Neural Photorealistic Video
An imagined future camera that makes video look exactly like life. The render is nearly clean: no grain, no shake, light compression and gentle sharpening.
A decade that hasn't happened yet, so we made it up, carefully
Every other position on the DECADE dial is history. This one is a guess, and we'd rather say so plainly. Nobody knows what a pocket camera will do in 2035. The fourteen decades before it are built from formats, films and sensors you could hold in your hand; the 2030s are built from a direction of travel.
That direction is clear enough to sketch. For thirty years each decade's camera has hidden more of its own limits. Tape noise went with digital tape. Grain went with sensors. Then the 2020s phone began merging many frames into one, so noise, blown skies and dark shadows faded too. Follow that line forward and you reach a camera with almost no visible fault at all: no noise, no shake, no rolling shutter, every highlight held, every shadow open. Our 2030 specification imagines exactly that, plus some features that already exist in early form today and could become routine: reconstructing depth, simulating a lens after the fact, relighting a scene once it's shot, fusing several cameras into one picture.
There is a second line, and it's just as old. Every time pictures got cleaner, people went looking for the old faults: retro filter apps in 2009, film presets in 2012, CCD compacts bought second-hand in the 2020s. It seems a fair bet that a future camera will offer to rebuild any past look on request, much as this app does. So the 2030 dial includes a neural film emulation and an "authenticity-first" capture that leaves the picture alone.
The details are deliberately modest. We gave the imagined camera the same building blocks as the real ones before it: a format, a color recipe, a cadence, a sound chain. Nothing on this page needs a breakthrough to be true, and nothing on it is promised to arrive.
Treat all of this as a story, not a forecast. The names on the dial are ideas, not products, and none of them is a prediction of fact.
Here the honest answer is: less than you might expect. Most of what makes the imagined 2030s special (total HDR, depth reconstruction, relighting, multi-camera fusion, synthetic optics, high frame rates) is part of the specification but isn't rendered by the app. What you get is the visible end of the line: a picture with no grain, no added shake, very light compression and gentle sharpening. Several entries look almost the same as your iPhone's own image. That is the joke, and also the point. The cleanest decade is the one with the least to show.
A few TECHNOLOGY entries do differ. Depth-Aware Capture adds a soft blur toward the top and bottom of the frame; there's no depth map, so it doesn't trace your subject. Computational HDR Next Gen raises contrast a little. Naturalistic Future Capture drops the sharpening and calms the color. Future Social Video adds a touch more compression. The specification gives most entries 60 frames a second and Adaptive Motion Capture 120; your file is 24 fps whatever you choose.
In STILL, most styles add a faint smoothing from noise reduction. Neural Film Emulation is the one with real texture: grain, a soft glow around highlights, dark corners and a matte fade. Adaptive Optical Bokeh and Computational Optical Rendering add the same top-and-bottom blur. Multispectral True Colour is the most saturated.
With SOUND on PERIOD, the voice passes through the full band, 20 Hz to 20 kHz, with no hiss, no motor and no wobble. In practice that means your audio comes back almost unchanged, and SILENT is true silence.
Go back to the 2020s, where these ideas start, or all the way to the 1890s. For how the looks are built, see how it works.
Twelve cameras on the TECHNOLOGY dial in CINE mode. Each frame below is the same shot, rendered by the app at DEGREE 7.
An imagined future camera that makes video look exactly like life. The render is nearly clean: no grain, no shake, light compression and gentle sharpening.
Imagined total HDR. The app doesn't reproduce HDR; the render is the clean future look with a little more contrast.
Imagined noise-free night video. The render is clean and very slightly softer; the night processing itself is not reproduced.
Imagined stabilization that predicts your movement. The render adds no shake and no rolling shutter of its own; it does not steady your hand.
An imagined camera that captures depth with every frame. The app has no depth map; the render adds a soft blur toward the top and bottom of the frame.
Imagined relighting after the shot. Relighting isn't rendered; you get the clean, near-untouched future look.
An imagined camera that changes frame rate with the action, up to 120 a second. Your file is 24 fps; the render is the clean future look.
An imagined camera that simulates any lens in software. Simulated optics aren't rendered; you get the clean future look.
Imagined social video of the next decade. The render is clean with slightly more compression.
An imagined future capture that resists processing. The render drops the sharpening and calms the color: the most neutral entry on the dial.
Imagined fusion of every camera on the phone into one. Fusion isn't rendered; the picture stays clean and noise-free.
Imagined exposure that never misjudges a scene. The render is the clean future look, with no exposure drift.
Twelve styles on the STYLE dial in STILL mode, with the print's own shape and border.

An imagined future photo that looks exactly like life. The render is almost untouched, with faint smoothing from noise reduction and gentle sharpening.

Imagined total HDR stills. HDR isn't reproduced; the render is the clean future photo with a little more contrast.

Imagined true-color night vision. The render is clean, very slightly soft, with smoothed fine texture.

Imagined optical blur rebuilt from depth. The app has no depth map; the render blurs toward the top and bottom of the frame.

Imagined color reconstructed from many wavelengths. The render is the most saturated of the decade, with slightly lower contrast.

An imagined camera that renders any lens in software. The render adds shallow focus toward the top and bottom and a faint smoothing.

Imagined relighting after the shot. Relighting isn't rendered; you get the clean, lightly smoothed future photo.

Imagined social photo of the next decade. The render adds a little compression and more sharpening; the skin smoothing in the spec isn't rendered.

A future camera's idea of film. The render adds grain, a soft glow around highlights, dark corners and a matte fade.

An imagined capture that refuses to beautify. The render has light smoothing, no sharpening and calmer color.

Imagined fusion of every camera into one photo. Fusion isn't rendered; the picture stays clean and lightly smoothed.

Imagined detail invented by neural upscaling. The render is clean with strong noise-reduction smoothing and some sharpening; the print is capped at 12 MP.
A computer voice reading one line, passed through this decade's chain by the app's own audio code (PERIOD, DEGREE 7), with the medium's noise bed. Headphones help.
Future Computational Audio. With SOUND on PERIOD, your audio passes through this decade's chain:
| Band | 20 Hz – 20000 Hz |
|---|---|
| Recorded | mono, like everything the app records |
| Noise bed | none |
| Wow · flutter | 0% · 0% |
SILENT here is true silence. ORIGINAL leaves your iPhone's audio untouched. How period sound works
The camera opens on a different decade every day, with one of these facts under the dial.