Is That Dress White and Gold Or Blue and Black?

HughPickens.com writes Color scientists already have a word for it: Dressgate. Now the Washington Post reports that a puzzling thing happened on Thursday night consuming millions — perhaps tens of millions — across the planet and trending on Twitter ahead of even Jihadi John's identification. The problem was this: Roughly three-fourths of people swore that this dress was white and gold, according to BuzzFeed polling but everyone else said it's dress was blue. Others said the dress could actually change colors. So what's going on? According to the NYT our eyes are able to assign fixed colors to objects under widely different lighting conditions. This ability is called color constancy. But the photograph doesn't give many clues about the ambient light in the room. Is the background bright and the dress in shadow? Or is the whole room bright and all the colors are washed out? If you think the dress is in shadow, your brain may remove the blue cast and perceive the dress as being white and gold. If you think the dress is being washed out by bright light, your brain may perceive the dress as a darker blue and black.

According to Beau Lotto, the brain is doing something remarkable and that's why people are so fascinated by this dress. "It's entertaining two realities that are mutually exclusive. It's seeing one reality, but knowing there's another reality. So you're becoming an observer of yourself. You're having tremendous insight into what it is to be human. And that's the basis of imagination." As usual xkcd has the final word. It would make the comments more informatively scannable if you include your perceived color pair in the title of any comments below.

Comments

White balance and contrast in camera.

[Z00L00K]

It's primarily an effect of the camera fooling around with the white balance and contrast/light to try to get a correct image and failed.

Add to it that the picture looks different depending on which display you have on your computer and you have a nice debate.

Re:White balance and contrast in camera.

[jrumney]

Not depending on which display so much, but with LCD displays, depending more on what angle you are looking at. Look at it straight on, and the dress is white and gold. Ask the person next to you, and they will tell you it is blue and black. Turn your screen towards them and the effect will be reversed.

Re:White balance and contrast in camera.

[drinkypoo]

When I see stuff like that, my brain corrects the image for me.

Are people who are seeing this dress as weird colors just defective?

(I'm defective in other ways, like a lack of modesty, and some respiratory issues, let's not get all twisted here)

Re:White balance and contrast in camera.

[itzly]

That doesn't explain why two people looking at the same photo swear they see totally different colors.

Re:White balance and contrast in camera.

[drinkypoo]

The thing is, in the provided picture, the dress actually IS white and gold, or at least grey and gold. Load it up in an editor and snip pieces of it out if you don't believe me, look at them on their own, compare them to color swatches. That doesn't make the dress any particular color. It makes the picture a particular color. The "white"/"black" part is banging right around 50%, which is clearly neither white nor black.

The camera diddled the image, maybe it was diddled even more before we actually saw it. Then we're all amazed that it doesn't look like the thing. But people have accidentally been taking pictures of things which don't look like things since time immemorial. They're called shitty pictures.

Re:White balance and contrast in camera.

[jpapon]

From any angle, it looks blue to me. Very distinctly blue. I'm actually somewhat baffled that anyone (nevermind the majority of people) perceives it as white and gold.

Re:White balance and contrast in camera.

[jpapon]

The thing is, in the provided picture, the dress actually IS white and gold, or at least grey and gold

I'm sorry, but it is definitely not. I just opened it up in GIMP, and the blue areas have hue values between 225 and 230. While yes, the saturation is low (30-40), that definitely still makes it blue, albeit a washed out blue.

Load it up in an editor and snip pieces of it out if you don't believe me, look at them on their own, compare them to color swatches

I did. It's definitely blue. Not highly saturated blue, but blue nonetheless. It's certainly not white/grey.

Re:White balance and contrast in camera.

[markdavis]

>"Not depending on which display so much, but with LCD displays, depending more on what angle you are looking at. Look at it straight on, and the dress is white and gold"

Well, in my case, when I look at the photo in any light, on any monitor, at any angle, at any time, I have and have always seen only light blue and brown/gold. There is no situation where it is either "blue and black" or "white and gold".

