Persistence of Vision

August 22, 2013 1:40 PM

Ever do a photo shoot with hundreds or thousands of similar images, only to find yourself scrolling through Lightroom really fast, rating the best shots and/or deleting the worst ones? Here's an optical illusion that makes the case for why you might want to slow down. (via)

The phenomenon occuring is described as "Persistence of Vision," whereby an "afterimage" is thought to persist in retina for approximately one twenty-fifth of a second, and in this case, invoking a distortion that sets in almost immediately. Scientists refer to this as the phi phenomenon or beta movement, whereby fixed objects create the illusion of motion; an illusion that of course makes motion pictures watchable, and that appears in everyday items like LED scrolling banners.

posted by phaedon (33 comments total)

that's crazy! the illusion is so much stronger than i would have ever imagined. as i read the article i kept my eyes on the left or right image and thought, what's the big deal? but then when i focused on the cross-hairs in the center - yikes! thanks for sharing this.
posted by rude.boy at 1:47 PM on August 22, 2013 [2 favorites]

All I know is that Patrick Stewart looks beautiful every time. The perfection of his face is immune to illusions, glamours, and distortions.
posted by 256 at 1:56 PM on August 22, 2013 [2 favorites]

but the effect persists much longer than 1/25 second; my uneducated surmise is that the illusion has much more to do with the limitations of peripheral vision.
posted by jepler at 2:01 PM on August 22, 2013 [3 favorites]

That is super freaky. Even when centring my vision on a single image I'd find my peripheral vision of the other image bleeding weird results into the image I was concentrating on.
posted by Mitheral at 2:01 PM on August 22, 2013 [1 favorite]

There are monsters in the corners.
posted by blue_beetle at 2:03 PM on August 22, 2013 [8 favorites]

So he's shooting the same old same old image and the insight is in something to do with the retina.....maybe shoot markedly different images would solve the problem ? Like something interesting ?
posted by sgt.serenity at 2:03 PM on August 22, 2013

THEY LIVE
posted by Flunkie at 2:04 PM on August 22, 2013 [1 favorite]

It'd be fun to try to reverse engineer this; take the regular images and distort them so that they look like what your peripheral vision suggests they look like, and then try to combine them with something that reverts them back to normal as they flick by.
posted by quin at 2:06 PM on August 22, 2013 [8 favorites]

And when the animation is slowed down or stopped, the effect is not lessened for me (though it's somewhat lessened after I look at a face and then back at the middle without the pictures changing in the meantime)
posted by jepler at 2:13 PM on August 22, 2013

I think it has more to do with the fact that your fovea has as much acuity as the rest of your retina combined. The more padding around the crosshair, the more the images are pushed into your (terrible) peripheral vision. Usually we scan over items with the much more accurate center of vision and then assemble a high resolution mental image. (IANA ophthalmologist)
posted by benzenedream at 2:23 PM on August 22, 2013

He's got the wrong name for the effect. This is not persistence of vision; it's a relatively newly discovered distortion of peripheral vision (which may possibly be related to this one or may be specificially related to how we process images of faces.)
posted by ook at 2:23 PM on August 22, 2013 [6 favorites]

I do experience the effect with peripheral vision, but...before I did that, I looked at the photos straight-on and thought a good many of the faces looked distorted, either from the actual plastic surgery or from the angle/lighting/pose or both.
posted by desuetude at 2:33 PM on August 22, 2013

I thought a lot of them are already pre-exaggerated/asymmetric (e.g. quite a few sneers) , and although I couldn't confirm it (a speed control would be handy) I would think that the changes from one to the next might make quite a bit of difference (forehead size, chin shape).

And am I missing it or has nobody (especially TFA) said exactly what they are seeing (blue_beetle is close)? I eventually see what looks like a parade of caricatures.
posted by achrise at 2:56 PM on August 22, 2013 [2 favorites]

There's no "there there."

