Twitter

Instagram View Richard Barrett-Jolley's profile on LinkedIn

Saturday, 28 February 2015

The Emperor's New Dress: #thatDress: A Case of Crowd syndrome?

The Emperor’s New Dress

If I were a clever writer, sadly I know I am not, I would write a parody blog providing scientific neurosciency explanation for why the King’s New Clothes do actually exist and simply cannot be seen by some people, but can be seen by others.  The kind of twine that is used is of a sufficient luminosity that some people simply fail to see it and despite the fact that he is really properly clothed, to some people he appears butt naked.  But Emperor's New Clothes is a very old phenomenon and that is exactly what has happened here.  I would have been the little boy that cried out from the crowd.. “but he’s naked!!”… and, I daresay, would have been dragged off and beheaded.  Hopefully those days are over, because here is my piece on #theDress #DressMeme.

Yes OK, I confess, I’m afraid this is yet another blog about the dress. I find whole phenomenon fascinating.  It is hardly surprising that someone who teaches the neuroscience of the eye, the visual cortex, rods, cones and opsins etcetera finds this phenomena interesting. However maybe you will be surprised at my perspective on the whole dress meme thing.
So let’s start with a summary of the facts (I think this is how Clouseau would begin in the Pink Panther movies?):  There is a dress apparently made by Roman that is black and royal blue. But a photograph of this dress has appeared where the colours are distorted and it looks gold and white or gold and pale blue.  Some people looking at the golden pale blue dress photo say that they see a black and royal blue dress. …and the internet has gone wild.  Scary.

There is also some confusion.  The question is not what colour is the dress, we know that in the high street store it is blue and black and it has just shifted colour to produce a funny coloured photo.  The squabble is that some people say that this specific picture looks blue and black.  There is no mystery about how digital photos of objects look different to the original, colour balance, colour temperature, exposure all explain that merrily.  The question is why some people say the gold looks black in this specific photo.




In response to this one scientist after another has trundled in to talk about perception differences between people, about the science of rods, cones, opsin proteins, wavelengths of light, processing in the visual cortex etc. but none of these “scientific” explanations explain the phenomenon.
My line is that the whole thing is nonsense.  It is probably a case of the “King’s New Clothes”.  There are a number of possible explanations from hoax to hysteria, otherwise known as “crowd syndrome”. If it is a hoax then how many how come so many people are a part of it? Well they aren’t really. My explanations are of hysteria, gullibility, or possibility sense of humour coupled to an innate sense of mischievousness in humans.  My only amazement is how few people can see the obvious flaws in the logic of the so-called scientific or visual perception anomaly theories.

Now why am I so confident that this is not a phenomenal about rods and cones and differences of perception? Because you can analyse the picture in heavy-duty imaging software and just get a simple answer on the RGB (Red, Green, Blue) levels of each portion of the picture. It’s true to say that when I see purple what registers in my brain maybe quite different from what you see when you see purple, but the problem here is that on a computer screen the world is less ambiguous.  “Black” in a digital photo means something quite specific.  It means that the RGB levels are either 0,0,0 or just close to this. For info, a typical digital photo consists of an array of pixels each with an RGB colour mix and each R, G and B value has to be a number between 0 and 255.  If they are all 255 that is pure white (255,255,255) and if we have black we have (0,0,0).  Of course (1,1,1) looks black and it gets less black as you move further away from (0,0,0). So there are a number of points around the RGB envelope that look black.  For example, below I have shown two very dark swatches. To me 40,40,40 looks pretty black. However, here’s the rub; we can analyse the gold in the photo and what do we get?
The figure shows the photo in question: and alongside it a swatch (A) coloured in the same colour OBJECTIVELY measured from the square using ImageJ software (NIH).  B and C are just examples of darker fills.  Below are three panels with the red, green and blue components separated, again in ImageJ, and the levels histograms from these presented below each.  For the record, the "blue/white" part is (126, 137,178).  There is no black guys.

So perhaps some people see 133,099,060 as black (“A” in the figure)?  …and always have… OK that’s possible, but that would mean they always would have done.  But millions of people have been viewing millions of gold dresses for years and how come no one spotted that gold looks black to some people before? It’s implausible. …and if it is true, well they are just colour blind in someway not observed previously.

Frequently optical illusions do occur, for example, running one colour, alongside another colour can give create an illusion that the second colour is different to its true colour. But you can “work it out” using objective analysis software (ImageJ et al.) as described above. Example: In theory… the red squares on the left look darker than those on the right? 





But are they? Simply analyse them in ImageJ:



Left:


Right





Histogram analysis shows that the mean colour intensity is 180 in both cases.  They really are identical. 


