Poly Haven photoscans on your battle map: what they add
A modelled prop is a shape someone built and textured by eye. A photoscanned prop is a real object, captured with real wear, real dents and a real light response baked into the geometry — and TerrainTale's clutter is built almost entirely from the second kind.
· 5 min read · TerrainTale
What a photoscan actually captures
A photoscan starts from photographs of a real object from many angles, reconstructed into a 3D mesh with a texture pulled from those same photos. The result carries genuine imperfection — a barrel's slightly uneven staves, a crate's real corner damage, a lantern's actual soot marks — the kind of detail a hand-modelled prop usually skips because nobody sculpts wear that nobody asked for.
About 225 props, from Poly Haven
TerrainTale's prop set is roughly 225 pieces sourced from Poly Haven: furniture, barrels, crates, chests, lanterns, candles, tools, weapons, statues, paintings, shrubs, plants, flowers, rocks, cliffs, a modular fort and a pier. All of it is CC0, the same licence as the rest of the library.
Why it reads as furniture, not decoration
Because a scanned prop already carries real wear and a real light response, it sits under dynamic lighting the way the object it came from would — a torch passing a scanned lantern catches its dented metal correctly, instead of the flat highlight a hand-painted texture would give it. Paired with the library's PBR-generated floors and walls underneath, a scene ends up looking like one consistent, lit space rather than a collection of separately styled pieces.
What scans are good for, and what they aren't
Photoscans excel at objects with real texture and irregularity — organic clutter, worn furniture, natural terrain like rocks and cliffs. They're not how the library's characters are made: the nine animated minis (knight, barbarian, mage, rogue, hooded rogue and four skeletons) are rigged KayKit models built for animation, not scans, since a scanned figure has no skeleton to animate.
Dressing a scene with scanned clutter
A handful of scanned props does more for a scene's believability than a large number of repeated pieces — a couple of crates, a lantern, a barrel with real wear reads as a lived-in room faster than a dozen identical, perfectly clean copies of the same asset. The building guide's advice to dress last still applies: get the shape and lighting right first, then place scanned props where the fight's sightlines and cover actually need them.
Spot-checking the difference on your own table
Drop a scanned crate and a plain colour-block placeholder side by side and rotate a torch past both. The scanned crate's edges catch light unevenly, the way real worn wood does; a flat-shaded placeholder just gets uniformly brighter or dimmer. It's a small thing on any one prop, but a scene dressed with a dozen scanned pieces accumulates that realism everywhere the light touches, which is most of what a player actually looks at during a fight.
None of this requires picking props by hand from a catalogue of licences and sources — everything in the library, scanned or generated, is CC0 and sits on the shelf ready to drag onto the table.
Props and surfaces built to sit together
A scanned prop and a generated PBR surface come from different pipelines — one from photography, one from material scans built into procedural tiles — but both carry the same kind of real light-response data, which is why they don't clash when placed together. A scanned crate on a generated stone floor reads as one scene rather than two different rendering styles bolted together, which is a common seam in libraries that mix asset sources less carefully.