Custom PBR materials vs stock textures.

Start with the consequence of being wrong.

If a background concrete is slightly different from the reference, the shot may still work. If a footwear knit, branded coating, hero terrain or product finish is wrong, the material can break recognition, approval or consistency across every asset that uses it.

The question is not whether custom is always better. It is whether the project loses something important when the surface is generic.

A library asset already solves the job.

• The material sits in the background and is not a recognition cue.
• The art direction allows broad interpretation.
• You need a common surface quickly and can adjust it internally.
• The license, resolution and map convention fit the production.
• No reusable generator or editable source is required.

Buying stock and spending an hour validating it is often cheaper than commissioning a surface that adds no meaningful distinction.

The material has a specific job.

• It must match a physical sample, concept, brand finish or approved reference.
• Scale, pattern placement or wear needs art-directable control.
• The same system must produce many credible variants.
• Outputs must follow an engine-specific packing or naming standard.
• A complete library needs consistent construction and parameters.
• The team needs editable SBS/SBSAR sources, not only flattened maps.
• Existing scans or materials need cleanup, de-lighting or reconstruction.

Test the candidate in the real pipeline.

1. Check physical scale and visible repetition at the intended camera distance.
2. Inspect roughness and normal response under neutral, moving light.
3. Confirm channel conventions, normal orientation, bit depth and license.
4. Apply the material to a representative asset in the target renderer.
5. Estimate the internal time needed to adapt, repair and document it.

If the adaptation cost approaches a focused custom commission, custom work is easier to justify.