The complaint is usually phrased badly: “movies don’t look like movies anymore.” That is not quite right. Some of them do. Some old movies looked terrible. Some new ones look gorgeous. The useful complaint is narrower: too many images are visible without being shaped.
That distinction matters because otherwise the argument turns into lazy nostalgia. The movies you saw between 10 and 20 tend to become your permanent definition of what movies look like. Robert Tolppi makes a version of this caveat in his video on “Netflix lighting,” and it is worth keeping in the frame. Nostalgia explains some of the reaction. It does not explain all of it.
Production norms changed.
Tolppi’s useful point is not that digital cameras ruined movies or that LED fixtures are fake light. That version is too easy. Modern sensors capture more color and contrast than older cameras could. LED units are brighter, cooler, lighter, more efficient, easier to rig, easier to hide, and easier to control. The tools got better.
That is where the trouble starts.
Better tools gave productions more room to avoid commitment.
The style of non-commitment
“Netflix lighting” is not a company style guide. It is a production behavior: high-key, low-contrast, front-filled, evenly exposed images where everything is usable and almost nothing is sculpted.
Blaming Netflix alone is too convenient. This shows up anywhere production culture prizes safe flexibility over visual authorship. Flat light is not always ugly. High-key images can be beautiful. The issue is not brightness. The issue is non-commitment.
An image can be correctly exposed and still have no idea attached to it.
Lighting is not just illumination. It is shape. A backlight separates a body from the background. Negative fill gives a face structure. A motivated source tells you how the world works. Contrast creates hierarchy. Shadow controls attention. Darkness is not a defect in the image. Darkness is information.
The point is not to make every frame noir. The point is to decide where the eye should go and what the frame is allowed to withhold.
Flatness is rational
This is where the complaint gets more interesting.
Of course productions drift toward flat lighting. It keeps everything usable. It protects coverage. It helps VFX. It survives HDR delivery, streaming compression, executive notes, uncertain downstream decisions, reshoots, pickups, changing edits, and the client who suddenly notices the background actor in the corner. If the set is broadly filled, you can turn around faster, cover more angles, save more takes, and leave more decisions for post.
That is rational. That is what makes it dangerous.
The problem is not that nobody knows how to make the image more interesting. The problem is that interesting images often require someone to make a decision before every downstream stakeholder has had a chance to be nervous about it.
A shaped image closes doors. If one side of the face is allowed to fall off, that is a decision. If the sun is backlighting the scene, the schedule has to respect the sun. If the key is motivated by a lamp, blocking has to respect the lamp. If the frame is built around shadow, somebody has to defend the shadow when the note comes back asking whether we can see more.
Blanket coverage feels safer because it preserves options.
Protecting options is exactly what can kill images.
Cinematography is partly the art of making irreversible decisions before the movie is fully known. That is uncomfortable under modern production pressure, because the modern machine is built to defer decisions as long as possible. More data. More takes. More notes. More recovery. More flexibility.
Flexibility is useful until it becomes the aesthetic.
Why Wicked is a useful example
Tolppi uses Wicked because fantasy musicals should be the easiest place to justify stylized, committed lighting. Oz is not a documentary location. If there is a genre where color, shape, theatricality, and motivated unreality should have permission, it is this one.
That does not mean Wicked as a whole is bad. This is not a review of the movie, and a trailer clip or selected scene is not a full artistic judgment. The useful point is narrower: when even a huge fantasy production can produce examples that feel broadly filled and ordinary, the issue is not lack of available technology.
The daylight problem Tolppi highlights is a good operational example. Put actors into harsh front sun and the image flattens: squinting faces, flat exposure, no separation. Put the sun behind them, then shape the face with bounce or soft fill, and the frame gets depth.
That is not a camera-spec argument. It is a craft habit. Where is the sun? Where is the key? What side of the face is allowed to fall off? What remains black? What is the viewer supposed to look at?
Craft is not vibes. It is a stack of decisions.
The production-tech pattern
The lighting argument connects to VFX, virtual production, and AI without turning into a complaint about any of them. Virtual production can restore intention by letting directors, DPs, production designers, VFX supervisors, and producers see context earlier. That was the useful point in When Game Engines Meet Hollywood: real-time tools matter when they move decisions upstream, where they are cheaper and more honest.
But the same tool can become another gray-safe machine. A volume can give actors and camera a real environment to respond to, or it can become a dim wraparound wall that encourages soft, general, noncommittal light because that is the easiest thing to keep from breaking.
AI image tools have the same trap. They can generate polished surfaces quickly. They can be useful for reference, planning, mood, iteration, and disposable exploration. But they do not know which constraint should remain. They do not know what should be dark. They do not know which imperfection carries the idea.
Better tools are useful when they collapse the distance between intention and result. They are harmful when they become buffers against intention.
That is also the point of Friction Is About to Become a Luxury. Not all friction is waste. Some friction is the cost of choosing.
Lighting has always understood that. Waiting for the sun is friction. Cutting light with flags is friction. Reblocking around a motivated source is friction. Defending a dark frame in review is friction. Some of that friction is the work.
The craft moat
When tools get better, taste matters more, not less.
That is the through-line from Phil Tippett to Toy Story to NVIDIA. The tool changes. The interface changes. The cost of output changes. The craft moves one layer deeper.
For lighting, that deeper layer is judgment.
The scarce thing is no longer exposure. The camera can see into the shadows. The lights can fill the room. Post can recover, isolate, mask, track, relight, denoise, compress, expand, and rescue more than ever. The machine can preserve more information than the old one could.
So the scarce craft is deciding which information does not belong.
Knowing when not to fill. Knowing what to leave black. Knowing where the viewer should look. Knowing when a face needs softness and when it needs structure. Knowing when the background should disappear. Knowing when a beautiful image is wrong for the scene and when a risky image is the only one with a pulse.
That is the moat. Not owning a better light or a newer camera. Judgment under production pressure.
Let the image choose something
The future should not look flatter because the tools got better.
Better cameras, better lights, virtual production, real-time review, AI-assisted planning, and smarter post tools should make stronger visual choices easier to execute. They should help teams see the decision earlier, defend it better, and carry it through the pipeline with fewer surprises.
If they only make choices easier to postpone, the image loses.
The point is not to make everything darker. The point is not to cosplay old movies or pretend older workflows were morally superior.
The point is to let the image have a point of view.