Matthew Cobb called my attention to this illusion that turns beautiful celebrities as ugly as a frog peeking through ice. Be sure to keep your eyes on the cross in the center!
From the YouTube description:
It won second place in the Best Illusion of the Year Contest, 2012!
It’s a new scientific finding called the “Flashed Face Distortion Effect”. [JAC: reference and pdf link below.]
From researcher Matthew B. Thompson’s website:
Like many interesting scientific discoveries, this one was an accident. Sean Murphy, an undergraduate student, was working alone in the lab on a set of faces for one of his experiments. He aligned a set of faces at the eyes and started to skim through them. After a few seconds, he noticed that some of the faces began to appear highly deformed and grotesque. He looked at the especially ugly faces individually, but each of them appeared normal or even attractive. We called it the “Flashed Face Distortion Effect” and wanted to share it with the world, so we put it on YouTube.
The effect seems to depend on processing each face in light of the others. By aligning the faces at the eyes and presenting them quickly, it becomes much easier to compare them, so the differences between the faces are more extreme. If someone has a large jaw, it looks almost ogre-like. If they have an especially large forehead, then it looks particularly bulbous. We’re conducting several experiments right now to figure out exactly what’s causing this effect, so watch this space!
Tangen, J. M., Murphy, S. C., & Thompson, M. B. (2011). Flashed face distortion effect: Grotesque faces from relative spaces. Perception, 40, 628-630 doi:10.1068/p6968
if you like that, check the Thatcher effect
http://en.wikipedia.org/wiki/Thatcher_effect
I’m amazed that this works for me, since I have monocular vision. I can use either eye, but only one at a time, and it generally takes an effort to use my non-dominant eye. So this isn’t the two eyes merging different images, this is the vision system (one or both eyes) merging two images. At least I’d be willing to bet that all you 3D folks could see the same thing with one eye closed.
Sure enough, it still happens with one eye closed. It really seems as though the brain is presenting a caricature! Facial recognition is pretty important for the purpose of identifying kin, and friends I suppose, especially since we don’t have such a great sense of smell. Maybe by flashing two faces we are forcing our brain to try to pick out the blood relative or the friend. By exagerating prominent features it tries to fit them to memories, in the same way political cartoonists make sure we know which president they are drawing. Gotta try this with a set of strangers, with family members, and other mixes.
Ohh, have to try this on people with prosopagnosia!
At the researcher’s website they do have an alternate video with random strangers, and the effect is just as pronounced. I also seem to notice, along with exagerated features (huge eyes, eyes reduced to slits, bent noses, shrinking or expansion of foreheads), that some areas become monotone in color, and that blemishes are noticable. Interesting to compare the google images for “eigenfaces.”
I have moderate prosopagnosia and I can see the effect. Couldn’t tell you who the faces were though.
Indeed we can – interesting!
Exactly. I have almost no vision in my left eye (from a car accident when I was 10) and I can still see this effect.
Cool! Maybe this effect could explain the weird “alien” faces people have reported in the past? Some of the larger eyed celebs look like those “little green men” sketches!
Everything turns ugly if you keep your eyes on The Cross
i see what you did there.
Or maybe it reveals the soul of the person in the picture … naw, I’m going with brain signal discombobulation.
+Trinity
I did not see it.
I did not see it.
wow. Definitely bizarre.
Can you make Karl Rove attractive using this??
Damn, that is strange! Since you can’t focus on any one image, it is tough to get get a highly distinct image of just what you “think” you’re seeing even though what you can perceive is really weird.
I’m not seeing anything weird about this.
Why does everybody appear so, cross?
(Best I could do on short notice.)
Yes! Very angry looking
Reminds us that our visual world is purely a construct of our neurons reacting to electro-chemical stimuli.
I would suggest that concentrating on the cross suppresses saccades, the rapid eye movements used to scan a scenes. I don’t suppose the suppression is complete but it could be enough to ensure that that much of the quality of the faces is “inferred” from what we have picked up from previous faces. Since the image has to be “complete”, it is distorted to fit.
Interestingly I have come across a couple of people who claim to see nothing unusual. It would be interesting to find out if saccades suppression was less complete among such people. For me it appears to “work better” the “harder” I concentrate or the bigger the screen, but that may just be me fooling myself while trying to confirm my hypothesis.
Incredible! Usually you would have to take drugs to get that effect.