The question is what we see in the photo, not what the dress ACTUALLY is- we can't know that because all we are allowed to see is a [poor] PHOTO of the dress, not the actual dress. And it is obvious the camera white balance and exposure is way off, trying to compensate for something, resulting in a photo with a probably very false representation.

Re:White balance and contrast in camera.

[wonkey_monkey]

The thing is, in the provided picture, the dress actually IS white [...] or at least grey [...]

What's your definition of "IS" here? Because the only objective measure I can think of is to look at the RGB values, and pretty much every pixel in the "blue" area has a B value about 20%-30% higher than the R or G value. That, to me, makes it about as objectively blue as it's possible to be.

Re:White balance and contrast in camera.

[jrumney]

The black is the bit that looks "gold". The "white" part is actually blue. Once I realised that the oversaturated light source behind was a reflection of the flash in a mirror (you can see more evidence of flash reflection on the dress itself), and not the sun shining through a window, my brain started processing the image totally differently, and I can no longer see it as white and gold unless I scroll the image so I can only see the top 20%.

Re:White balance and contrast in camera.

[beuges]

This morning, I saw it on my phone in my darkened bedroom, and it was clearly blue and brown. Just now, I opened the Washington Post link on my 24" screen in a sunlit room, and it was clearly white and gold. I then found the link that I had seen on my phone this morning (not Washington Post, so I wanted to confirm that it just wasn't two different pictures that I was looking at), opened it up, and it was white and gold there too. Went back to my bedroom and closed the curtains, and it remained white and gold for a bit, but after I left the room (after my eyes had adjusted a bit to the darkness), it was blue and brown again. The picture on the Washington Post was also now blue and brown. Now that my eyes have adjusted to the sunlit room again (and the white Slashdot background), I switch back to the Washington Post tab, and it's white and gold again. My wife (who's now gotten fed up with following me around to look at this picture under different lighting conditions) has had pretty much the same experience as me.

So it appears to be linked to the lighting conditions that your eyes are adjusted to when seeing the image initially... even after they've adjusted to the ambient light, the brain appears to stick to the image it created initially.

Re:White balance and contrast in camera.

[itzly]

I did. It's definitely blue. Not highly saturated blue, but blue nonetheless. It's certainly not white/grey.

Agreed, but white objects in the shade under a blue sky are usually blue. Look at the white snow under the trees, for instance. http://khongthe.com/wallpapers...

Re:White balance and contrast in camera.

[Anonymous Coward]

I saw it as hot pink and emerald green.

Unfortunately, I made the mistake of telling that to my psychiatrist, and that's why I'm typing this from a padded cell.

Re:White balance and contrast in camera.

[Fnord666]

So it appears to be linked to the lighting conditions that your eyes are adjusted to when seeing the image initially... even after they've adjusted to the ambient light, the brain appears to stick to the image it created initially.

Here is a pretty good explanation of why this might happen.

Re:White balance and contrast in camera.

[grnbrg]

Put it into to Photoshop and eye-dropper the colours. They are quantitatively light blue and dark brown.

But they can perceived as either blue and black or white and gold.

Re:White balance and contrast in camera.

[Camel Pilot]

I am fairly convinced that those that claim the dress to be "distinctly blue" are just trolling the rest and extracting some weird sort of pleasure from it. The interesting aspect of this story is that there are sooo many people willing to troll others.

Re:White balance and contrast in camera.

[EvilSS]

So it appears to be linked to the lighting conditions that your eyes are adjusted to when seeing the image initially... even after they've adjusted to the ambient light, the brain appears to stick to the image it created initially.

Here is a pretty good explanation of why this might happen.

Something is wrong. You said "pretty good explanation" but you then linked to Gizmodo. These two things are mutually exclusive.

Re:White balance and contrast in camera.