Try covering up one of the images and the crosshair completely. Then stare at the procession of images in the remaining panel. After a while, distorted images show up on their own. Seems like there are more in the panel in the left for whatever reason.
posted by lord_wolf at 3:06 PM on August 22, 2013

Ah, okay -- I gave it a few more tries and after a while, some -- but by no means all -- of the non distorted images appeared a little off. But some people remained beautiful no matter what.
posted by lord_wolf at 3:08 PM on August 22, 2013

ook is right.
Previously
posted by Valued Customer at 3:10 PM on August 22, 2013 [1 favorite]

I was really hoping this was going to be about a new and ingenious app that did the drudge work for you, picking the best one out of dozens of similar shots. I would love such a thing.
posted by Flashman at 3:12 PM on August 22, 2013

That sounds like a hard AI problem. The solution to which could make you a lot more money than that garnered by a photo app.
posted by Mitheral at 3:21 PM on August 22, 2013 [1 favorite]

I read the post and followed the instructions and I'm not sure what I should be seeing. There doesn't seem to be any explanation of what is obvious to everyone else. Could someone tell me what is supposed to be happening? I see lots of pairs of faces, and some text that says "Isn't this weird?"
posted by monkeymadness at 3:34 PM on August 22, 2013 [2 favorites]

That came really close to giving me a panic attack.
posted by Ice Cream Socialist at 3:39 PM on August 22, 2013

Could someone tell me what is supposed to be happening? I see lots of pairs of faces, and some text that says "Isn't this weird?"

The faces appear grotesque and distorted.
posted by mr_roboto at 4:23 PM on August 22, 2013

Wow! There's only one Nicolas Cage. But if you stare at the crosshairs then they all turn into Nicolas Cage!
posted by Kabanos at 4:32 PM on August 22, 2013 [3 favorites]

I think we're seeing the faces distorted because of filling-in. The features are mostly, and we're seeing them in our peripheral vision, so the different features start fading out and the rest of the face is pushed in to fill-in the gap. Then the process gets interrupted by a different face or by our eye movements, so another feature gets faded out.
posted by Joe in Australia at 4:55 PM on August 22, 2013

That came really close to giving me a panic attack.

Me too, though moreso because it reminded me of the hard drive with 500Gb of unprocessed raw images, along with a shoebox in a changing bag in a road case containing maybe 10 rolls of 35mm, 20 rolls of 120, and maybe 30 sheets of 4x5, all unprocessed, some shot almost two years ago.

Also, as I discovered earlier this month, rodents got into the storage locker with my boxed-up darkroom and chewed a hole in my bottle of fixer. Well, I hope they at least all got wasted and had a good time.
posted by [expletive deleted] at 5:07 PM on August 22, 2013

monkeymadness, try moving closer. I didn't get the effect at more than a foot away from my laptop screen. (But yes, the effect is that the faces look distorted.)
posted by capricorn at 5:30 PM on August 22, 2013

Hijacking this thread for a book recommendation, but if you find this interesting you might love Oliver Sack's The Mind's Eye. Or most any of his books

Sight is crazy fascinating.
posted by graphnerd at 5:32 PM on August 22, 2013 [1 favorite]

Could someone tell me what is supposed to be happening?

At first the faces appear normal, but then the eyes become large and drift towards the center of the face. All expressions turn cruel and exaggerated, though being outside of your focus they're blurry anyway. After that everyone's skin takes on a strange scaly texture and the whites of the eyes disappear, leaving them dark and enormous and glassy, and noses flatten to slits. Many people place the palms of a hand on their neck. Suddenly the faces have normal skin and noses again, but the eyes are staring and the mouths are stretched into improbably wide grins. Looking away from the computer, everyone in the office is staring at you with the same expression, idiot grins beneath baseball-sized bug-eyes. You excuse yourself and go to the bathroom. It's comfortingly familiar and blessedly empty, but when you enter the toilet stall and close and bolt the door, wiping sweat from your brow, you turn around to find the toilet is made of black plastic - or is it stone? or glass? - and you look into the bowl and it goes down, down, down...

The alarm clock rings. You reach for it, knocking it over. It rolls onto its face, and on the back you see the battery door open - and empty.

