"Math error" found in foundation of color theory.

[Doug A]

https://petapixel.com/2022/08/11/math-error-found-in-foundation-of-century-old-color-theory/

 

This could impact a lot of things. Not sure this is the correct place to post this.

Thanks,

Doug A

[Andy Perrin]

I work with a color scientist pretty regularly, and what they "discovered" is in fact well-known in that field (which has advanced a lot since the days of Max Planck!). In fact, the way color scientists measure the "distance" between two colors is with the CIE2000 ∆E:

https://en.wikipedia.org/wiki/Color_difference#CIEDE2000

 

That is only a quasimetric, which of course is not fully a General Relativity-style metric like these people were looking for. But what it does have going for it is that it takes into account all kinds of subtle effects in human vision. Physicists are CONSTANTLY going into other people's fields and reinventing the wheel, and this seems to be yet another example of some physicists thinking they've discovered some fundamental error when that field has known about it for eons. [Edit: apparently they are psychologists? I think otherwise I would leave the statement as it stands...don't go into other people's fields and tell them they are wrong unless you are very very sure.]

 

ETA: Looking at the original article, it's a little more subtle than petapixel seemed to imply. They actually are contesting the CIE2000 ∆E and saying they know why it doesn't work over long distances in color space, which is more interesting than what I thought, which is that that they were claiming Planck's work needed an update. I should have checked the original. My bad!

[Doug A]

Good to know @Andy Perrin. So, nothing is likely to change due to this "discovery".  In a Star Wars voice, "There's nothing to see here".

 

Thanks for setting things straight,

Doug A

[Andy Perrin]

Yeah, I don’t think so. I’ll ask my friend if she heard anything. The reason I know about this in the first place was that when she was learning how to do the color correction stuff, she came to me for help on the programming, so we both read up on it together. 

[dabateman]

Yes I read that too.

I am not expecting any changes. Some manufacturers would test devices with actual people to change them. So it was related to their response and since shifted to more general response as more people bought into those products. 

[Andy Perrin]

David, I don’t understand your reply - this was about color measurement theory, not devices?

[dabateman]

The petapixel article thinks that now that this math error has been solved, devices will be more vivid and have better colors. 

Quote from Petapixel:

"The new discovery has the potential to create more vibrant monitors and screens, as well as printed materials and textiles."

 

I tried to look at the PNAS article,  it wasn't clear what track it was. Actually submitted and reviewed,  contributed or just published from a member.  My postdoc was with an Academy member,  so I know the differences between the 3 tracks. The authors I think are in psychology. 

 

[Andy Perrin]

Ah, yeah, I scanned over that as the usual blather. The article does look like they think they found an error in deltaE2000, though, not what petapixel made it sound like, so the junk about Planck is misleading. Having better color spaces is great but since most of us use sRGB online, which is known to be garbage, it will be a long time before any devices are affected.