How and why does this optical trick work?

First, look at this, making sure to look at the center like it says:

So, how and why does that work? What’s happening in me eyes/brain combined with those pictures that produces that effect?

I’m not seeing anything particularly out of the ordinary. Could be just me, though, I also can’t make out the 3D images in the Magic Eye pictures.

Please describe whatever the “effect” is that you’re asking about.

If you focus on the cross in the middle, the facial features take on weird proportions.

I can’t give a detailed explanation. It looks as if features from the different images are getting combined with one another in the wrong way. However, I cannot tell whether it is always features from the two simultaneous pictures that are getting confused, or if (always or sometimes) it is features from the previous picture getting mixed up with the current one.

Anyway, it is a weird effect, and a nice illustration of how visual perception is a much more piecemeal, less holistic process than is usually realized, analyzing a scene feature by feature instead of taking a picture, all-at-once, into the brain. It also shows what a crucial role unconstrained eye movements play in normal, accurate perception.

I thought so, too, but if I close one eye, and stare at the cross, I get the same effect. I’m thinking it must be something different, but not sure what.

Cover the cross and one side with your hand and look at the other. The pictures are distorted via some photoshop effect. Some subtly, some not. They start out normal and get weirder as you go.

So it’s not an “illusion” but more of a “trick”.

You do not have sharp visual acuity at the area of peripheral vision, and so, your brain takes whatever information it gets and makes a best guess. And even if the best guess is off, you hardly notice in normal daily life since it’s always updating.

For the illusion above, the flashing pictures at the periphery give too much contradictory information for a realistic representation, and so, your brain makes distorted images based on a jumble of images.

Even if you hit ESC to stop the pictures from changing, block out one side from view and view it with just one eye on the crosshair, you’ll ‘see’ that the image in the periphery isn’t in focus and becomes distorted after a while (as long as you don’t glance over at it and view it directly).

Nope. Not at all.

Hit Escape to stop the gif and you’ll see each celebrity’s image is normal. The illusion still works with one picture changing and with one eye.

I agree that they seem to be manipulated, but I’m not quite sure how. I think that it’s some sort of mucking about with the power spectrum, similar to those images that turn from Marilyn Monroe to Einstein when you get close to them.

Ah hah, ^that^ sounds right to me! So:

By not looking directly at the images, it throws off the lizard people’s holographic camouflage technology that all those famous people use to hide their reptilian features.

You may think you are seeing all things within your cone of vision at all times, but in fact your brain is only actively refreshing a very narrow part (what you’re focusing on) and filling in the fringes with what it “thinks” is there.

Since the pictures change frequently, there’s a little lag in this memory, so your brain is giving you a fluctuating mix of those things that are on the fringes of your focal point. Your brain is optimized for visual tasks related to faces, and knows that those images to either side are supposed to be faces, so it composites what it thinks/knows should be there into something resembling a face. This lag and best-guess compositing leads to the grotesque imagery you “see”.

Facial recognition is baked deeply into our brains by evolution, otherwise we wouldn’t be here. All your caveman ancestors who couldn’t quickly distinguish a friend or mate from an enemy or tiger got screened out of the gene pool early on, so we’ve come to excel at seeing and recognizing human faces reflexively without expending much mental effort on the task. As a side effect, we make sense of random stimuli (like the images at the fringes of your visual focus in this example) by assuming that they’re faces. It’s the same reason people who want to see Jesus often find him burned into their toast, but never see a house or a 1981 DeLorean appear there.

I’d be curious to see a similar demonstration using common objects that are not faces. I’ll bet the effect would vanish.

You can find this on YouTube if you want. Then you can stop it and pause/rewind/FF all you want. Nothing’s been distorted at all.

I was saying in another thread that I want to show this to my 7 year old. The faces get distorted in a way that’s very typical of caricatures or the way people mess with celebrities via photoshop (big teeth, dragging eyebrows or cheeckbones up). I’m curious how someone who hasn’t been exposed to those things would react to something like this. She wouldn’t recognize those distortions as the typical caricatures that we see all the time. What would she see? Something different? The same thing we see? Totally normal pictures?
Also, she wouldn’t recognize those people to begin with.