[Baloroth]

Yeah, and that's kinda the whole point: everyone who looks at the photo is automatically (and completely subconsciously, without realizing it) applying color-correction to the dress, based on the brain's similar experiences with color-correcting and the visual clues in the picture. What makes the picture interesting is that it's so close to the edge between white/gold and blue/black that different people can perceive it differently, even on the exact same screen. Actually, I've seen it both ways, though I believe the picture that I saw as white/gold was ever so slightly lightened (based on a totally not scientific color picking of the blue areas). The picture was also a smaller version, which may have made the difference. The point is, the picture is a fascinating example of how what humans really perceive is not what they're actually seeing, but a heavily interpreted version based on context and various visual clues. In fact, humans would be effectively blind without that processing (imagine face blindness, but for everything you see).

Re:White balance and contrast in camera.

[serviscope_minor]

Well, the best explanation is that our brain does very, very hefty colour correction on its own. The reason is probably so we can ignore the irrelevant information, i.e. the light colour and any sensitivity lag between the different cones and identify objects.

If you've ever tried to do colour computer vision you will be *astonished* at how good we are at it.

Anyway, my guess is that the reason for this one appearing as two different things is it's very close to some threshold. It's possible to interpret it as more or less normal colours under a blueish light (i.e, sunlight), which makes you perceive the blueish grey as white. Or, you perceive it as blue under washed out much more orange light (e.g. incandescent bulbs) in which case you perceive the blueish grey as blue. The "white/blue" colour is objectively blueish grey if you check the actual hex values: it's mostly grey but the blue channel is a bit higher, but not by that much.

Our brains seem to model the complete co varying of the colours so once you decide what the light colour is, this effects the perception of all the colours. Black is nevertruly black and always reflects some light, so once the brain has decided the blue is super washed out by excessive light and that light is orange, that makes bright orangish things perceived as black since that's how black would appear under those circumstances. Given the rather good corrections of the picture, it appears those people are correct (though I'm on the white/gold side).

These two perceptions are very, very different and so it's kind of a binary thing, either you perceive one or the other. Now, I suspect that this photo is very close to the tipping point which is why some people perceive it very differently from others.

Either way, it's a fascinating example of colour interpretation.

This one is also excellent:

http://en.wikipedia.org/wiki/F...

Re:White balance and contrast in camera.

[The Rizz]

I can only see it as blue, period. Not trolling - I really cannot see this as white in any circumstances, even the XKCD "color balanced" bit I still see it as blue (albeit a much lighter blue on the left).

Maybe this has more to do with some effect similar to color-blindness? There are more types of color-blindness than just the standard red-green, and they have different effects on how colors are seen (and some kinds of color blindness can actually make you see other colors better). Maybe this dress has highlighted a previously unknown type of color-blindness?

Re:White balance and contrast in camera.

[serviscope_minor]

Well, I'd hesitate to call you colour blind since you are in fact correct. The dress really is blue and your brain is somehow undoing the mangling that's been done by the camera and lighting to arrive at the correct colour. I can't unsee it as white and gold however.

You can think of it all hinging on the blue/white stripes.

Objectively they're grey with a little blue in (check with a colour picker). All colour interpretation is ambiguous since you have do undo the effect of lighting and uneven colour sensors (eyes included). If your brain decides the light is white with a bit of blue, then that implies the underlying colour must be white and so the other colours must be gold.

If however your brain makes the call that it's strong orange light then that implies the stripes are in fact blue. That further implies the other stripes are dark grey, usually intrepreped as black.

There are other optical illusions designed to trigger this different intrepretation such as this:

http://en.wikipedia.org/wiki/F...

What's unique about this is it's not made to trigger two different intrepretations in one person, but instead by chance triggeres quite evenly the different intrepretations in different people, leading to big debates.

