You do not need an expensive 3D scanner to get a model of a product. A phone with a decent camera, some well-placed light and the right capture method are enough. The principle is photogrammetry: you take many photos of the object from different angles, and 3D processing finds the shared points between them to rebuild the shape and texture. Model quality depends almost entirely on capture quality, so it pays to get the capture right from the start.
What you need
- A phone with a good camera (most recent phones are fine).
- A neutral, non-reflective background: a table, a sheet of card or a cloth with no pattern and no shine.
- Soft, even light: near a window on an overcast day, or two or three diffused lamps. Avoid direct sunlight and flash.
- Optionally, a turntable, manual or motorised.
Pick a method: turntable or walking around
Walking around the object
The object stays still and you move in a circle around it. This suits large items (chairs, tables, sculptures) or anything you cannot rotate. The golden rule: do not move the object relative to its background during capture. If you touch or turn it halfway through, 3D processing can no longer match the photos.
Turntable
The phone stays fixed, ideally on a tripod, and the object rotates on the turntable. It is convenient for small and medium items. Here a plain, uniform background is essential: only the object should change between photos. If shelves, sockets or other fixed features appear behind it, the processing treats them as part of the scene and the result breaks. A white or grey card behind the turntable fixes this.
How many photos, and from which angles
For most products, 30 to 80 photos are enough. How you take them matters more than the count:
- High overlap between photos: each frame should share a large part with the previous one. Take small steps between shots.
- Several heights ("rings"): one full circle close to the object's level, one from roughly 45 degrees and one from higher up to cover the top.
- Fill the frame with the object without cutting it off at the edges.
- Keep the object in focus in every shot. Blurry or shaken photos lower quality; delete them.
- Consistent exposure: if your phone allows, lock exposure and focus so brightness does not vary between frames.
If you also need the underside, you can shoot a second series with the object turned on its side, where the capture method allows it. Otherwise the base usually stays uncovered, which does not matter for most products that sit on a surface.
Video instead of photos
You can also record a slow video around the object, from which frames are extracted. Move slowly and steadily to avoid blurred frames, and cover the same heights as with photos. Video is quicker to shoot; individual photos usually give you more control over sharpness.
Difficult objects and how to prepare them
Photogrammetry needs surfaces with visible texture. Some objects are hard to scan:
| Object type | Problem | What you can do |
|---|---|---|
| Shiny, metallic, mirrored | Reflections change with every angle | Very diffuse light; temporary matte spray if the product allows |
| Transparent (glass, clear plastic) | You see through the object | Temporary matte spray or powder; otherwise choose another method |
| Very thin (wires, strings, leaves) | Too little detail to reconstruct | Closer photos, contrasting background |
| Completely plain, no texture | Hard to find shared points | Contrasting background and soft side lighting |
Always test any coating on a hidden area first and use it only if it can be removed without harming the product.
Common mistakes
- Too few photos, with big steps between them: processing cannot find enough shared points.
- A single ring: the top of the object is missing or distorted.
- Harsh shadows: shadows get baked into the texture and show on the model from every angle.
- Object moved during capture: the photos no longer match each other.
Uploading and processing on 3DView.ai
On the 3D Scan page you follow the capture guide, upload your photos or video, and 3D processing on our servers builds the model. The result is a textured GLB and USDZ model, ready for the web viewer and for augmented reality. For more on formats, read GLB, USDZ and OBJ explained.
If the product is not at hand or you just want a quick idea, you can use AI generation from one photo or a text prompt. Keep in mind: AI is fast for a preview; for the best result, scan the real object. The full comparison is in product photos vs 3D scan vs AI.
After the scan
Check the model in the viewer: shape, textures, any holes. Then add it to your product page with the code from the e-commerce page (see also adding 3D models to Shopify and WooCommerce) or download it. To see how the model looks in a shopper's room, read about augmented reality for online stores, and for the bigger picture, our guide to 3D product visualization. Available plans are on the pricing page.