Lets you add noise to an attribute using one of 5 types of FBM based noise. The range of the noise is kept consistent regardless of the values of octaves and roughness. It also avoids the issues that built-in noise nodes have for input positions located around {0, 0, 0}.
This page also gives details on the corresponding VEX noise functions that you can use in your own VEX code.
The table below shows the different types of noise available, and the corresponding Vex function that is being called. The
| Displayed Name | siLib Function | Range |
|---|---|---|
| Perlin | si_noise <sig> | -1 -> +1 |
| Simplex | si_xnoise <sig> | -1 -> +1 |
| Orig Perlin * | si_onoise <sig> | -1 -> +1 |
| Sparse Convolution * | si_snoise <sig> | -1 -> +1 |
| Alligator * | si_anoise <sig> | 0 -> +1 |
* These types of noise do not support Vector4 types as input. The 4th component of frequency and offset will be greyed out.
<sig> A two letter suffix designating the return type and input type e.g. “fv” means output type is Float, input type is Vector3. See below for more details.
All of the noise functions but one, are centered around zero and approximately cover the range from -1 to +1. In practice, you’ll find it rare for values to exceed ± 0.9.
Alligator noise is different in that you could consider it to be an abs(noise) function. In other words it is centered on zero with a range of -1 to +1, but any negative values are made positive to give the final range of 0 to +1. Note that this is very different from being a distribution that is centered at 0.5 and having a range of 0 to 1, as outlying values are less common.
Since some of the noise functions only work with certain input/output data types, the drop down for the Noise Type parameter dynamically changes to only let you select valid options. This is managed using a Python script stored in the HDA, and the script is automatically copied to any Parameter node that you connect to the noisetype input, or when you use either of the “Promote Parameter” or “Promote Subnet Input” options from the middle click port menu. This means that your parameter interface will show the valid noise types automatically.
The following VEX functions are made available so you can easily use these noise functions in your own VEX scripts. To use it, you need to import the “si_noise.h” file like this:
#include "si_noise.h"
If you get errors including this file, check that your HOUDINI_VEX_PATH is set (use hconfig on the Houdini command line) and make sure siLib is installed properly.
The code block below shows the prototypes for the VEX functions. Which input and output type you need will define which function you need to use.
You’ll notice that the last three noise types only support Vector3 types as input. The Turbulize Attribute node wraps these functions
for you to emulate support for the ff and vf signatures.
// Perlin Noise
float si_noiseff ( float pos, freq, offset; float rough; int octaves)
float si_noisefv ( vector pos, freq, offset; float rough; int octaves)
float si_noisefp (vector4 pos, freq, offset; float rough; int octaves)
vector si_noisevf ( float pos, freq, offset; float rough; int octaves)
vector si_noisevv ( vector pos, freq, offset; float rough; int octaves)
vector si_noisevp (vector4 pos, freq, offset; float rough; int octaves)
// Simplex Noise
float si_xnoiseff( float pos, freq, offset; float rough; int octaves)
float si_xnoisefv( vector pos, freq, offset; float rough; int octaves)
float si_xnoisefp(vector4 pos, freq, offset; float rough; int octaves)
vector si_xnoisevf( float pos, freq, offset; float rough; int octaves)
vector si_xnoisevv( vector pos, freq, offset; float rough; int octaves)
vector si_xnoisevp(vector4 pos, freq, offset; float rough; int octaves)
// Orig Perlin Noise
float si_onoisefv( vector pos, freq, offset; float rough; int octaves)
vector si_onoisevv( vector pos, freq, offset; float rough; int octaves)
// Sparse Convolution Noise
float si_snoisefv( vector pos, freq, offset; float rough; int octaves)
vector si_snoisevv( vector pos, freq, offset; float rough; int octaves)
// Alligator Noise
float si_anoisefv( vector pos, freq, offset; float rough; int octaves)
vector si_anoisevv( vector pos, freq, offset; float rough; int octaves)
| Name | Description |
|---|---|
| Signature | Specifies the input and output types for the VOP node. |
| Noise Type | The class of noise to apply. The available options will change depending on the setting of the Signature parameter and/or the data type of the input/output connections. |
| Position | The position to evaluate the noise at. You will usually want to connect a node to this input. |
| Frequency | The frequency of the noise. |
| Offset | A vector to offset the center of the noise. Unlike some of the built-in noise nodes, you shouldn’t see any issues when position and offset values are set around {0, 0, 0}. Internally, this is achieved by reflecting each successive layer of noise in each axis. |
| Octaves | The number of layers of noise to apply. Each layer uses double the frequency of the previous layer. |
| Roughness | Each layer of noise applied has its amplitude multiplied by this amount to progressively reduce the affect. A roughness of 1 means that each layer is applied with exactly the same amplitude as the first. A roughness of 0 will mean that only the first layer of noise has any affect. |
| Bias | Optionally allows you to control the distribution of the noise using a Bias shaping function. |
| Gain | Optionally allows you to control the distribution of the noise using a Gain shaping function. |
| Amplitude | The amplitude of each component of noise. |