New in Vizera: Masking — change one detail, not the whole image

The image is done. The presentation is tomorrow. And the client wants a different armchair.

Until now that meant re-rendering — and hoping the rest would stay the same. With the new masking it means: mark the area, describe the change, done.

Masking changes only the area you mark. Perspective, architecture, light and every other part of the image are preserved exactly — even after several edits in a row.

How it works

A change takes a few seconds. You see the result directly in the image and can immediately re-describe the area if it is not right yet. You do not need a 3D model for this — masking works on the finished image, whether render, photo or screenshot.

Why quality no longer degrades

Anyone who has edited an AI image several times in a row knows the effect: each pass recomputes the entire image. Textures get softer, colours shift, details disappear. After five or six steps the image looks different in places nobody wanted to change.

Masking avoids this. Only the marked area is recomputed; everything outside it still comes from the original image. That is why image quality holds up even after many edits — and you can approach a result step by step instead of having to get everything right in one go.

Four examples from practice

Replace furniture

Armchair, table or lamp — the rest of the room stays untouched. The armchair is new; floor, light and the art on the wall stay as they are.

Swap materials

Wood instead of concrete, stone instead of plaster — the geometry stays exactly the same. Ideal for facade variants without rebuilding the scene.

Add elements

A plant, a lamp, a picture — exactly where it is missing. Mark the empty spot and describe what should be there.

Add people

People for scale and life — at exactly the marked spot. Interiors in particular start to feel lived-in rather than staged.

Where masking reaches its limits

Very small details, highly reflective surfaces and transparent materials such as glass remain demanding — especially when a change would have to repeat in a reflection. In those cases a fresh generation gets you there faster.

Close-up faces are tricky too. For people in the image, medium and longer distances work far more reliably — which is usually the better choice for scale and atmosphere anyway.

Try it now