# Roughness

A roughness node describes a 2D surface slope distribution. It may model a broad range of surface finishes, including anisotropic and measure-based ones.

# Zero

### Summary

The Zero roughness model corresponds to an ideal, perfectly smooth, specular surface.

### Link

# Cosine

### Summary

The Cosine roughness model is a very simple one : the probability of a given slope is proportional to its cosine. It corresponds to a very rough surface. It is strictly equivalent to phong with the exponent parameter set to 1.

### Link

# Phong

### Summary

The Phong roughness model is based on the specular part of the classical Phong BRDF, controlled by an exponent parameter.

### Link

# Beckmann

### Summary

The Beckmann roughness model is based on an isotropic gaussian distribution of slopes. Its roughness parameter corresponds to the standard deviation (RMS) of the slope. A value of zero corresponds to a perfectly smooth specular surface.

### Link

# Ashikhmin-Shirley

### Summary

The Ashikmin-Shirley roughness model is an anisotropic generalization of the phong model, based on the paper "An Anisotropic Phong BRDF Model".

### Link

Ashikhmin-Shirley roughness reference

# Map

### Summary

The map roughness allows giving a custom roughness distribution as an image, representing the normal distribution as a function of theta (vertically) and phi(horizontally).

### Link