Finally, the Scape terrain editor prototype is now officially open sourced and freely available for download! This article on Scape is meant to provide an aggregated collection of links to previously published documentation and downloads, as well as to the new source code and binary package downloads.
The Bernoulli distribution can be generalized to support more than two discrete states, where each state's probability is generated by linearly interpolating between two given extremes, evenly dividing the difference from one extreme to the other. This article describes the properties of this discrete staircase distribution, together with a method to generate random samples from it in constant time.
In this fourth article on the Scape terrain editor, the actual brush-based editing pipeline is described from beginning to end, which uses either the CPU or the GPU to update the areas affected by a user's brush strokes. Besides many pipeline optimization details, it also covers Scape's direction noise feature.
This is the third article in the series on Scape, a GPU-based terrain editor, picking up where I left off in the previous article on procedural noise techniques. In this article, I'll discuss two novel procedural noise-mixing algorithms that are capable of generating terrain types that seem to be heavily eroded.
Scape is a heightfield terrain editor I developed as part of my thesis work at W!Games in 2008. It allows the user to sculpt large terrains using procedural brushes in real-time, using GPGPU techniques for the bulk of the work. In this first article on Scape, its (procedural) terrain rendering techniques are explained.
Armed Mine is a creature I created for Doom 3 that uses procedural animation to drive motor-controlled physics joints. Animation-driven physics gives a lot of flexibility, but can be difficult to get right. Especially when the engine doesn't support this out of the box. Read how the Doom 3 physics engine was extended and the Armed Mine was implemented.
Although the repeated games used in the field of game theory are relatively trivial, they do allow for strategy analysis, comparison and optimization within a well-defined framework. This article reviews existing learning strategies like Q-learning and leader strategies like Godfather and Bully, extends the latter to improve on their weaknessess, and presents detailed tournament results.