Explain this photo

299792458Mps Luke Mandle sent this photo (click it to embiggen) of his four-year-old son making pancakes. Can you explain why his eyes are closed on the left but open in the reflection?


    1. Yeah, t.v screen closer.
      It’s like that space/time thing where lighting that strikes 2 poles simultaneously appears so from the standpoint of someone sitting still looking at them head on, whereas if we were to travel towards one of the poles with the other standing behind, the lightning would appear to strike the closest pole first.

  1. This looks shopped. I can tell from some of the pixels and from seeing quite a few shops in my time.

  2. resolution pixels on the chip load from L to R (or R to L) and the split second refresh caught both sides of the blink?

    My camera phone seems to go from bottom to top and is apparent taking pictures from a moving car.

    1. Agreed. The image does not load simultaneously and an eyeblink is really fast. The same thing happens with mechanical shutter film cameras at long exposures.

    2. I am betting on this being the case. Also if you notice the picture is taller than it is wide, so the picture was probably taken vertically (which would coincide with the bottom to top of your camera phone)

  3. Rolling shutter maybe? In the time it took the sensor to read left-to-right across the field, the kid opened his eyes.

  4. It’s taken in portrait orientation, and SLR (or DSLR) shutters go up and down. So one side of the picture is just from just slightly before the other.

  5. best guess: the digital camera doesn’t capture the image all at once, but rather sweeps from one side to the other, and he just happened to snap the pic either just at the beginning or just at the end of a blink.

  6. Because he blinked, and the CCD is scanned from one side to the other. Or he’s a demon. And who lets a 4 year old make pancakes un-attended? with a spoon no less! And I imagine he’ll learn to put a shirt on when he fires up the bacon.

  7. What nickelrocket said. a number of ccds used in camera phones use a method of scanning which loads pixel information sequentially on capture. the speed of the ccd can introduce artifacting (e.g. when taking a pic of a spinning propeller), as well as what’s seen here.

    decent overview here on how delays can occur, and effects they have on images. the one captured, however, is a much better example, I think.


  8. His eyes are open in the reflection because someone recently viewed the same site I did just this morning. That’s not Photoshopped (at least not the eyes) nor a mystical event. It’s a picture of the same child on the laptop monitor in an identical pose, . . . but with open eyes. The most common oddity I’ve seen using desktop photos of surrounding environments is to take a photo of the region behind the monitor and then make that the desktop image. This creates an illusion of a vacant area where the monitor is expected to be, and trust me, turns a lot of heads once people realize what they’re seeing.
    So, again, I suspect photo/desktop, but only the creator can confirm or deny.

    1. It’s a reflection: The child’s hand in the reflection stops at the television border’s reflection.

      It’s a rolling shutter effect of a CCD / CMOS image sensor being polled sequentially by the controller.

      Though I am intrigued by ian_b’s explanation.

  9. Agree with post #3, CMOS mostly and especially on cheaper devices is not instantaneous but scanned out of the chip. What model of camera was this taken with?

  10. It’s a photo from an iPhone?

    There have been past articles on boingboing dealing with the way certain cameraphones use photo-gating to capture pictures. My guess would be that the kid blinked right as the photo was taken, and the sensors on the left caught the blink, while the ones capturing the reflection missed it.

    Or it’s just a trick of the angle?

  11. Occasionally you can catch your own reflection doing this. Stare at a mirror, look away, then look back really fast. Repeat hundreds of times. Eventually you’ll start noticing the reflection isn’t quite following you anymore. Start screaming “I see you! I see you!” Congratulations, you’ve flushed out your devil spawn reflection!

  12. No no! It’s a new SUPERFAST shutter, and because the light in the TV screen reflection takes twice as long to get to the camera, we’re actually witnessing the birth of time-travel photography!

  13. No flash. The exposure would take a little longer, and with the l-to-r or r-to-l pixel capture it just captured it at slightly different times in the (very quick) blink.

  14. Come on now… He’s got his eyes closed in the reflection, too. Nice bit of suggestion there. Mark could be a stage magician.

    1. >> simple. the child is the devil.

      As an Internet Expert with extensive credentials and over eighteen (almost nineteen!) months experience in Netflix, IMDB, and Wikipedia, I am going to concur with my colleague’s opinion on the matter, but I will take it one step further. He is not the devil, he is the son of the devil, the so-called anti-Christ.

