The Noise SOP displaces geometry points using noise patterns. It uses the same math as the Noise CHOP.
Parameters - Noise Page
/group - If there are input groups, specifying a group name in this field will cause this SOP to act only upon the group specified. Accepts patterns, as described in Pattern Matching.
/attribute - This menu sets which attribute of the geometry the Noise SOP acts on.
/function - The noise function used to generate noise. The functions available are:
- Sparse - Produces high quality, continuous noise based on Sparse Convolution.
- Hermite - Quicker than Sparse, but produces lower quality noise.
- Harmonic Summation - Sparse noise with the ability to control the frequency step of the harmonics. Takes the longest to compute.
- Brownian - Works like a bug in random flight. With Num of Integrals at 2, its acceleration is changed randomly every frame.
- Random - (White Noise) Every sample is random and unrelated to any other sample. It is the same as "white noise" in audio.
- Alligator - Cell Noise.
/seed - Any number, integer or non-integer, which starts the random number generator. Each number gives completely different noise patterns, but with similar characteristics.
/period - The approximate separation between peaks of a noise cycle. It is expressed in Units. Increasing the period stretches the noise pattern out.
Period is the opposite of frequency. If the period is 2 seconds, the base frequency is 0.5 cycles per second, or 0.5Hz for short. Hz refers to Hertz, the electrical and audio engineer of the 19th century, not the car guy.
If the Type is set to Random, setting this to zero will produce completely random noise. Otherwise, the period should be greater than zero.
/harmon - The number of higher frequency components to layer on top of the base frequency. The higher this number, the bumpier the noise will be (as long as roughness is not set to zero). 0 Harmonics give the base shape.
Harmonics with a base frequency of 0.1Hz will by default produce harmonics at 0.2Hz, 0.4Hz, 0.8Hz, etc. (up to the number of harmonics specified by the Harmonics parameter).
/spread - The factor by which the frequency of the harmonics are increased. It is normally 2. A spread of 3 and a base frequency of 0.1Hz will produce harmonics at 0.3Hz, 0.9Hz, 2.7Hz, etc.. This parameter is only valid for the Harmonic Summation type.
/rough - Controls the effect of the higher frequency noise. When Roughness is zero, all harmonics above the base frequency have no effect. At one, all harmonics are equal in amplitude to the base frequency. When Roughness is between one and zero, the amplitude of higher harmonics drops off exponentially from the base frequency.
The default roughness is 0.5. This means the amplitude of the first harmonic is 0.5 of the base frequency, the second is 0.25, the third is 0.125. The harmonics are added to the base to give the final shape. The Harmonics parameter and the Roughness parameter must both be non-zero to see the harmonic effects.
/exp - Pushes the noise values toward 0, or +1 and -1. (It raises the value to the power of the exponent.) Exponents greater than one will pull the channel toward zero, and powers less than one will pull peaks towards +1 and -1. It is used to reshape the channels.
Num of Integrals
/numint - Defines the number of times to integrate (see the Speed CHOP) the Brownian noise. Higher values produce smoother curves with fewer features. Values beyond 4 produce somewhat identical curves. This parameter is only valid for the Random noise type.
/amp - Defines the noise value's amplitude (a scale on the values output).
Parameters - Transform Page
Translate / Rotate / Scale / Pivot
/t[xyz] /r[xyz] /s[xyz] /p[xyz] - The Translate, Rotate, Scale and Pivot parameters let you sample in a different part of the 3D noise space. Imagine a different noise value for every XYZ point in space. Normally, the Noise SOP samples the noise space from (0,0,0) along the X-axis in steps of 2/period.
By changing the transform, you are translating, rotating and scaling the line along which the Noise SOPs samples the noise space. A slight Y-rotation is like walking in a straight path in the mountains, recording your altitude along the way, then re-starting from the same initial location, walking in a slightly different direction. Your altitude starts off being similar but then diverges.