Lore

Flat, shadowless, high-contrast base render

A deliberately minimal export from Blender that serves as the structural and colour input for an AI image-to-image pass. Blender is no longer asked to look good. It only carries what the AI cannot be trusted to invent: the layout of the scene and the colours the artist chose.

Division of labour

Settings

Why shadows are off

The downstream steps read this image in several ways: Florence captioning, depth and canny extraction, and the latent for image-to-image. Baked shadows and low contrast would blur object boundaries and could be misread as geometry or dark material. High contrast keeps objects separable for the ControlNet maps (depth-and-canny-controlnet-maps-from-input) and the caption (florence-2-detailed-captioning).

Known limits

Related

redundant-colour-carry-through-in-img2img, Free AI preparation workflow (outline and depth from a viewport render), Sketch-to-photo conversion.