      I hate to be the bearer of bad news to the parents of this child, but my professional recommendation is to be observant and note any unusual manifestations of his identity in the form of hounds with glowing red eyes, agitated Roman Catholic priests hiding in nearby hedges, and sing-songy otherworldly chanting originating from the Indian burial ground upon which your palatial upstate New York mansion is built.

      If your toaster starts giggling at you in the voice a distressed, young girl, take it back to Target. That’s a well known recall issue with a particular brand of Krups appliances and not related to your child being the Stygian templar of a hellish reign of mass murder and the utter destruction of all that is right and holy in our world. I hope you kept the receipt.

      Also, allow me to be the very first to pledge undying loyalty and support to my future master and all-powerful overlord savior of the glorious coming paradise on earth. All hail!

      Alternate theory: it could be that the bacon the kid is frying has randomly and unexpectedly done the work of the LHC for about $1.38 and the young lad has inadvertently split the Higgs boson next to a lovely pair of eggs over easy. If this is the case, that means a portal to a parallel universe (probably not an evil one, as the alterna-child in the television is goatee-free) has likely been opened inside your television. Fortunately, the portal is probably localized to the Food Network, so don’t watch that channel. The lives of everyone in the universe depend on this.

      Seriously, bro. Don’t watch the Food Network.

      Also, I guess this would mean that your kid is not the son of the devil, so please disregard that, but I’m still sticking to the “hail, master” part just to be safe.

      Having said that, we are all just game sprites in tranCendenZ, so I wouldn’t worry about it too much anyway.

      Unless… we’re still *IN* the game. In a dream. Of the game. In the matrix.

      Burma Shave.

  15. My initial thought before reading any comments was that it was due to the fraction of a second in difference it takes for light to reach the camera from two separate distances. But on further reflection it could be a combination of that and the type of camera/shutter that was used to capture the image.

    Explanations aside, it’s a cool photo. And I don’t think it’s ‘shopped.

  16. Nothing to add that nickelrocket and others haven’t covered except:

    Nice cabinetry!! I love it!

    @Luc: did you look at the bigger version? (click on the photo.) It’s quite obvious that his eyes are open in the reflection.

    1. Ah, you’re right, in the large version it’s easy to see his eyes are open in the reflection. Doh! In that case it must be rolling shutter.

  17. This could be an artifact of a focal plane shutter (or its digital analog). If it moves from left to right (or vice-versa) there is a time difference between when the left and right sides of the detector are exposed.

  18. Low end digital cameras scan the scene top to bottom (left to right), in a few miliseconds. It works like a flat bed scanner, but faster. If you pan that camera from left to right while taking a picture, I guarantee that vertical lines in the picture will lean one way or another due to the image being taken line by line rather than in an instant. I believe it’s called a rolling shutter. So if some one blinks while the left is being scanned and has there eyes open during the right side being scanned the reflection’s pose will be different than the actual person’s pose.

  19. Looks like an optical illusion to me. Everything in the reflection is darker and blurrier. I think we’re just seeing the darker, slightly smudged reflection of his eyelashes, and our brain is filling in the details, like a form of pareidolia?

    Just my guess.

  20. ps: Metal utensils in what looks like a non stick pan! The teflon flakes have already caused his soul to mutate! Hurry before he starts harvesting others!

  21. Shop. From the angle where the picture was taken, and the position of the TV, there is no way that would be the reflection IRL.

  22. It’ll be using a rolling shutter, which means that the pixels at the top of the picture are recorded before the ones at the bottom (or vice versa). It seems likely (as the picture has a portrait aspect rather than landscape) that the camera was being held on its side at the time to record the picture.

  23. The reflection is much darker, making the shaded of the eye appear larger in comparison and causing the eyes to appear more open. It’s a visual illusion, not a fancy technology effect I’d wager.

  24. i’ve seen this trick before. it’s funny how easily we are tricked into the fantastic, and then when you hear the logical answer it all seems so obvious. it is just a simple case of finding twin boys, making one live inside a television, while the other lives in a home, and then a simple matter of choreography.

  25. I agree that it’s some factor of transmit time of the light sensor. However, the photo suggests that the camera was turned to a portrait orientation, which means this wasn’t some left-right phenomenon but a perfectly normal top-to-bottom raster scan.

  26. This is why you should never EVER put a TV in your kitchen. What happens in an eyeblink can take a lifetime to make right.

  27. my god. this is really simple, c’mon!! they are NOT open they seem to be due to that the reflection is pushed in as an image and what seems to be the line of the closed eye if you press it in, it will create a “dot”. like an open eye.

    1. He’s either actually making pancakes or pretending to make pancakes. Kids shouldn’t fry bacon without shirts until they’re at least six.

