Triplanar mapping can be visually and performance-wise improved by switching from linear to non-linear texture weight blending. Raising the per-axis weights to a power (e.g., pow(weights, 8)) sharpens transitions, eliminates visible layering on smooth meshes, and drives small weights toward zero — enabling a shader optimization that skips texture lookups for negligible weights entirely. The higher the exponent, the more aggressive the sharpening and the more effective the skip optimization becomes.
Sort: