7
\$\begingroup\$

Okay, seriously, I'm still a little obsessed here. We've got this blue-black/white-gold dress and a very interesting difference of perception. None of the explanations are really satisfying to me — I understand about white balance, but usually most people seeing a daylit photo incorrectly set to tungsten respond with "what's that weird blue tint everywhere? some kinda instagram filter". Here, though, to some people, the perception is so strong that even the idea of a blue tint is sometimes discounted.

This is fascinating to me as a photographer, and it seems like it should be fascinating to a lot of us, especially because we obsess so much about color calibration and getting colors to be accurate — yet here, the same image, seen by people from the same culture in the same conditions on the screen, clearly is perceived differently.

So, what am I doing over on meta? Well, what I really want to ask is "Can anyone reproduce this binary perception effect with a totally different subject?" I don't care how you do it... with an originally blue/black object or white/gold, or with gelled lights, or whatever — just, reproduce the effect in another scene. By doing that, we should be able to identify what elements are really essential to the effect.

But, fundamentally, "can you do this?" is more a challenge than a question. What do you think — okay to ask? Or should I just let it go?

\$\endgroup\$
4
  • \$\begingroup\$ I don't have a strong option for or against, but it is interesting to note that the question got community flagged for answer volume. If nothing else, that shows a hot topic worth pursuing. \$\endgroup\$
    – Joanne C Mod
    Commented Mar 1, 2015 at 19:32
  • \$\begingroup\$ Alright, I'm gonna do it while the topic is still hot. :) \$\endgroup\$
    – mattdm
    Commented Mar 1, 2015 at 19:33
  • 2
    \$\begingroup\$ We'll see how people take it. I suspect your reputation level can take the suffering if the response is poor and I'm beginning to wonder if these kinds of questions have more merit for us now. \$\endgroup\$
    – Joanne C Mod
    Commented Mar 1, 2015 at 19:35
  • \$\begingroup\$ I think I missed the news cycle — shoulda asked 'em both at once. People are so over it now, it seems. "Geez, what are you still on about that for? That was yesterday's news." :) \$\endgroup\$
    – mattdm
    Commented Mar 2, 2015 at 14:06

2 Answers 2

1
\$\begingroup\$

Yes, I will.

Adding filler text.

\$\endgroup\$
1
  • 1
    \$\begingroup\$ Too late. :) (more filler text \$\endgroup\$
    – mattdm
    Commented Mar 8, 2015 at 18:38
0
\$\begingroup\$

what I dont get is if people actually see white, or make assumptions that it is white and just looks blue - trying to second-guess the technical part of imaging.

It is a blue dress thats blue in the image. I could understand if it were white and imaged to be blue or vice versa - then it would be a "Haha" thing. But its just a bad image of a blue dress that gives it a brighter blue and a yellow cast.

\$\endgroup\$
4
  • \$\begingroup\$ That's my perception too. Yet, I had people swear up and down that it was definitely white in the image, and when pressed "well, sure, it's white with a slight blue tinge if you look at it right". \$\endgroup\$
    – mattdm
    Commented Mar 13, 2015 at 23:43
  • \$\begingroup\$ that is so weird. \$\endgroup\$ Commented Mar 14, 2015 at 17:22
  • \$\begingroup\$ See for example graphicdesign.stackexchange.com/a/49005/433, where the answerer argues "the 'true' colors in the image do lean towards gold and a very cool off-white in terms of the actual pixel values". (And see my comments to that answer.) Anyway, mind still blown (not by the image itself, but to the reactions). \$\endgroup\$
    – mattdm
    Commented Mar 14, 2015 at 17:55
  • \$\begingroup\$ No matter how much I stare at it I can't see even a blue tint. It just looks like a white and gold dress. \$\endgroup\$
    – Paul
    Commented Jan 5, 2017 at 5:36

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .