Digital Light & Color

Happy New Year to one and all.

Just saw this article about a "perceptual color space for image editing", and wondered if it was of any interest...

https://bottosson.github.io/posts/oklab/

Thanks for bringing it to my attention.

It looks somewhat promising as a color space -- not sure if it will ever be widely adapted though. I was particularly impressed by the way in the examples it seems to avoid the Abney effect, namely the tendency of blue to turn purple as you blend it with white.

Off the top of my head, using it as a general purpose color space to edit in seems cumbersome and would require a rewrite of pretty much every transformation. Also, no file format or display hardware supports it so one would have to convert back to some RGB-based color space when reading or writing files or when displaying or printing images. This can be problematical since it is a device-independent color space and the conversion back to RGB may result in unrealizable or out-of-gamut colors. On the other hand, it is possible it could be used in some specific applications such as generating gradients, tinting black and white images, compositing images, adjusting saturation, or converting color to black and white. I would have to experiment with it to see if any improvements would be significant and worth the extra computation.

I was disappointed the article did not include real-life images to be able to see if I could "see" a difference, but I think I understand how you see it could be useful. The 2nd article on his blog (https://bottosson.github.io/posts/colorwrong/) had more examples of differences between other spaces.