This operator computes 1D, 3D, and 4D Voronoi noise, which is similar to Worley noise but has additional control over jittering (i.e. how randomly the points are scattered through space) and returns the actual locations of the two nearest points.
Voronoi noise works by scattering points randomly through space according to a nice Poisson distribution, generating cell-like patterns. The generated noise is not anti-aliased. For best shading results, use the anti-aliased Celluar Noise instead.
Though this operator is slightly more expensive than Worley noise, the fact that it computes the actual point positions allows it to overcome some of the artifacts of Worley noise, such as getting even widths along the cell boundaries.
You can look at dist1 as the amount of generated noise (see other pattern generators such as Boxes or Stripes), which can be connected to a mixing bias (see Mix), a displacement amount (see Displace Along Normal), or other float inputs.
NOTE: The returned distance and points are in the noise space, so need to be divided by frequency and offset to move back to world cooridinates.
The seed associated with the first closest point is also returned. The seed is pretty much guaranteed to be unique for every point, meaning that it is unlikely that two points close by will have the same seed associated with them.
If the periodicity (
period) input is connected, periodicity will be
factored into the noise computation.
The relative costs for computing noise of different types is roughly:
Cost | Noise Type -----+------------------------- 1.0 | Perlin Noise (see Periodic Noise operator) 1.1 | Original Perlin Noise (see Turbulent Noise operator) 1.8 | Worley Noise (see Worley Noise operator) 1.8 | Periodic Worley Noise (see Periodic Worley Noise operator) 1.9 | Voronoi Noise 2.1 | Sparse Convolution Noise (see Turbulent Noise operator) 2.3 | Alligator Noise (see Turbulent Noise operator)