Skip to content

Comment on Retrospective on 10 Years of colour-science

Comments

I got into a bit of color science when I worked on a physically-based, spectral renderer some 15-ish years ago. Truly a fascinating and complex topic.

It's relatively easy to make a spectral renderer, that part of the math is pretty straight forward.

But generating fairly accurate and good-looking images that can be displayed on a low dynamic range RGB monitor from the result is something else.

This is really a complex topic and we certainly do not have the answers to many questions! No longer than two days ago we were discussing about the fact that we need a Display Rendering Transform (DRT) (typically using a Sigmoid-Like function) to exhibit images on a display and there is not a full understanding as to why. If we think about the dynamic range of the scene being vastly different in many cases than that of the display, it makes sense, but even with very good display, which can cover the scene dynamic range, we still need a DRT. Part of the answer could be in the viewing conditions and surround being different but many of the answers lead to preference.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.