More on Space Filling Algorithms…

A few years ago I wrote about Procedurally Distributing Game Objects, essentially how to sprinkle objects in game space so that they might appear randomly distributed, they are always somewhat equi-distant, thus making the density of object’s distribution in game space is nearly homogeneous.  These types of uniform distributions are frequently used in  techniques that use Monte Carlo simulations, such as ray tracing algorithms. A nice paper from Pixar’s Per Christensen, Andrew Kensler and Charlie Kilpatrick  – Progressive Multi-Jittered Sample Sequences adds three new sequential distributions for use is generating space-filling distributions. The paper goes in depth about measuring the distributions including some nice Fourier analyses – a great example of how to measure distribution and “randomness”.