I think I spotted Shrek at one stage.. haha
It reminds me of the cartoons drawn of politicians, which exploit the way our brains pick up on unusual characteritics of a persons face and exaggerate it for maximum effect
I guess that’s what is happening here as it’s slightly out of the field of vision so we’re homing in on unusual facial features and exaggerating them
Jim Carey looked the same as ever tho 😉
It also works with only 1 eye. Now will it work to distort 2 unlike objects next to eachother or does it only work if the 2 objects are either somewhat similar or have at least one obvious point of comparison on them (such as the eyes + nose in the case of the faces).
For me, they are not only distorted; they are symmetrical as well. Has anyone else noticed this?
Surely this effect is due to the blind spot in each eye which is where the optic nerves go to the brain. It only happens for a second while your brain filters the image, to compensate for the blind spot.
Great effect.
Of course the blind spot restricts what we actually see. But normally eye movements compensate for this.
Maybe Walt Disney used this effect to create his early moving cartoons —- for example the Ugly Sisters ! and the ungly angry stepmother.
Creators of the early moving cartoon figures
may have noticed this effect too.
It took me three attempts before I ‘got’ it. The first time I couldn’t stop myself trying to ID the celebs (I do recognise several of them, I think). The second time I was trying to compare the pairs of people in the photos. The third time I was staring hard at the cross (you really do have to concentrate on that) and the effect hit me, extremely strongly. Yes they’re bizarre. Huge staring eyes, is what I noticed.
And for me, it doesn’t need a pair of photos, it still works if I cover one of the pair and stare at the cross. But by far the strongest effect is with people with large eyes, which appear bizarrely magnified. Even people who I think look attractive, it happens to, e.g Halle Berry.
And, I can pause the display, look at a normal-looking person, then ‘zoom in’ on the cross, and the effect appears, though mostly, it’s the eyes that it happens to.
My guess is, the visual system is weighted to pick out facial features, eyes particularly, and this is what’s happening – in an impaired view of the face (because we’re not looking directly at it), the eyes get identified first and there isn’t time for full ‘scanning’ of the rest, so the system just presents what it’s got so far, which is the most prominent features.
An optometrist’s thoughts:
Blind spot? No, or it would only work on one side for people missing an eye – and we know upthread that that isn’t the case. In fact under this theory it shouldn’t work at all for normally sighted people, because the left eye fills in the right eye’s blind spot and vice versa.
It seems much more likely that this simply makes it obvious to us just how poor our peripheral vision really is. We can only see detail in a tiny part of our visual field, and create the perception of a broad field of vision by scanning constantly and piecing the details together.
This illusion stops us from using our central vision, and so only the grossest details register. For example, the small, half-closed eyes of some of the male faces fall below the threshold of perceptibility and disappear altogether, but the large, heavily made up eyes of some of the female faces cross that same threshold and override other, smaller features nearby, appearing much larger (and uglier) than they really are.
Incidentally, our peripheral colour vision sucks too. Stare at one corner of a room, stand someone in another corner just within your field of vision, and have them produce a brightly coloured ball the size of a tennis ball without telling you what colour it is (they should keep the ball still). Then try to identify the colour. So long as you keep your eyes still it’s impossible to do.
…and the high speed flashes could simply rob the brain of the time required to process the images sufficiently to normalise them. Even then the normalisation is only relative – for me, the opening faces that are held stationary for a while still appear slightly distorted if I focus (religiously! 🙂 ) on the cross and nothing else, at least compared to when I look at their eyes.
Having dutifully kept my eyes on the cross and seen the resultant uglification, I looked at the thing again, this time looking at the faces, not the crosses, and they struck me as being in the main just as ugly as they were the first time…
We marvel at the length of a giraffe’s neck, or the streamlining of a Great White’s bod. Over the eons evolution has produced even more magnificent extremes, only for such over-specialization to be relegated to an inexorably ballooning extinction list. But so automatic is the process of identifying a face that one rarely pauses to marvel at what a gargantuan computing task it is. The absolute amount of distance between width of nose and separation between eyes that helps us distinguish between various faces is mere millimeters. Yet the impression upon us of each face we know of is profound. It is no doubt a phenotypic expression of a species very social at its core. Oliver Sacks happens to be at the lower end of the spectrum of an amazing ability that we otherwise take for granted (see his New Yorker piece on it). I’m curious if this illusion spooks the facial recognition process in people like him as it does in most of us.
What I find extraordinary is the way we recognise people. Not just faces (which obviously can’t be just a ‘matching a template’ feature since no two photos of a person’s face will ever be precisely the same), but also cues from the way they move and the way they speak, extrapolated to match a pattern that we’ve never precisely seen before. There’s some massive pattern-recognition process going on (which is probably part of the reason why we’re so chronically ready to perceive patterns in any random phenomenon).