Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Do not optimize ray trace method for far away colors
It turns out, when testing far away _and_ near colors, that optimizing with the smallest angle from origin (achromatic) can yield better results for far color, but worse results for near colors. This is just due to how the perceptual spaces curve through the linear RGB space. But by leaving things as they were previously (tracing from achromatic to the sample and using 4 iterations instead of 3) results are more consistent near to far. Colors nearer to the gamut are often more accurate, but far colors are still below ∆h of 2. It is better to have more accurate near and far colors than gain some speed to converge far colors faster.
- Loading branch information