So since the digital photo clearly has no black in it what are the other explanations:
  1. Confusion. People have seen that the original dress is really black and they are making the classical exam mistake of NOT answering the question. The question is not “what colour is the dress?”, the question is “what colour does the dress appear in this particular digital photo?”.
  2. Crowd syndrome.  People are going with the flow.  Some people will be prepared to say what they think they should say, what they think the crowd expects them to say, to an extraordinary degree.  If we have an interview situation about this dress, with the interviewer asking what colour the guests see, s/he will be embarrassed if no one says black.  So some people will inevitably just have to throw a life line by saying it looks black to them, even though it doesn’t.
  3. Sense of humour.  Well we are all having fun aren’t we, and that will require some people to perpetuate the myth by saying that the gold parts of the photo appear black to them.  Or the fun would stop.
…but visual illusion it isn’t :-)
Of course if 10% of sighted population do see 113,99,060 as black they will think this a really embarrassing diatribe.  Such is life.  I’ve done worse. There was even that one time……





17 comments:

  1. David Weinkove1 March 2015 at 21:43

    I hear what you are saying but I don't think you acknowledge what is happening here. The people like me that say it is blue and black, can see that the black is more gold than black and that the blue appears closer to lilac. The amazing thing is the majority of people think this dress is pure white without any blue at all, with bright gold trim - completely different from what the blue-black people see. I recognised this because I once (and only once) saw the dress as white/gold - then later it went back to the lilac-dark gold. And I have seen blue-black on the same screen as someone who sees white-gold so it is not about image quality/brightness etc.

    ReplyDelete
    Replies
    1. Thanks so much for your comment!!! Please I ask you a question? If so... Can you see any difference between the black of panel A and the black of panels B and C?
      Thanks again
      R.

      Delete
    2. David Weinkove3 March 2015 at 12:15

      I can distinguish them easily and only C appear black. A appears golden brown, B as grey

      Delete
  2. I think it's not a case of confusion, crowd syndrome or sense of humor.
    It is a difference in perception, not due to the viewer's different wavelength sensitivities, but rather pattern matching of the brain.
    Here's my argument why:

    I've been "oblivious" to this viral picture till today when I found it on Wikipedia.
    I was just starting to read the encyclopedia entry after glancing at a clearly black and blue dress.
    It is at some point I realized I had been exposed to the image, on the side of my news feeds, but dismissed it as an ad for dresses.

    But I distinctly remembered the image as being white and gold, so I turned to my news sources to confirm.
    Lo and behold, the white and gold dress was there.
    A clearly white and gold dress, so I suspected it was different image from the Wikipedia article.
    Turning back to Wikipedia I could no longer see the blue and black dress anymore.

    I've studied some color composition and perception books, so I ran the different hypotheses through my head:
    -the surrounding environment(my room's lighting, the colors of the webpage featuring it)
    -the colors of yellow and blue are complimentary in our perception(perception is dominated by what you recognize first - we usually stick/are stuck to our first perception)
    -the poor/over exposure makes us mentally process the image(overexposed blue makes it a washed out grey, so does underexposed white)

    I can now see the dress in both ways.
    Just like that turning ballerina image.
    It takes some effort to do so though.

    If I focus on the lower part of the dress, the darker lines which lie in proximity become black.
    The lighter lines become blue by comparison.
    Then, looking at the big picture, I can see the black as being washed out at the top, acknowledging it's a very overexposed picture.

    If I focus on the top at first, I see that large golden part.
    There's no way of seeing it as black now.
    As I look below it, the lines and the shoulder to the side, being lighter, are white.
    So looking at the big picture, I see the white and golden dress, darkened in perception by the fact that it's the side opposing a powerful light source.

    Regarding your question to David.
    I could only consider the C as a "black".
    A is dark gold, B is grey.
    C is not a perfect black, the truly black elements around it make it clear.
    The R,G,B values on the side might have also influenced my perception though :)

    ReplyDelete
    Replies
    1. Thanks for taking the trouble to commenty... But just to clarify.... You can see black on this image of the dress within the box on THIS page if you look at it in the right order as you describe?

      Delete
    2. Yes, on this page as well. Never a pure black. Always washed out

      Delete
  3. This post suggests that you were not in a room with multiple people when the photo was shown.

    The phenomenon I experienced was that someone showed me the picture, without first telling me the color of the real dress, and to me the photo looked unquestionably to be of a blue and black dress. Meanwhile, someone else was shown the picture in my presence, also without knowing the color of the real dress, and answered that it was unquestionably a white and gold dress.

    A group of people all viewing the photo together were clearly, distinctly, and even somewhat emotionally divided nearly half and half over the color. Everyone agreed, after being shown the other pictures, that the real dress was blue and black, yet about half remained indignant that they were not able to see blue or black in the first photo.

    This doesn't sound like confusion, crowd syndrome, or sense of humor to me (nor were any of those remotely evident in the room).

    I have seen the color analysis breakdowns, but it remains the case that several of my friends and family still emotionally attest to not being able to see any blue in the photo (or even though isolated color swatches, which they perceive as neutral gray), while I and others in the group am still not able to perceive the dark stripes in the dress as being made of detectably gold-colored fabric. I realize that the color you (and others) sampled is indeed a very dark gold, but my brain insists on interpreting its occurrence in the photo as semi-glossy black cloth showing a silky luster in warm-toned (yellowish) ambient light.

