Cluster operator

The Cluster operator allows you to partition particles into groups and then save the resulting group index to a custom data channel, for later use in other operators.


Cluster Rollout

Cluster Type

  • Voronoi cells: cluster groups will take the shape of voronoi cells, derived from nearest-point proximity groupings.

  • Turbulent noise: cluster groups will be derived from 3D turbulent noise values.

  • LDNP: cluster groups will be derived from LDNPs (local distributions of neighboring particles). This is an approximated k-means clustering method.

  • Random: cluster groups will be chosen at random, based on a min/max range of values.

Cluster Values

  • Max unique: the maximum number of cluster groups.

  • Bias: the amount of bias that group indices will have towards lower values.

  • Channel: the custom data float channel to save the resulting group indices to.

Display

  • Show cluster points: show the individual cluster points (voronoi cell centers) in the viewport.

  • Visual clusters: display each particle as a random color, based on their cluster group index.

Uniqueness

  • Seed: the seed value for all varied parameters.

Voronoi Rollout

Voronoi cell point locations

  • Particle cloud center: the center of the point cloud will be located at the center of the particle cloud.

  • Object pivots: point clouds will be located at the pivot points of input objects.

Objects

  • Input object list: the list of scene objects to use as point cloud centers.

Voronoi Cell Points

  • Count: the number of points in the point cloud.

  • Sphere/Box: the shape of the point cloud bounds.

  • Scale X/Y/Z: the scale of each point cloud axis.

  • Scale mult: the overall scale multiplier for the point cloud.

  • Variation %: the per-particle percentage of variation to apply.

Deviation

  • Fuzzing: the amount of implicit jitter to add to each particle prior to point-cloud proximity searches, which will blur the borders between the resulting cells.

  • Noise amount: the amount of implicit perlin noise to add to each particle prior to point-cloud proximity searches, which will shift the borders between the resulting cells.

  • Noise Scale: the scale of the perlin noise.

Voronoi plane normals

  • Scale X/Y/Z: the amount of scaling to add to voronoi cell walls, along each axis. Adjusting these values will stretch the overall shape of the resulting voronoi cells.

Noise Rollout

  • Noise mode: controls which noise algorithm will be used.

  • Noise texmap: the texmap that will be used by the noise texmap mode(s).

  • Noise preview: a preview image showing a 2D representation of the selected noise mode.

  • Strength: the strength of the noise (a multiplier on the default noise range of [-1, 1]).

  • Frequency: the speed at which the noise will evolve over time.

  • Scale: the scale multiplier for position values sent through the noise algorithm. Smaller values create larger noise patterns.

  • Roughness: controls the amount of extra detail applicable noise modes will generate.

  • Lacunarity: controls the scale of successive noise octaves for applicable noise modes.

  • Gain: controls the relative intensity of successive noise octaves for applicable noise modes.

  • Iterations/Octaves: controls the number of overlapping noise patterns that applicable noise modes will generate.

  • Phase: provides manual control over the evolution of the noise over time.

Noise Settings Rollout

  • Min level: clamps the minimum value of the resulting noise values.

  • Max level: clamps the maximum level of the resulting noise values.

  • Fuzzing: the amount of implicit jitter to add to each particle prior to noise calculations, which will blur the borders between the resulting shapes.

  • Offset X/Y/Z: the amount to offset particle positions in space prior to sampling their noise values. Changing these values will shift the overall noise pattern through the particles.


LDNP Rollout

  • Radius: the radius of the particle neighborhoods.

  • Variation %: the per-particle percentage of variation to apply.
  • Fuzzing: the amount of implicit jitter to add to each particle prior to point-neighborhood proximity searches, which will blur the borders between the resulting cells.


Random Rollout

  • Min: the minimum random cluster value.

  • Max: the maximum random cluster value.