← All resources

The Perlin Problem: Moving Past Square Noise (NoisePosti.ng)

Auto-discovered

Blog Intermediate

NoisePosti.ng

Visit resource ↗

Description

NoisePosti.ng's January 2022 article opens a two-part Perlin Problem series on visible square bias in classic gradient noise for terrain and textures. It clarifies coherent noise, Perlin, and fractal Brownian motion versus library implementations.

The essay ties bias to hypercubic grid interpolation, compares value and value-cubic variants, and walks Simplex-type alternatives including OpenSimplex families, gradient-set quality, and contribution radius for ridged fractal stacks.

Domain rotation rotates 3D sample coordinates so grid-aligned artifacts leave the observed slice, with notes for 2D and animated use, optional math, and sample code; part two covers naming and tutorial guidance.

For intermediate procedural authors in any language or GPU path, the free post helps decide when to keep Perlin, adopt Simplex-type noise, or rely on rotation alone.