Module noise 

Coherent noise functions: Perlin, simplex, fractal Brownian motion.

Unlike hash functions (which produce uncorrelated “white noise”), coherent noise produces values that vary smoothly — nearby inputs yield similar outputs. This is essential for generating realistic time-series data, spatial fields, and any workload where adjacent coordinates should have correlated values.

The permutation table is built at init time from a seed. The noise evaluation runs at cycle time.

Inputs are u64 coordinates mapped to a float domain via scaling. Outputs are f64 in [-1, 1] (raw noise) or [0, 1] (normalized).

Structs

FractalNoise1d

1D fractal Brownian motion: layered Perlin noise with decreasing amplitude at each octave. Produces rich, natural-looking signals. Output is f64, roughly in [-1, 1]. Lacunarity is fixed at 2.0 and persistence at 0.5 (standard FBM parameters).

FractalNoise2d

2D fractal Brownian motion: layered Perlin noise in 2D. Produces terrain-like spatial variation. Lacunarity is fixed at 2.0 and persistence at 0.5 (standard FBM parameters).

Perlin1d

1D Perlin noise.

Perlin2d

The perlin_2d node.

PermTable

A permutation table for noise functions. Built from a seed at init time, immutable thereafter. The table is doubled (512 entries) to avoid modular indexing.

Simplex2d

The simplex_2d node.

Functions

fbm_1d

1D fractal Brownian motion: octaves layers of Perlin noise, each at twice the frequency and half the amplitude of the last.

fbm_2d

2D fractal Brownian motion over Perlin noise, as fbm_1d.

perlin_1d_algo

Evaluate 1D Perlin noise at a given point.

perlin_2d_algo

Evaluate 2D Perlin noise at a given point.

simplex_2d_algo

Evaluate 2D simplex noise at a given point.