the foundational argument that the brain's visual system has built-in preferences for specific compositional and luminance relationships, and that the artists across cultural traditions who have produced durably popular work have, in effect, been discovering those preferences empirically and exploiting them. The painter does not need to know the neuroscience. The painter needs to know what the eye responds to. The painter learns this by spending many hours looking at how viewers respond to specific compositional choices. The result is a slow accumulation, across the artist's career and across the tradition's centuries, of compositional rules that the modern neuroscience can now name in terms of which pathways they activate.

Game design has begun to do this in the last fifteen years. The early stylized indies (the original Limbo, Braid, Sword & Sworcery) were, in retrospect, the small market signals that the magnocellular work was missed by the audiences who had been playing photo-realistic games for years. The signals were initially dismissed as nostalgia for old visual styles or as compensation for indie-budget constraints. Both readings are partly correct and substantially wrong. The visual styles those games used were not chosen because they were old or because the studios could not afford photo-real production. They were chosen because they did perceptual work the photo-real production had stopped doing. The audiences responded. The wave grew. Twelve years later, Expedition 33 is the most accomplished commercial entry in the wave's painterly subset, and the wave has stopped being an indie phenomenon.

The implication for the next decade of the medium is straightforward, even if the surrounding commercial conversation has not absorbed it. The photo-real arms race is, on the available perceptual evidence, hitting a point of diminishing returns. Each new generation of consumer hardware adds polygon counts, texture resolution, lighting fidelity, and frame-rate that the parvocellular pathway notices and the magnocellular pathway does not. The audience continues to buy the hardware. The audience also continues to gravitate, in steadily larger numbers, toward the stylized games that engage the visual cortex more completely. The two are not in commercial conflict yet. Both kinds of game are profitable. The longer arc, if Livingstone and Zeki are right about how the visual system works, points toward a medium in which photo-realism continues to be the default for certain categories (sports, military shooters, simulation), and stylized rendering becomes the default for the categories where the visual experience is part of the product the audience is paying for.

The frame the reader should walk away with is, in plainest form: when a game looks deliberately stylized, the style is usually not decoration. The style is doing perceptual work the photo-realistic default has stopped doing, and the work registers in the body as the feeling the casual conversation has been calling "atmospheric" or "vibey" or "painterly" without quite being able to say why. The next time a stylized game arrives and feels different in a way the reader cannot immediately explain, the reason is probably Livingstone's. The magnocellular pathway has been given something to do.

There is one further observation worth registering, because the Sandfall studio's specific narrative choice deserves to be named. The game's antagonist, called the Paintress, paints a number on a monolith each year, and everyone above that age dies. The expedition's task is to oppose the Paintress. The fiction is, in its strange way, the same proposition the game's visual choices have been making: the painter who composes the world has substantial power over the people inside the composition, and the people inside the composition have a small but real ability to push back. The expedition is, in the game's narrative, doing what the audience is doing in the visual register. Both are noticing that the composition has been made by someone, and both are arguing with the composition the maker has chosen.

That is the small accomplishment the game is doing beyond its surface. The game is not just a beautiful-looking RPG. The game is one of the better-realized contemporary arguments that the perceptual register the photo-real default has been suppressing is still available, still wanted by audiences who have rarely been told why, and still capable of producing the specific aesthetic experience the broader cultural conversation about games has not yet learned to name.

The reader who finishes this essay can go back to any of the stylized indies of the past decade and notice, with the new vocabulary, what the design was doing. The reader can also pick up a Bonnard or a Vuillard reproduction and notice, looking at it, the same thing the game is doing in real-time 3D. The eye that has been trained by the photo-real default to skim has, in the looking, briefly slowed down and started to do the work the painted image is asking for. That slowing-down is the gift. It travels. The next stylized game will activate it. So will, with a little luck, the next painting the reader walks past in a museum hallway.