That's what happened to me, anyway, but like people were saying I don't think this is persistence of vision, which is a more subtle effect.
posted by 23 at 7:28 PM on August 22, 2013 [21 favorites]

Can anyone explain why Tom Cruise's wheels appear to be turning in reverse?
posted by Catch at 7:31 PM on August 22, 2013

Eventually they all start to look like Mad Magazine caricatures.
posted by ShutterBun at 9:01 PM on August 22, 2013

To me the question is, why isn't this effect prevalent in normal life? After all, we see people out of the corner of our eyes all the time, and if we're daydreaming, maybe we're focused on a point while a crowd is moving in our periphery. So we should experience this phenomenon normally, but we don't. Why not?

Here's my guess.

I think the reason this is happening is because, first of all, our brain magnifies anomalies taking place in our peripheral area, since our visual acuity is weaker outside of our fovea. That way, for example, if a wild animal is heading our way, we don't have to wait to identify its detailed features to know to move the hell out of the way. As soon as a small weirdness is detected, it is magnified into significance and that attracts our attention. Once the animal enters our fovea, and we see it's just a puppy, then that detailed image overrides the highly compressed scan we initially took, and we can shut off the alarms.

Now the other thing is that our brains are very creative when it comes to scale and coloring. For example there is a well known checkerboard illusion where a dark square and a light square can be exactly the same shade but due to our internal processing, we perceive them differently. It's not a voluntary thing; we can't shut it off if we try. The processing is hardwired.

Thirdly, because we use binocular vision, our two eyes tend to move, focus and process in the same way simultaneously. What we do with one eye, we do with the other one as well.

Combine these three phenomena: peripheral high-sensitivity, automated post-processing and tandem binocular effects, and what happens?

In the real world, everything works normally. The end. However, in the case of these celebrity photographs, they are all taken from slightly different angles, with different lighting conditions, exposure, focal length, saturation etc. In real life that would never happen. Two people standing next to each other in our field of vision are under the same sky, at the same time of day, with shadows facing the same directions, and so on. Here they are not, but they are still side-by-side on the page. So our mind says, "They're next to each other, they must be part of the same scene." And our crude peripheral signal processing tries to normalize the two images by applying the same correction algorithm to both.

Except that doesn't work. Since they're all randomly jumbled together, our peripheral vision doesn't know what to do with the differences it detects. One person is desaturated, the other is normal. Apply correction, what happens? The desaturated person is corrected to be normally saturated. No problem. But the same filter is applied to the already normally saturated person, who then gets supersaturated and looks like a splotchy HDR nightmare. Or, one person is shot with a 85mm lens, the other is shot with a 35mm lens. The 35mm person gets normalized, but mentally doing the same to the 85mm person's face makes it look squished like a pancake. And it's all going by quickly, so we don't have time to make a second pass of the data. Our initial crude peripheral processing does all the heavy lifting and it is just not designed to deal with two totally unrelated images at once. It magnifies the tiny differences in the viewpoints of the photos, and creates monsters.
posted by xigxag at 10:58 PM on August 22, 2013 [2 favorites]

I solve this problem by running Lightroom on a computer that is too old and slow to rapidly switch between images. My happy beachball cursor distracts me from the horrors of vision persistence while the picture slowly resolves in the background. Take that, professional photogs!
posted by caution live frogs at 5:11 AM on August 23, 2013

They look like grotesque caricatures to me. Fascinating.

(Or, what Shutterbun said)
posted by Acey at 11:41 AM on August 23, 2013

So he's shooting the same old same old image and the insight is in something to do with the retina.....maybe shoot markedly different images would solve the problem ? Like something interesting ?

On a professional photo shoot, you shoot until you nail it. Depending on the circumstances of the shoot (model, subject matter, client, moon phase, etc.) you can easily wind up with hundreds of similar looking frames. If you're good, the golden frames are at the end of the series - part of being a pro is developing a feel for when the shot is in the bag.
posted by oxidizer at 12:34 PM on August 24, 2013