For you, the XKCD one isn't extreme enough to push you in different directions. Interestingly high level intrepretation has a big part. I can't cite anything but I can give a personal anecdote. The other day I saw an odd green thing on the floor in my hotel room. It was actually my backpack lit with green tinged light, but was crumpled in an odd shape so I couldn't tell what it was. When I figured it was my backpack, I could no longer see it as green (the actual colour is black). I probably spent 5 minutes staring at it trying to re-see the green with no success. It's a good illustration that the brain uses information from all levels including high level information such as recognising a specific object that I know is in the room to intrepret light colour.

Re:White balance and contrast in camera.

[rastos1]

Put it into to Photoshop and eye-dropper the colours.

Excuse me. Is GIMP good enough for that functionality?

--
On the internet nobody can hear you being subtle

Re:White balance and contrast in camera.

[Solandri]

I've been doing white balancing and level adjustments on my photos for over 20 years now. The dress as shown in the photo is white/gold (actually more of a light-blue/gold). You can confirm this with eyedropper measurements in Photoshop.

However, if you look at the sliver of background which appears to the right, you can tell the photo is badly overexposed. If your eye spotted this and your brain compensated for it by interpreting the pic as what might see if you stepped out into bright sunlight after being in a darkened room, the dress will appear blue/black.

Re:White balance and contrast in camera.

[Solandri]

This dress thing is actually a simplified version of a really powerful and freaky illusion which unfortunately only works in a darkened room. You take a quilt with different color squares. You can verify in white light that the different colors span the rainbow.

On top of this quilt, you lay a black piece of cardboard which has a cutout allowing you to only see one color patch. Say it's a greenish patch. In the darkened room with the single white light, you can confirm that this single visible patch is green.

Then you put colored filters in front of the light until the patch appears to be red/orange. The light is red/orange, so the green patch now appears red/orange. Makes sense right?

Then you remove the black cardboard so you can see all the color patches. And suddenly that red/orange patch appears green again. Cover it up with the black cardboard again and it'll appear red/orange. Remove the cardboard and it appears green.

What's going on (and also what's going on in the dress photo) is that your brain is using the information of surrounding objects (in the dress pic, it's the background visible to the right) to gauge the amount of light and the color of the light. It then corrects for the color of the light to white balance what you see, and for the brightness to compensate for over/under exposure. So the colors and intensity your eyes see remain the same, just how your brain interprets them changes.

Re:White balance and contrast in camera.

[Baloroth]

Everyone who looks at the photo will see the dress as *some* color. That's the point. It's not a conscious judgment, you see what your eyes/brain are telling you is there. Some people see it as white, some people see it as blue (and some people have seen it as both).

There isn't actually enough information in the picture to make that call.

But your brain does it anyways, because that's what your brain does.

Re:White balance and contrast in camera.

[tehcyder]

People seeing it for the first time, on the same device, in the same room for several hours, see different colors.

Your experience is not enough to draw any conclusion from.

If you don't see it as blue and black, you are clearly some sort of alien lizard-being.

Worryingly, that means that only about a quarter of the people around me are actually human.

Back-end image file manipulation?

[swb]

I'd swear I saw completely different images of these dresses posted, at extremes of the color controversy and neither was at all ambigous as to what color it was.

I wonder what the likelihood is that two or more images were served to clients, either at random or by some algorithm, to further the controversy? I can see one single ambiguous image that could go either way, but most of the examples I saw looked to be tweaked for maximum color association.

If you served tweaked images to clients so that "everyone" saw a different image, including people who saw different images at different times muddying their memory of what they saw over time, you could really amplify the controversy since people would actually be seeing a different image.

Re:Back-end image file manipulation?

[werepants]

Of course you could, but what's more likely - some elaborate scheme to create a viral controversy that's tied to no obvious material benefit, or a picture that just happened to be taken with a shitty cellphone that gets interpreted differently by different viewers? What's more, lots of people looking at exactly the same image, at the same time, on the same device (for instance, my wife and I) came to opposite conclusions.

The summary and linked xkcd comic do a completely accurate job explaining the phenomena, no conspiracy theories required.

Color means many things

[ganv]

This reminds me of the great entries from the competitaion to explain 'What is Color':

https://vimeo.com/87968614 http://www.centerforcommunicat...

