srneely@watcgl.UUCP (Shawn Neely) (02/24/86)
In the SIGGRAPH '85 proceedings Ken Perlin describes a solid texture function called Noise() which has, among other properties, statistical invariance under rotation. An algorithm is given for computing the function by smooth interpolation of pseudo-random values on an integer lattice. This method gives a texture which has a slight but noticeable grain along the coordinate axes. Has anyone else experimented with this function? I have played with the notion of a "gravity field" whereby the function value at a given point is determined by its distance to neighbouring pseudo-random "masses", but I still have undesirable correlation with the coordinate system. Any suggestions?