3b5410aef9
This method takes a given normal, and computes a random ray in the unit-hemisphere described by that normal. We use cosine-weighted importance sampling because it leads to better convergence and is a nice micro-optimisation (from four trigonometric operations to only two). |
||
---|---|---|
beevee | ||
pathtracer | ||
.gitignore | ||
.gitlab-ci.yml | ||
.rustfmt.toml | ||
Cargo.toml |