By the way, I see white/lavender and brown. It would be very interesting to know what lighting/image manipulation was done to get those colors out of a dark blue and black dress.

Perception

[Dan East]

First off, the picture is crap. It's overexposed and the white balance is off by a mile. My 10 year old Razr flip phone took better pictures than that.

However, there's still a human perception factor going on. I had looked at the picture on my laptop, and it was clearly white and gold. Then later I pulled the exact same picture up on my iPhone to show it to someone, and it looked black and blue. I then concluded that the picture looked different on my laptop than my phone due to differences in the display. When I got back home I pulled the picture up on both my phone and laptop to do a direct comparison, and both, including on my phone, looked white and gold again.

So I think it depends on whether your eyes are currently adapted to dim indoor lighting or bright outdoor lighting, in addition to the backlight on your device also changing the hue depending on if it's automatically full bright for outdoors or dim for indoors.

Re:Perception

[qubezz]

I think that understanding photography and exposure is the key to recognizing the color in the picture - if you are familiar with photography, you can see that the background is light and almost blown out, and can use that as a reference for how the entire photo was lit when it was taken. However, if your brain doesn't process the context of the photo, and you evaluate it based on the blue background of a web site or a dark room around you, maybe you have the optical illusion that it is white. I can not unsee it as blue because I recognize the photo's lighting.

Re:Perception

[itzly]

However, if your brain doesn't process the context of the photo

I did notice the overexposed background, and I interpreted as a picture taken of a white/gold dress in the shade against a sunny background with a blue sky.

Re:Perception

[Kjella]

As such it's not the brightness, it's the color temperature. The spectrum of light ranges in color temperature from 1850K at sunrise/sunset to 6500K on an overcast day to 15000K under a clear blue sky. The eyes adjust to this, if you look at someone using a cell phone at night it'll probably seem to have an eerie blue glow as it has daylight color temperature. So with "nightvision" the dress looks blue/black, with "dayvision" it looks white/gold.

The people trying to read the RGB values to determine the "truth" forget that the color space assumes you have a D65 white point. Basically your LCD screen is trying to show you correct colors for overcast daylight. If you stare at the red sunrise/sunset or the blue sky for a while and then look at the LCD screen, your color perception will be off. Apparently this picture is in just the right sweet spot to confuse a lot of people.

Re:Perception

[zoffdino]

I'm a graphic designer and photography enthusiast. I'm typing this on a NEC PA242W color-calibrated monitor, in a near-dark room. That dress is white and gold. the white part has a blue tint but I wouldn't call it blue. The colors look the same on my iPhone 5S. When I bring the iPhone outside, the blue tinge is more apparent (a short of light sky-blue) and the gold/brown turn darker, somewhat into the black category.

The different isn't in the screens, it's in your eyes, caused by environmental light. A sunny day at noon can be 100x brighter than even a well-lit room with floor-to-ceiling windows. If it's sunny in your location right now, try this: find a view point where you can frame both the sky and a patch of dirt land (no grass or foliage). Put the camera in manual mode, pick a shutter speed, manual daylight white-balance (6500K) and a low ISO, start with a large aperture (like f/4) and gradually step it down (like f/22). The sky will appear more blue and the ground will appear darker.

That's exactly what our eyes do. In darker places, the pupils open up to accept more lights, the highlights (blueish-white) gets push up to white but mid-tones and shadows are preserved. In bright places, the opposite happen: lower mid-tones and shadows are pushed to near-black, highlights are pulled down to reveal the blue accent.

Re:Perception

[Lunix Nutcase]

That dress is white and gold. the white part has a blue tint but I wouldn't call it blue.

No, it's not. The original picture is overexposed and washes out the colors as shown in the picture I linked which contains a second picture showing the actual colors.

Re:different from Cornsweet

[itzly]

I have exactly the same, except I can not get it to look blue/black.

sigh

[s.t.a.l.k.e.r._loner]

This "____-gate" shit needs to stop. Now.

