What you need
Accurate room mapping requires a supported iPhone Pro with LiDAR and iOS 26 or later. Results can vary with lighting, reflective surfaces, occlusion and how much of the room the camera can see.
Visera comparison — measurement vs generation
LiDAR-measured vs AI-generated paint visualisers.
Both promise to show you a colour before you buy the tin. They get there in genuinely different ways, and that difference shows up exactly where it matters: light, texture, sheen and the openings in your wall. Here's a plain comparison, including where an AI preview is still a reasonable tool.
Download Visera on the App StoreThe core difference
One measures your room. The other generates a picture of one.
A LiDAR-measured visualiser, like Visera, scans your room with Apple ARKit and the LiDAR scanner on iPhone Pro first, building a real mesh of its walls, floor, ceiling and openings — then renders paint colour live into the camera feed on that measured surface. An AI-generated visualiser starts from a single photo and uses an image model or a segmentation mask to guess which pixels are wall, then repaints those pixels or generates a new image altogether. One is working from a measurement of your actual room. The other is working from a description of it.
Side by side
Where the two approaches diverge
| What it starts from | LiDAR-measured (Visera) | AI-generated |
|---|---|---|
| Wall geometry | Measured directly by LiDAR as a 3D mesh, then classified by ARKit. | Inferred from pixels in a photo by a model's guess at "wall". |
| Light | Your room's real light, live through the camera, at the moment you look. | Whatever light the model imagined when it generated the image. |
| Wall texture | Plaster, brick or panelling stays visible under the new colour. | Often flattened or smoothed away by the repaint or the generation. |
| Sheen (flat / eggshell / gloss) | Renders on the measured surface under real light and reflection. | Tends to render every sheen the same, since no physical surface is simulated. |
| Windows and doors | Classified and excluded from the paintable mesh automatically. | Guessed at from image edges; more prone to painting over openings. |
| What it needs | A supported iPhone Pro with LiDAR and a short scan of the room. | Any phone and a single photo — no scan required. |
| Best used for | Testing how a specific colour will actually look on your specific wall. | Quick mood or palette exploration before you're near the room. |
Why it matters
Why this matters for an actual paint decision
Paint colour is not just a hex code — the same colour can look markedly different under morning light versus a lamp at night, on a textured wall versus a smooth one, and in flat versus gloss finish. A visualiser that isn't built from a measurement of your room can't show any of that, because it was never given your room's light or surface to begin with; it's rendering onto a description of a wall, not your wall. That's why a colour that looked right in an AI-generated preview can still look wrong once it's actually on the plaster.
In fairness
When an AI-generated preview is still useful
An AI-generated visualiser is a reasonable tool for the stage before you're standing in the room at all — exploring a colour family, a mood or a palette direction from a photo on your phone, or on a device without a LiDAR scanner. The distinction this page is making isn't that AI-generated previews are worthless; it's that they're an inspiration tool, not a measurement of your actual wall, and treating one as the other is where the disappointment happens.
Important details
What LiDAR measurement requires
What it doesn't replace
Even a LiDAR-measured preview is a visualisation, not a certified colour match. Confirm a colour on a real sample before you buy the tin.
Questions
LiDAR vs AI — questions people ask
Which is more accurate, LiDAR or AI paint visualisation?
A LiDAR-measured visualiser is more faithful to your specific room, because it renders colour into your live camera feed on walls it has actually measured, under your room's real light. An AI-generated visualiser produces a new image that resembles your room but isn't built from a measurement of it.
Why do AI paint visualisers get light and texture wrong?
Most work from a single static photo and use an image model or a segmentation mask to repaint pixels it identifies as wall. The model isn't simulating a physical surface under real light, so it tends to flatten texture, invent or ignore shadows, and render flat, eggshell and gloss paint identically.
Is LiDAR slower to use than an AI paint app?
It takes a short scan of the room first, whereas an AI app can work from one photo. That extra step is what buys the accuracy: once the room is mapped, you can try as many colours as you like live, with paint automatically stopping at windows and doors. See how the scan works.
Can an app use both AI and LiDAR together?
In principle, yes — AI techniques can assist with tasks like colour matching from a photo, while LiDAR handles the room geometry. Visera's core paint and furniture preview is LiDAR-measured rather than AI-generated; that's the deliberate distinction this page is about.
Which should I use to test a paint colour before buying?
If you want to see how a colour will actually look on your wall, in your room's own light and texture, a LiDAR-measured visualiser like Visera is the closer approximation. An AI-generated preview is a reasonable starting point for exploring a mood or palette before you're in the room at all.
Keep reading