On why studying a real surface is the most underrated step in material creation - and why experience doesn’t make you immune to skipping it.
Most artists think they know what copper looks like. They've seen it their whole life - door handles, pipes, coins. So they open Substance Designer and start from memory. That's where it goes wrong.
You Think You Know, But You Don't
Memory doesn't retain that copper in shadow reads differently than copper in direct light. It doesn't retain that edge wear looks fundamentally different from surface wear. Or that verdigris doesn't spread uniformly - it collects in recesses, on horizontal surfaces, anywhere moisture sits. Photo reference does.
What Reference Gives You
The most valuable thing reference gives you isn't color - it's logic. When you study a real surface carefully, you stop seeing a material and start seeing a system of rules.
Where does wear concentrate? On the high points - edges, ridges, anything that gets touched or contacted. Where does oxidization start? In the recesses, the cavities, the areas that trap moisture and see less air. Where does dirt settle? The lowest geometry, the sheltered horizontal surfaces.
These aren’t arbitrary decisions made by an artist; they’re physical laws. Reference shows you what those laws produce in practice, and once you see them, you can't unsee them.
How to Study a Surface
Not all reference is equally useful. A quick image search gives you an impression. What you need is detail - close-up shots, multiple lighting conditions, ideally studio photography with controlled light.
When studying a surface I look for:
- Transition zones - where does clean end and worn begin? Is it a sharp edge or a gradual fade?
- Edge behavior - are worn edges brighter (polished from contact) or darker (oxidized from exposure)?
- Surface texture variation - a single material rarely has uniform roughness across its entire surface
- Texture frequency - is it mostly macro variation or high frequencies driving the overall feel? This affects color, roughness, and normals - and how those elements play off each other.
- Color under different lighting - metals shift hue significantly between shadow and direct light; getting this wrong is what makes a material look like a texture rather than a surface
Into the Build
Once you’ve studied the reference, the blend logic writes itself. You know which material states you need. You know which mesh data drives each transition - curvature for edge wear, ambient occlusion for trapped oxidization, a grunge map to break up anything that would otherwise look too uniform.
Without reference, those decisions are guesses. With it, they’re observations.
That’s the shift. It moves material creation from decoration to description.
© 2026 Stefan Groenewoud - All views are my own, not those of my employer.