That Dress, and Its Implications for Color in Photography: Di/Tri/Tetra-Chromats as Audience?

Hi,

For those not Facebook-aware, here you go:

http://www.wired.com …e-one-agrees-color-dress/

So, about 1/4 of our population, The Minority, sees the Correct colors. Most of my fiends see the white and gold and are totally astounded anyone could possibly see anything else. Well, I must be seeing things cause I see a blue and some kind of mushy rusty/muddy color. Then this article started to pop up:

https://www.linkedin.c​om …olors-p-prof-diana-derval
[EDIT: see kaola yummies' post below. it appears this article has been debunked.]

Orig: [That article implies that it would be the tetrachromats (~25% of our population) are the ones seeing the correct colors. (Just FYI--I am red/green color deficient, but I do see at least 37 individual colors in that test, though on my crappy work computer screen. I need to recheck on my calibrated screen at home.]
[EDIT: see kaola yummies' post below. it appears this article has been debunked.]

So all this got me thinking... it looks like we have 3 different and prominent populations of color-seeing people. And that dress clearly shows that we have at least 2 populations at war with each other, in respect to visualization of color...

... so that tells me, even if we spend all the money and the time in the world to make your colors exactly "correct", a large part of our population may see completely different colors!

So what does this mean? Do we care?

...
ps--Aw heck, just saw there are at least 2 other threads on this. Well, anyway, I think the focus of this thread is on its implications and is different than the others, but I guess we'll see...

Thanks for posting the Linked In link. I see 36–38 colors in the array; it's hard to tell whether some adjacent bands are really different. The frequency of tetrachromats is a surprise. I thought they (we?) were rare. I did less well on another test, recently posted, that required rearranging swatches.

Yellow doesn't annoy me when it's on birds or flowers, but I don't wear it, because I don't think it looks good on me (or on most people).

I've always seen the dress as light grayish blue and muddy brown. It never switches to anything else. But is seeing its colors accurately the reason it doesn't switch? Other articles explain that it switches because it looks like either a white dress in shadow or a blue dress in bright light, and human brains automatically correct it to one of those. I don't get how missing some slight nuances in the blue and the brown would cause switching. I don't get how seeing all the nuances would prevent switching. Whatever hues people see in the dress, they all have years of experience seeing medium tones in different lights. If their brains tend to jump on one of the two possibilities and restore the colors as perceived to the way they think the dress looks in normal light, I'd expect that the fabric looking precisely this particular blue, not a slightly different blue, wouldn't be required. Their brains should make the jump regardless. Even if the photo were black and white, they should see the dress as lighter or darker than the tones in the image. Gray looks different in light and shadow as much as colors do.

Vision scientists are excited about this dress uproar because they've never before found such a radical difference in color perception among viewers without identified color deficiencies. I now wonder whether this kind of thing doesn't happen all the time while going unnoticed as people assume that everyone sees what they themselves see.

It does seem that large segments of an audience for photos will miss some color in your images or will see some that you miss.

In the main thread, I asked questions of the switchers, mostly not answered, to find out the conditions that affect what they see. It may have only made me look obsessional. I wish others were more interested in citizen science.

Xyclopx wrote in post #17456650
https://www.linkedin.c​om …olors-p-prof-diana-derval

I see 37 (and I see the dress in blue/bronze).

That means we should take more time to concentrate on composition, emotion, light direction, subject, background and other meaningful things instead of getting those "correct colors" and perfect white balance.

I'm not surprised that perceptions differ among viewers, but I'm very much surprised that perception shifts for individuals. I'm usually able to see light balance color shifts in nature--I can see white in open shade with a bluish cast, or white with a greenish tone reflected by foliage. But I can also let my brain ignore it, as most people do, and understand the color as white, not green or blue.