Altered or not, they aren’t the most flattering pictures of those people, and the pairs are probably selected to exagerrate the effect. I perceive some blending of the images, Picard’s bald head noticeably affecting the other picture. You can just click the stop button on your browser to check the images individually.

But it happens to me even if I cover one side of the pictures, as mentioned above.

I don’t see much distortion with one side covered. Have you stopped the pictures to look at them? Some of them are odd expressions and angles. I do notice that all the eyes stand out even with one side covered, while they are flashing by like that. When I stop the pictures it doesn’t seem that way. I also viewed from more than 3 feet away from the screen and the effect was reduced. I think at that distance both eyes are picking up images that aren’t from such distinctly different angles. The brain massages the input from the eyes a lot (and there’s some processing going on before the info even hit the brain). With flashing pictures, funny faces, different viewing angles for each eye, and brain enhancement there’s a lot of room for optical illusion.

Here’s another illusion(a new take on the curveball illusion question). It is supposed to be based on the transition from central vision to peripheral vision. I think in this case by focusing on the center of the screen everything is peripheral vision and my brain is interpreting things differently as a result. The flashing and dual images may just be enhancing that effect.

That’s because the sequence of pictures alternates the “distortion”–which is a manipulation (or simply a change) of the brightness, contrast and indoor-outdoor exposure setting–back and forth. So while one picture has one particular brightness/contrast/color key, the one on the other side has different, “opposing” keys.

I believe that when the picture on the right, for example, is with outdoor lighting, the one on the left is indoor lighting, and then the next pair of pictures is reversed, or something along those lines. Your brain is trying to be both indoors and outdoors at the same time, and ends up perceiving “distortion.”

I can’t help noticing that most of the photos are of people with very distinctive features. Not distorted, as in Photoshop, but actual faces with “extreme” features.

What has closing one eye got to do with anything? The open is eye will still be receiving peripheral information from two pictures simultaneously.

However, if find that if I cover the picture on one side with my hand, but still fixate the cross, the effect still appears, which tells me that it is not so much a matter of mixing up features from the images on either side (though that may still happen - the effect may be stronger with two simultaneous images), but from the current image and the one (or even more than one) preceding it. I mentioned this possibility in my original post.

If you look directly at the pictures on one side or the other, instead of at the cross, or just allow your eyes to move freely, none of the pictures look anything like as weird as they soon start to do when you fixate the cross. A few of them do look a little odd, with huge eyes, for instance, but that is probably due to the facelifts and other cosmetic surgeries that celebrities seem to indulge in, rather than photoshop.

However, it is possible, I think, that a much milder version of the effect is occurring even when I look straight at the pictures: that is, to a (much more limited extent) there is still some tendency to carry over aspects of the previous picture to the next one. Presumably this is just because they are changing so quickly, at a speed greater than any our face recognition systems (some components of which no doubt work faster than others) have evolved to handle.

… Yes, I am finding that if I stop the changes with the Escape key, even the relatively minor weirdness that I see when I look straight at one of the sets of changing pictures does not afflict any static face.

…OK, on further experimentation with the escape key I have found one woman who looks kind of weird even when she is static. I do not know who she is, but her eyes seem unnaturally wide. I am fairly confident this is just (very slightly) botched cosmetic surgery, though. It is nothing like the degree of weirdness I experience when fixating the cross, or even when fixating the flashing images themselves.

If anyone can figure out where this come from or who who posted it, or, better still, if there are any references to it in the scientific literature, I would be grateful if you could let me know.

Nonsense. They are just a random collection of frontal portraits of celebrities (probably because celebrity photos are easiest to find online) roughly normalized so that each face appears at about the same size. I am confident that this effect would appear with any faces, provided they were not all very alike.

Celebrities, especially female ones, do tend to have somewhat extreme faces, but I am confident the effect would work with ordinary people too.