Adding color to the previous example, we obtain tinted Perlin Clouds; allowing the color values to overflow past 1 gives us Perlin Plasma. A shift in coordinates reveals that the Frocessing implementation of Perlin noise is symmetric about all three axes (not necessarily implying that it’s incorrect if you stay in one octant…), producing Perlin Kaleidoscope and Perlin Oriental Rug.

Here are two instances of the same demo, modeled after the one Seb Lee-Delisle showed last fall, built with two different Flash implementations of 3D Perlin noise. The one on the right uses Ron Valstar’s Perlin class (as does Seb’s demo), which appears to be a faithful port of Ken Perlin’s Improved Noise reference implementation in Java, and therefore reliable.
The one on the left is built on the PerlinNoise class from the Frocessing library by Tomoaki TAKANAWA. This library is in an early stage but it is intriguing nonetheless. Note, though, the different responses of the two instances to changes in the parameter values. The Frocessing implementation seems to be faster but perhaps a bit buggy. Anyway I’ll be watching the development of Frocessing with interest.

UPDATE: I have rebuilt the Valstar example using Mario Klingemann’s optimized version of Ron’s class.

Frocessing Resources

• Takanawa’s Frocessing demo
  His code is presented as a Flash frame script; here is a .as version of it suitable for Flex Builder. He also notes that for performance reasons, Font and Image implementation will have to wait for the Astro release. How close can Flash ultimately come to Java performance?
• Frocessing Wiki (Japanese)
• ASDocs
• Frocessing source code (library and samples)