But I can't see or understand a single individual's shift of perceived color in a single image based on suddenly knowing the real-world conditions, and that hasn't been something apparent in my experience with clients. If I failed to correct for the color balance of lighting, no amount of explanation after the fact has ever convinced one of my clients that if she looks hard enough, her skin isn't really green.

I'm able to see the dress in both blue/black and white/gold. For me, it's a bit like looking at an optical illusion, like a drawn cube being seen to poke out of the page, then shifting my preception so it's poking into the page. If I focus for a couple seconds, I can change my perception of it.

I also counted 38 on the chart at the LinkedIn post.

I saw all 39 on my phone and can only count 37 on my LCD monitor. Maybe I need to calibrate my monitor.

going to put this here, worth a read if you are interested in human perception. i'm personally not a fan of the short snippets of half thought out brain spasms, but it might be a jumping off point for those interested, particularly the last link in this "article".


http://www.buzzfeed.co​m …-about-autism#.rklzjeZVzQ

and one of my favorite quotes, not because of the author, but because of the content.

"I'm for mystery, not interpretive answers. … The answer is never the answer. What's really interesting is the mystery. If you seek the mystery instead of the answer, you'll always be seeking. I've never seen anybody really find the answer, but they think they have. So they stop thinking. But the job is to seek mystery, evoke mystery, plant a garden in which strange plants grow and mysteries bloom. The need for mystery is greater than the need for an answer."

"The Art of Fiction" - interview by Robert Faggen, The Paris Review No. 130 (Spring 1994)

During the day that the dress was shown on the BBC news programmes, to me it was clearly the same as on the right-hand photograph, with strong blue and black. I saw this several times throughout the day. On the evening news, with what appeared to be the same programme edit, the dress appeared as it does in the centre photograph, with weak-looking colours. I presume that they had changed it and were messing about.

I can see all the colours in that chart. I don't know what that means, but I don't believe that they showed the same photograph during the day as they did in the evening.

hmm, no one is saying that they see < or = 32 colors, which should be ~75% of the population per the article. maybe it's bs?

i've only seen answers between/including 37 and 39 here and on facebook.

Well one of the responders went at the article pretty hard:

"Brian D. Flynn
Online Course Producer at General Assembly
If this is the quality of research one can expect from Prof. Diana Derval, then DervalResearch must be a terribly incompetent organization. This wildly oversimplified article does a disservice to a branch of science that's actually really amazing.

Most humans have 3 color receptive cones in their eyes, which recognize red, green, and blue. Every color that we see is created from mixing these colors.

About 2-3% of women (and only women) have a 4th color-receptive cone. This ability is limited to women because color receptive cones are connected to the x chromosome (which is why far more men are colorblind than women).

Of these women with 4 color receptive cones, many just have a second red cone and thus do not have enhanced color vision. A small percentage of these women have a 4th cone that recognizes yellow, providing the POSSIBILITY for them to recognize an additional spectrum of colors.

Out of this small percentage of a percentage of women, only about 10% are able to pass a test indicating that they can actually recognize extra color sensitivity as a result of their mutation. Only a handful of women have been discovered with this ability so far. NOT 25% of the population, which is just dumb.

Further, even if that were true, no test on your computer could properly test for tetrachromacy, as all color on all computer screens is created from red, green, and blue light. A good test would need to look for differences in yellow light. Not to mention the differences in color created from different browsers, screens, and internal settings. Sorry to burst everyone's bubble.

If you'd like to actually learn about this fascinating concept from a source that actually does solid research, check out this excellent podcast: http://www.radiolab.or​g/story/211119-colors/"

Another:

"Joseph Cacioppo
DVM/PhD Student at University of Illinois College of Veterinary Medicine
1) The citation in your book to the portion discussed above is incorrectly written; you cite Kimberly et al. 2001, when the citation should be Jameson et al. 2001 (Kimberly is her first name).
2) You can only be a human with four types of cones if you're a woman because it's X-linked, as Dr. Jameson discusses in her paper, unless you have Klinefelter syndrome or something similar."