    ReplyDelete
    Replies
    1. Thanks! Frankly I don't think I ever find myself in a room full of people without any humour present though!!! ...thanks for taking the time to comment. Live long and prosper!!
      R.

      Delete
  4. This comment has been removed by the author.

    ReplyDelete
  5. I thought it was Crowd Syndrome as well and that it was obviously gold and white. I did the photoshop analysis as well and found a few strange things. When you inverse the colors (Ctrl-I) I had expected the white to turn black and the gold to turn to some other medium color (blue, perhaps?) ... I was shocked to see that the "white" turned gold, and the "gold" turned white, which means to me that the "white" wasn't nearly white at all, and that the "gold" was dark enough.

    Plus my 7-year old daughter says that it's blue and black, without previously hearing of it before.

    Wayne

    ReplyDelete
    Replies
    1. Thanks... No the White is clearly not 255,255,255... But I havent measured that. I wouldn't really know what to expect from inverting 113,99,60 in Photoshop. I think there are actually a few different algorithms for that!
      We should probably skip using subjective terms like "gold" and "blue" and just use colour profiles. The song "Woman in Red" can be "The Woman in 250,0,0".

      Delete
    2. I meant to say the "white" is clearly not 255,255,255.... Sorry. I'll measure it next time I have my PC up.

      Delete
  6. Interesting points. I had a debate with an artist friend who is steeped in color theory. He mostly expressed annoyance an the whole thing and felt that there really was no interesting visual phenomenon here at all. To him it was not a neuroscientific conundrum but merely the result of people's mental laziness and almost-willful inability to penetrate beyond their first impressions. (And not even a particularly interesting instance of it since this sort of thing happens in all domains of life.) He made the example that in many lighting situations snow appears blue but that it would be a chore to get people to recognize what their own eyes were telling them because they just *knew* that snow was white. I am skeptical of this for the same reason that David Weinkove has stated -- that people are insisting not only that the dress appears white but that they can't see any blue in it. It is rather bizarre to me that anybody would interpret this even as a white dress that had been color shifted toward blue in the photographic process, but I could at least accept that. What I find utterly incredible is that people would say they can't see blue in it, when to me it is (generalizing here) a medium-toned Tufts blue. (The portions of the dress that could on their own be characterized in my mind as possibly white are only in the highlights, roughly 5% of the photo.) Similarly, while I can accept that the majority of people might initially say that snow in shadow appeared white to them, I can't accept (without evidence) that the majority of people would completely resist the notion that there was at least a significant amount of blue light coming off of the snow. Your position is somewhat closer to my friend's, and I disagree with both of you, but one area in which I think we all can agree is that the questions that are being asked are fairly sloppy. "What color is this dress?" is open to a great deal of interpretation and also primes people to get defensive about their answer. I don't think we'll make progress in figuring out what, if anything, interesting is going on here until we begin formulating the question in more rigorous ways. Unfortunately at this point it will be hard to find people who haven't already been primed.

    ReplyDelete
  7. Thank you so much for your comments. Actually the blue/white thing? That doesn't excite me all that much! ...(126, 137,178) appears a little different to me depending on the ambient lighting and the screen brightness etc. My issue is whether people seeing 113,99,60 together on the same page at the same time with ordinary brightness and their eyes fully open ;-) see differentially "black" or "gold". To me, to add to my silly blog above, it is also a strange co-incidence that this viral phenomenon surrounds a corrupt picture of a dress that does have black banding in the flesh. If it was just a "gold" dress that some people suddenly saw as black and then suddenly gold I would be less skeptical. But remember there are lots of properly exposed jpeg images of this same dress registering its native "dark" colouring (R.G.B) 5,4,6. That is a coincidence isn't it? So different versions of the same dress appear online at both 5,4,6 ("black") AND 113, 99, 60 (ambiguous/black/gold) and THIS is the image the world believes is seen in different colours?
    Now returning to the "blue". Its pale blue to me. When I first saw it, at night, on a dog walk on a mobile phone, in the middle of an un-illuminated park (Liverpool is a very safe city!) I thought it was white, then when I saw it in day light I could see it was pale blue. Not a big deal. But you know what, you have cheered me up so much you can't believe it!! I literally thought I was the ONLY person on the entire planet that believes this is just viral madness and not an amazing visual processing phenomenon... but you actually know someone else that thinks that too?! Wow!! Say "Hi" for me. So now its 2 of us vs 7 billion or so!!!

    ReplyDelete
  8. Why the funny font? Is this my fault? How to I fix it? I feel the credibility seeping away!

    ReplyDelete
  9. i love reading this article so beautiful! great job!
    Eid Collection 2017

    ReplyDelete

Be Nice!