This story is being featured as part of our “Yahoo Best of 2015” series. It was originally published on February 26, 2015.
It was the mind-bender that blew up the Internet: the mystery of the dress colors.
It started Thursday with a Tumblr post by Swiked. She posted a picture of a two-toned dress. “Guys please help me,” she wrote, “is this dress white and gold, or blue and black? Me and my friends can’t agree and we are freaking … out.”
Within hours, the image went viral.
It’s been driving people crazy. What colors are the dress? Is it blue with black lace, or white with gold lace?
You see one or the other, and you probably can’t believe anyone else would see anything different. Each person is certain that his or her perception is correct.
So what’s going on?
Theory 1: It’s a hoax.
I see blue with black. Anyone who claims to see white with gold has got to be trolling me. Right?
Wrong. Because it’s easy to find somebody you know, somebody you trust, somebody in your own family, who sees the photo with the “wrong” colors. For example, two of our own Yahoo Tech writers — straight shooters, truly good people who would gain nothing by humiliating me in public — swore to me that they saw the dress as white and gold.
Theory 2: We’re using different kinds of screens.
Do phones and laptop screens show two different things? Do different kinds of LCD panels show different colors?
No. Because in many, many cases, two people sitting side by side, looking at the same screen, have seen different colors.
Furthermore, many people have seen a different pair of colors when they’ve looked at the same photo on the same device later. Some say they can force their brains to go back and forth!
(This tweet from @Tmoldovan sums up the frustrating bafflement: “Not a hoax, not a trick. Not brightness, nor emotions, nor angle. My kids see it as black and blue, my wife and I white and gold.”)
Theory 3: It’s science that you don’t understand.
It wasn’t long before people began passing around technical articles about color perception, such as this page of color-perception illusions and this article from livescience.com. (“These receptors, called melanopsin, independently gauge the amount of blue or yellow incoming light, and route this information to parts of the brain involved in emotions and the regulation of the circadian rhythm.”)
That’s the explanation. Right?
Wrong. Because if people identify colors so wildly differently, we’d have arguments about color names all the time. It wouldn’t take some stupid photo of a dress in 2015 to reveal this phenomenon.
Theory 4: Half of you are just nuts.
People soon started using Photoshop’s eyedropper tool to identify the colors of the photo for sure. This site, for example, “proves” that the colors are blue and brown by using Photoshop to sample the stripes.
Also, it didn’t take long for someone to hunt down the actual dress. As in, for sale, from a British store called Roman Originals. It is, in real life, blue and black:
That’s it, then! Mystery solved! Right?
Wrong. Because we’re not talking about the JPEG image, and we’re not talking about the original dress; we’re talking about the photograph of the dress. And that means that factors like white balance are at play. “The backlighting throws off the sensor” of the camera, explains one commenter.
Theory 5: Your brain might be compensating for shadow.
On Reddit, ChrisConlon offers a tantalizing theory that avoids the pitfalls of the others listed here: Some brains assume that the photo was taken in shadow, and other brains don’t. He refers to the diagram shown here (created by Claire Hummel). You see what’s inside the shaded rectangle, and your brain infers that the dress looks like what’s on the outer portion:
“I’m pretty sure this is an illusion similar to the checker-shadow illusion,” he writes. “Some people are seeing it as a white and gold dress in a bluish shadow; others are seeing it as a black and blue dress in a yellowish light, as seen here.”
(The checker-shadow illusion is this one, where square B looks much lighter than square A, thanks to the “shadow” cast on it. In fact, both squares are the same shade of gray. You can prove it to yourself by covering up all but a tiny bit of each square.)
Over at the official Flickr blog, resident data scientist Bhautik Joshi gives a similar take, explaining how the human eye adjusts to what it’s seeing according to either indoor or outdoor lighting. This means two different people looking at the same picture of the same dress may experience opposite on-the-fly vision correction and perceive different color arrangements.
Now we’re getting closer; read on.
The colorblindness theory
In an effort not to lose the last of my marbles, I wrote to Dr. Stephen McLeod, chairman of UCSF’s ophthalmology department. He gave an explanation that nobody had yet offered:
“I think it’s really just a sensitive test of red-green color deficiency, which is pretty common. I’ll bet most of the people calling it black are men. It’s consistent with cone deficiency, and red-green would do it for this hue.”
That notion makes a lot of sense, actually; the dress-color test works exactly like those Ishihara colorblindness tests, where some people can see a number hiding among the dots, but people with red-green colorblindness cannot:
“If you go to the gold/black area [of the dress photo] and blow the image way up, people who see gold will see that the clusters of pixels are basically reds (pinks) and greens (olives),” McLeod went on. “They merge to give the summated color. If you have red-green deficiency, you’ll see just black.”
He’s right on that last part. I’m red-green colorblind. I see black in those areas. And I cannot see white/gold in the dress image no matter how hard I try.
If McLeod’s theory is true, you’d expect that many more people would see white/gold than blue/black — and, in fact, that’s what the numbers show. In a BuzzFeed poll of more than 1.2 million people so far, 73 percent said they saw white/gold.
But wait: 27 percent of the poll respondents saw blue/black in the dress. However, only 7 to 10 percent of the male population has red-green colorblindness. So what about the rest of the blue/black-seers?
And what about people whose perception switches back and forth between white/gold and blue/black?
“I’m guessing that across the population, the phenomenon is a combination of red-green deficiency and context,” McLeod says.
In other words, there’s a second factor at work: the context of the lighting and time of day.
Color-extraction theory says that our brains correct color according to the visual context. As Wellesley College neuroscientist Bevil Conway told Wired: “Your visual system is looking at this thing, and you’re trying to discount the chromatic bias of the daylight axis. So people either discount the blue side, in which case they end up seeing white and gold, or discount the gold side, in which case they end up with blue and black.”
All the experts agree that this is one of the most intriguing color-perception puzzles to come along in quite a while. It works only because of these particular colors in this particular photograph: a freak of photography, light, and color.
The experts also agree that it’s not an especially beautiful dress.