Article writer's response:

"Prof. Diana Derval
Expert in Neuromarketing
Dear Joseph thank you for your post and checking the book! Yesterday only men could be dichromats and only women could be tetrachromat, today the same people say that some women and be dichromat [direct quoting here not my typo] and some transgender men can be tetrachromat - in the meantime we continue our research and thanks to all the people who shared their experience and feedback."

koala yummies wrote in post #17457386
Well one of the responders went at the article pretty hard . . .

Wikipedia's article on tetrachromacy refers to lines of research that suggest different frequencies of "seeing something extra in the spectrum" in the human population. It's very rare, or it's rather common, or nobody knows.

Don't computer screens produce a range of yellows with different wavelengths, and isn't that good enough?

koala yummies wrote in post #17457386
"Prof. Diana Derval
Expert in Neuromarketing
Dear Joseph thank you for your post and checking the book! Yesterday only men could be dichromats and only women could be tetrachromat, today the same people say that some women and be dichromat [direct quoting here not my typo] and some transgender men can be tetrachromat - in the meantime we continue our research and thanks to all the people who shared their experience and feedback."

wow, thanks for reading the comments and finding this... so, she didn't mention none of that stuff about "yesterday" either. even her answer is misleading/wrong.

well, anyway, i edited the original post to say it was debunked.

but i think the same question remains: The Dress proves not everyone perceives the same colors. does that matter to how we do color adjustments during post-production?

Xyclopx wrote in post #17457597
wow, thanks for reading the comments and finding this... so, she didn't mention none of that stuff about "yesterday" either. even her answer is misleading/wrong.

well, anyway, i edited the original post to say it was debunked.

but i think the same question remains: The Dress proves not everyone perceives the same colors. does that matter to how we do color adjustments during post-production?

Not a new issue. In the philosophy of aesthetics, it's been a matter of debate for 3000 years whether my "red" was ever the same as your "red." In the same vein, it's interesting that it's only been in the last thousand years that the color "blue" has been discovered, and some primitive people on the planet today still haven't discovered it. Yet, we know that what we call "blue" (let's agree for the moment that what you call "blue" is the same thing I call "blue") has always existed...yet it appears to have been the last color humans actually identified as something distinctive.

Photographers lost control of that when clients started judging images from their own computer screens. Virtually nobody outside imaging professional and enthusiast circles views images on anything close to controlled (much less actually accurate).

Personally, I usually go deliberately "wrong" with my portraits, slanting toward a tone that matches the mood of the photograph.

What I think that the dress issue tells us is that if the photograph is both exposed incorrectly, and has very bad colour balance, that you can hit just the right combination of effects to cause this to happen. I think that it will be incredibly difficult to hit that exact same situation again. Given a correctly exposed and colour balanced image it is pretty unlikely to become an issue in general photography. This is especially likely where the photographer is also the person who is doing the processing, as they will have been there so should know what colours to aim for in the processing.

Now of course everyone will see the colours in all our images slightly differently. Even then though if you ask them about "normal" ranges of colours that one sees a lot, such as in nature, they are going to give the same "names" to the same colours. Thats simply because they will have been taught the names of colours by being shown those colours. It is only if we start asking people to describe exactly what shade of colour something is that we will start to see differences.

Alan

BigAl007 wrote in post #17457646
What I think that the dress issue tells us is that if the photograph is both exposed incorrectly, and has very bad colour balance, that you can hit just the right combination of effects to cause this to happen. I think that it will be incredibly difficult to hit that exact same situation again.

That'd be an interesting thing to test with different colors and so on. A photo needn't be inaccurate to bring on the effect. If the dress were really the colors that nonswitchers see in the photo, the light blue and the brown, and it were photographed correctly, against a light background, the photo would look just like this one and would cause switchers to switch.

We still won't know what they see. If they say "This brown is like coffee beans but a little redder," they're talking about their experience of coffee beans, not the investigator's.