Case Studies

The method applied to real projects

All three pillars brought together — starting point, deterministic decisions, creative latitude for the AI, workflow, result.

Highland Cottage

F1

Highland Cottage

From Blender skeleton to photographic result in three steps.

Minimal Blender geometry, maximum AI interpretation — and a composition preserved exactly through every step. The study shows how far the RAY-L workflow carries with minimal effort.

Granola Bowl

F2 · Food

Granola Bowl

Why the material in Blender is the real tool for ControlNet Canny.

Food photography presents a specific challenge: the actual subject doesn't exist in Blender at all, and the materials that make a bowl look attractive produce noise in Canny instead of edges. The study shows how changing materials in Blender — not adjusting thresholds — turns a noise field into a clean edge map. Plus: why Flux is the stronger choice for organic, moist subjects, and how a T5XXL prompt makes a food shot genuinely appetizing.

Interior visualization

F3

Interior Visualization

planned

Same room, two lighting moods, one Blender scene. How far a single scene setup can be varied — without touching the composition.

Architecture visualization

F4

Architecture Visualization

planned

Exterior shot, changing light, complex geometry. Where the limits of the workflow lie — and where they don't.