Comparison over multiple screens

[MassEnergySpaceTime]

Earlier tonight, I compared the same picture on the same site using 3 different computer monitors side by side and 3 different tablet screens.

To me, the white/blue part of the dress is sort of a pale light blue on all 6 screens. But they're different shades of pale blue. On one screen, the blue stand outs a little stronger. On another screen, the blue seems more faded towards white.

For the black/gold/tan part of the dress, on some screens, the tan color seems more faded, making the darker part stronger, and I COULD call it black. I know it's not PURE black, and it's not as black as that cow patch thing in to the left of the dress. But I could call it a shade of black. On other screens, the tan part stands out more, and I would definitely not call that part black. I don't know if I would call it "gold", but I would call it tan/light brownish.

So I think the screen settings is one variable that contributes to what colors the user thinks they see in the dress picture.

For the situations where different people are looking at the same screen or printed photograph, my guess is that the variability comes from the color/brightness/etc sensitivity of their eyes. For example, in my own eyes, one of them sees the wall in a brighter shade of white (and possibly slightly red tinted) than the other eye. Perhaps those who aren't as sensitive to blue might see the blue/white part of the dress as a shade of white, and call it white.

I guess this picture is just one of those freak pictures where the colors are at some borderline that could be interpreted as one shade of color or another.

Slashdot can do video, but not an image?

[wonkey_monkey]

Couldn't Slashdot post the image in the summary? I mean, it's not even like it'd be an illustration; this article is literally all about a specific image.

You can do it with annoying autoplaying videos, so why not a simple JPEG?

Re:XKCD shows you

[Sesostris III]

The BBC also show the actual dress:

Optical illusion: Dress colour debate goes global

I see white and gold, although the actual dress is blue and black.

Color Illusion

[physicsphairy]

The XKCD plot just makes me see gold and white at different levels of brightness. But I did find this color illusion featuring yellow and blue. The dogs are actually the same color, which you see if you look at them individually through a small aperture
http://i.imgur.com/sh5NwCK.jpg

Make it pretty obvious that at some point your brain switches from wanting to see blue to wanting to see yellow based on the color context. It would appear some of us are slightly different in where transitions like that occur.

Dressgate??!!!

[Anonymous Coward]

Dressgate? Seriously? Just kill me already.

Actual Dress Colour

[XB-70]

The actual colour of the dress is "No, That Dress Does NOT Make You Look Fat."

Re:Stupid post

[dcollins117]

If it makes you feel any better about this issue appearing on a tech site, it's part of what makes computer vision hard. Colors change under different lighting conditions so how an algorithm treats color information when identifying an object or analyzing a scene is an interesting problem.

Re:Frankly, my dear, I don't give a damn

[serviscope_minor]

Then please hand in your nerd card on the way out along with the idiots who marked this "insightful".

How the brain perceives colour well is fascinating, and our subconscious is amazingly good at it. The "gold" colour is objectively gold (check in the GIMP), and yet some people's brains manage to correctly interpret it as black.

If you're not interested in that and the underlying algorithms behind how our brain works, then frankly I wonder what the hell you're doing here.

Or possibly you just think you're too cool to be interested in popular stuff, in which case, grow up.

W&G, was Re:White balance and contrast in came

[Anonymous Coward]

Here is a pretty good explanation of why this might happen.

Why is it that my mod points always expire right *before* I want to use them? I used eight of fifteen purely on posts where I only sort of wanted to mod. I had to give up commenting to do so. Here, I log in so as to mod -- and my mod points are gone. And I have no real interest in commenting!

Grrrr.

Anyway, regardless of the general quality (or lack thereof) of gizmodo, this was a decent explanation. It points out that in the picture, the colors are pale blue and dark gold. However, the original dress is a darker blue and black. The colors in the picture are incorrect. People who see it as blue and black are seeing past the problems with the picture while those who see white and gold are being fooled by the bad colors in the image.

