AI room scenes for furniture from one photo: scale, shadow and catalogue

Furniture is harder to photograph than most products. Moving a corner sofa to a studio, placing it in a living room set and doing it again for every colour option costs time and transport. That is why many furniture makers have just one photo of a product, taken in the workshop or the warehouse.

AI can take that photo and show the same product in a living room, a bedroom or a garden. In furniture, though, two mistakes show at once: the product at the wrong size and the shadow in the wrong place. This guide covers the source photo, scale and shadow checks, materials, and a workflow for catalogues and online shops.

The source photo: how to shoot furniture

A warehouse photo is enough, as long as the product is complete and sharp. The boxes in the background do not matter. The product does.

• Get the whole product in the frame, legs included. A cut-off leg gets invented in the scene.
• Shoot slightly from the side, at roughly a three-quarter angle, so the front and the depth both show.
• Hold the camera close to seat height. A sofa shot from high up looks squat in a scene.
• Use daylight. A warehouse door or a large window is enough. Fluorescent light can turn colours green.
• Take close-ups of the fabric, the wood grain, tufted buttons and metal legs.
• For sets, also shoot each piece on its own. A three-seater, two-seater and armchair can hide each other in one frame.
• Write down the measurements: width, depth, height and seat height in centimetres.

Scale: is the product the right size in the scene

A photo does not tell anyone the real size of a product. Without measurements a two-seater can look like a three-seater in a scene and a small side table like a dining table. Buyers choose furniture to fit their own room, so this is the most expensive mistake.

In every scene compare the product with the fixed things around it:

• Doors and windows. Is the sofa's back height sensible next to the window sill.
• Other furniture. Is the coffee table close to the sofa's seat height, does the bedside table look tiny or huge next to the bed.
• Small objects. Do cushions, books, cups and lamps look natural in size next to the product.
• The floor. Does the size of the rug, parquet or tile pattern make the product look bigger or smaller than it is.

If you are unsure about a scene, state the product's measurements clearly under the image or on the product page. Put the measurements on the page, not inside the image.

Shadow, light and floor

Furniture with the wrong shadow looks as if it is floating. It is the detail that gives away an AI scene fastest.

1. Contact shadow. Under every leg and base, where it touches the floor, there should be a thin dark shadow.
2. Light direction. The shadow should fall away from the window or lamp in the scene. If the window is on the left, the shadow stretches to the right.
3. Reflections. Polished wood, glass, marble and metal should reflect the room, and what they reflect should match the scene.
4. Floor fit. If the legs stand on a rug, the rug should dip slightly, and the perspective should match the floor.

Do this check at full size, not on a thumbnail. A shadow error disappears when the image is small, but a buyer who zooms in on the product page sees it at once.

Materials and workmanship

In furniture the buyer pays for the material. If solid walnut looks flat like chipboard in the image, or bouclé looks shiny like velvet, the product looks cheaper than it is or sets the wrong expectation.

On wood check the direction and tone of the grain, on fabric the size of the weave, on leather the creases and the stitch lines. Count tufted buttons, drawer handles and legs one by one against the original. Use a separate photo for each colour option instead of having one fabric colour turned into another. Woven rugs and finely patterned upholstery are the hardest surfaces, so check them at full size on their own.

Which scene goes where: catalogue, shop and marketplace

The same product needs different images in different places. Most furniture brands use a plain-backdrop image together with a few scenes.

Take extra care with scenes in an export catalogue. Seeing the product in a room that looks like the buyer's own home makes it easier to picture it there. A grand interior that feels foreign to the target market does not do that job.

Google openly accepts furniture staged in a room in product listings. Its lifestyle_image_link page gives "furniture staged within a room" as an example and says such images can be shown in free listings. Its image best practices recommend that the product fills 75% to 90% of the main image and that images are 1500 x 1500 pixels or larger. Images made with generative AI must carry metadata that says so. We could not check other marketplaces' official rules, so we have left them out.

A workflow for catalogues

When you prepare a catalogue of dozens of products with AI, order matters as much as the images. Work in this sequence.

1. List the products: code, name, colour options and measurements.
2. Take one clear photo for each product and each colour.
3. Write a scene plan: which product in which room, in which style. One style across the catalogue holds the pages together.
4. Make the images or have them made, then run the scale, shadow and material checks above.
5. Show the result one last time to the person who knows the products best, usually the production lead.
6. Name the files by product code and keep the AI metadata.

If you want to compare with a shoot, our guide product photo shoot or AI images works out the cost per image.

With Bulsu Labs Studio

Bulsu Labs Studio makes furniture images with AI from one photo per product: a white or plain backdrop, and room or outdoor scenes if you want them. You write the measurements and the scenes you want with your order. Every image is checked against the original product, and one where the colour, pattern or logo changed is generated again at no cost. 50 images cost €100 and 100 images cost €175, paid once, delivered within 3 business days at the latest. A general checklist for the source photo is in our guide to AI product images from one photo.