Empirical Comparison of Three Existing Methods for Simulating Embers in Computer-generated Fire

Abstract

This study compares computer-generated embers using three existing methods. While there exists research related to fire propagation, flame rendering, and smoke generation, there is nothing that discusses embers. This study focuses specifically on the generation of an ember's path when it is animated. The three methods selected were the Lattice Boltzmann Model, a 3D fractal tree, and a particle system that uses third-order polynomials. The three methods were compared using three different metrics: memory requirements, time requirements (computational cost), and a subjective visual representation. A different method performed best in each of the three metrics. The particle system had the lowest memory requirements. The 3D fractal tree had the least computational cost. The Lattice Boltzmann Model had the best visual representation. Over all, when rendering embers in a real-time environment, the 3D fractal tree is the best choice. However, if pre-rendering your images, then the Lattice Boltzmann Model is best.