Actual dress is the blue and black one on the left in this picture: http://media.gotraffic.net/ima...

Re:Can someone please answer

[smooth wombat]

I said elsewhere that this is a scam for the following reasons.

Her Tumbler account (Swiked) shows the initial photo asking what color is the dress. A day or so later she posts a second picture of someone wearing the blue/black version, stating:

this is the dress as i saw it on the day of the wedding. blue and black. It's just that one photo bUT it's so weird???!??

Here's the thing, she does not say the second photo is the same dress ON THE DAY of the wedding, only that the photo shows what she saw.

What most likely happened was she took the first picture on a different day and because it's white/gold, it took on the cast of the lighting to give a blue hue and her friends had a disagreement over the color which started the whole thing, especially if someone is red/green colorblind.

Second, if you look at the top edge of the dress in the first picture you can clearly see what would be considered "virgin" light, i.e. light which is not reflected but directly falling on to the dress. If the claim is that our eyes are fooling us because of the bright background or because we can't be sure what the lighting situation is, this light should produce a different color under it, just as light shining across a rippled surface can produce different colors.

Except it doesn't. There is no transition from the slice of light to the shaded portion. It's one continuous tone. In fact, as the picture shows, there are multiple shadows of different angles and lighting conditions which should produce different colors, but they don't. For instance, under the cape/shawl toward the upper right, the area under the shadow is darker but not a different color.

Further, if our eyes are being deceived by the bright background, covering up all but a small portion of the dress should reveal the true colors because then there would be nothing to confuse us. Except that doesn't work either, the dress stays as white/gold with a blue hue.

And finally, if you look at the edging in the middle of the picture then down to the lower left corner (our left), you can see shadows under the gold edging. If the dress in the picture was blue/black you would not see such distinct shadows as are shown in the picture.

So, the dress in the original picture IS NOT the same dress she shows a day or so later but is the same style.

One final thought. The opposite of blue on the color wheel is gold and the opposite of black is white. However, if you look at what is blue and black in the "good" picture, they are not the correct parts. The bodice is blue which means in the reversed portion the bodice of the original picture should be gold and the edging should be white. Except that is not what is shown. The bodice is white and the edging is gold.

And for the record, I am not red/green colorblind. I pass all the Ishihara color tests without issue.

Re:Can someone please answer

[serviscope_minor]

I said elsewhere that this is a scam for the following reasons.

Except a good chunk of slashdot, absolutewrite and a few other also completely unrelated forums and IRC channels are in on the "scam" and have a bunch of people who have joined the conspiracy to pretend it's blue and black. Or white and gold, in which case I'm in on the scam and hereby declare I got my note from a shady black vehicle with blacked out windows this morning at precisely 5:50am at the dedicated drop point.

It's not a scam, because it frankly doesn't matter what the original colour of the dress is.

The interesting thing is it's sufficiently close to some average threshold of human perception that nearly half the population perceive it as completely different from the other slightly-more-than half.

At that point it wouldn't actually matter if it was a 'shopped image of a dress covered in purple unicorns.

Just making shots here...

[kefalonia]

The eye pupil is known to exhibit interesting behaviour at times,
http://en.wikipedia.org/wiki/P...
one notable being photic reflex (which also affects a quarter of a population)
http://en.wikipedia.org/wiki/P...

IMHO, human vision is still incompletely understood at whole population (global) level,
with all sorts of exceptions and special trade-off cases being documented:
http://en.wikipedia.org/wiki/T...
http://discovermagazine.com/20... ### check this one!

Finally, let's not forget, that it is well known that manly colour vocabulary is 4-bit, while females have true colour sets ;-O
http://www.smithsonianmag.com/...
http://io9.com/5919311/some-wo...
https://www.google.be/search?r...

Last but not least: make sure you see the image of the OP in fractional ways (say, top 10th of the image),
along with another person that sees it in the alternative mode. You may come up with surprises. ;-)