Noise Options - PolyhedralDev/Terra Wiki
Wherever you see a "noise options" section in your config, you can define a noise sampler there, using the following values:
The type of sampler to use. Included samplers:
- Linear - Linear normalizer.
- Normal - Normal (normal -> continuous) normalizer.
- Clamp - Clamp normalizer.
- Expression - Paralithic expression noise function
- Image - Pull noise from an image.
- DomainWarp - Domain-warp a function with another function.
- FBM - Fractal Brownian Motion.
- PingPong - Ping Pong fractal type.
- Ridged - Ridged fractal type.
- OpenSimplex2 - OpenSimplex2 noise, regular variant.
- OpenSimplex2S - OpenSimplex2 noise, smooth variant.
- Perlin - Perlin noise.
- Value - Gradient noise.
- ValueCubic - Cubically interpolated gradient noise.
- Cellular - Cellular (Voronoi/Worley) noise.
- WhiteNoise - White (random) noise.
- Gaussian - Gaussian (normally distributed random) noise.
- Constant - Constant value.
- Kernel - Apply a kernel to an input function.
A value to add to the seed of the noise function. Noise functions are seeded with the world seed, adding a salt allows multiple otherwise identical noise functions to produce unique results in the same world.
The frequency of the noise function. Higher values = higher frequencies = quicker changes in noise values. Frequency has no effect on white noise (other than effectively being a second salt).
Number of dimensions of this noise function. Must be either 2 or 3.
Fractal Noise Types
The following options are supported by all fractal noise types:
Defines the number of fractal octaves to generate. A fractal octave is a single noise function. For example, 5 octaves would create 5 noise functions stacked on top of each other.
Function to apply fractal to.
Controls the fractal gain. The gain sets the amplitude multiplier of each subsequent noise function.
For example, the default gain of
0.5 would cause an input noise function to have noise added with half the amplitude, then one
quarter, then one eighth, etc.
Controls the fractal lacunarity. The lacunarity sets the frequency multiplier of each subsequent
noise function. For example, the default lacunarity of
2.0 would cause the first octave to have the base frequency,
the second to have 2 times the frequency, the third to have 4 times, etc.
Sets the strength of the "ping-pong" algorithm. Higher numbers will produce more defined stripes.
Sets the weight of each octave. Higher values will cause octaves to produce higher weights
at higher values. Note that values outside of
[0, 1] will not maintain
[-1, 1] output.
Options for the
Cellular noise variant.
distance sets the distance function to use for calculating the cell border. Supported distance functions:
Euclidean- Euclidean distance.
EuclideanSq- Calculate distance based on the square of Euclidean distances. This is faster than
Euclidean(as it does not require a square root operation) and produces effectively the same result.
Manhattan- Manhattan (Taxicab) distance.
Hybrid- Average of EuclideanSq and Manhattan distances.
return sets the return type of the cellular noise function Supported return types:
CellValue- Returns the value of the cell itself.
Distance- Returns the distance to the nearest point.
Distance2- Returns the distance to the second nearest point.
Distance3- Returns the distance to the third nearest point.
Distance2Add- D1 + D2
Distance2Sub- D1 - D2
Distance2Mul- D1 * D2
Distance2Div- D1 / D2
Distance3Add- D1 + D3
Distance3Sub- D1 - D3
Distance3Mul- D1 * D3
Distance3Div- D1 / D3
NoiseLookup- Return the value of the cell's center point when passed into the
lookup defines another noise function to use in the
NoiseLookup return type.
Default: OpenSimplex2 function with no fractal, and 0.02 frequency.
Options supported by all normalizer types:
Function to normalize.
Linear redistribution redistributes the input from
[min, max] to
Minimum input value
Maximum input value
Normal redistribution redistributes the input from a normal distribution with mean
mean and standard deviation
standard-deviation to a continuous distribution with range
The mean of the input function.
The standard deviation of the input function.
The clamp normalizer clamps an arbitrarily ranged input function to a maximum/minimum value. Any value above the maximum will return the maximum, and any value below the minimum will return the minimum.
The source function. This function will be used to retrieve noise values.
Function to use for domain warping. The result of this function will be added to the X, Y, and Z inputs of the
The amplitude of the domain warp. This value is the maximum an input point may be translated (provided that the warp
function is bounded from
The path of the image to use, relative to the config root.
The channel to get values from. Valid channels are:
RED- Red channel of the image
GREEN- Green channel of the image
BLUE- Blue channel of the image
ALPHA- Alpha channel of the image
GRAYSCALE- Average of RED GREEN and BLUE channels.
The input coordinates are multiplied by this value, allowing for image scaling.
A 2D array representing the kernel to apply to the input function.
Examples of common kernels.
Factor to multiply all kernel values by. Useful for keeping kernels readable. Defaults to 1.
Noise function to apply the kernel to.
The equation to use for this noise function. The variables
z, along with
any defined in the
variables key, are registered, along with the noise functions defined in
A map of noise functions. These functions are registered for use in the equation.
A map of variables. These variables are registered for use in the equation.