  28. It’s just a side effect of capturing someone’s soul by photography. If a reflexion of that person exist at that moment, it continues to live within the reflecting agent for the time it take for the body to zombiefy, then disappear.

  29. Nah, it’s the shutter wipe thing. You people suggesting photoshop or LCD monitor trickery, use Occam’s Razor. Would Mark post a photoshopped picture and say “OMG, HOW DID THEY DO THIS?”

    For an interesting example of shutter wipe applied to a rotating object, check out this iphone pic of the propellor of my cessna:

  30. My guess is his eyelids are faster than the speed of light.

    I for one welcome our new Young Mutant overlords.

  31. Someone with a little more time than I have this morning (and/or wiki-ing experience), should add this photo to the Rolling shutter wiki that GrumpySteen and fataltourist mention above.

  32. 4-year-olds can make pancakes? I’m totally having kids, now. I had no idea they could be put to useful work so early.

    “After you’ve cleaned up my pancake dishes, weave me a dazzlingly intricate rug, child!”

  33. My guess (if not ps’d) is it’s shutter lag…. or rolling shutter…. & the kid blinked at precisely the perfect moment.

    The phenomenon is seen when shooting an slr with flash at faster than rated flash sync speed…. you will only see half or part of the frame bc the shutter blocks the image from being recorded…. or a photo of an old tv at faster than 1/60th /sec.
    Without flash the camera will expose the whole frame but with a shutter moving 1/30th of a second or less, the movement from top of frame to bottom is faster you’ll get 2 possible exposures at once.

  34. Easy test:

    With the same camera, drop a horizontal 2×4 as you trip the shutter (or even very shortly before you release the shutter). See if/how it distorts the board as it falls and the picture is taken.

    If it is a sliding shutter or some other such artifact, you will record it. If the average blink is 300-400 milliseconds, going from full open to closed would be about half that, and you can calculate how much distortion you would expect to see in .15-.2 seconds.

  35. The image on the TV is from a video camera and is slightly in the past due to the lag time of the video camera and the TV. Hence we see the kid blinking just a moment ago and we also see him in real time (after the blink) standing at the stove.

    1. I got it backwards. I meant to say that the blink is in real time and the open eyes image on the TV is pre-blink.

  36. I don’t think that rolling shutter, shutter wipe, or load across the chip explain this. As Vnend said, a half a blink takes onwards 150-200 milliseconds. That’s fast but still much slower than the shutter. Unless a tripod was used, the shutter speed of the camera would have had to be greater than 1/30 of a second (also confirmed by the fact there is no blurring of the boys fingers/hands/features that were in motion during the picture). A 1/30 second is three times shorter amount of time it takes for the half blink.

    In short, rolling shutter does not explain this.

  37. Off-topic: It appears BoingBoing comments have reached that critical mass where there’s too many comments for many commenters to feel like they have to read them all before posting. Thus many, it appears, have switched into “write-only” mode and we have many essentially duplicate posts.

    I’ve seen this on a lot of forums, and it’s a positive feedback cycle – more people making comments without reading the discussion, which means a longer discussion with more duplicate posts, which means less people willing to read a good percentage of the comments before commenting themselves. As people with something interesting to say realize that their comments won’t be read, they stop posting. Then I stop reading.

    The only cure is aggressive moderation (eg. as done on SomethingAwful).

    On-topic: we’ve seen this very common effect on this site before (remember “bendy propellers”?), but it’s still a cool photo.

    1. I think that the bigger problem on BoingBoing is actually due to their aggressive moderation policy where posts don’t make it onto the comments page immediately. User posts seem to take a few minutes and anonymous posts take even longer. It’s not that people aren’t bothering to read others’ comments before posting – it’s that they never even get the chance to.

      1. Matt Deckard
        Spot on about the “moderating” lag and its differential for anon vs signed in comments.

        I’ve previously posted a comment that was “original” and by the time it appears (even when signed in, let alone when anon) there are other comments already above mine saying much the same thing.

        I’ve even seen a comment that I’ve posted when signed in, appear in the thread and look like an “original” comment and later on some late-appearing anon comments are inserted above making me look like I didn’t bother reading their comments before I posted.

        Look at the time intervals (or lack of them) between comments above to see how accurate your assessment is.

    2. We should all have to answer three questions about the comments already posted before being allowed to post ourselves, but obviously within a fairly short time limit, so that our own comments remain timely.

      Or just post any old rubbish, and then score unoriginality points for the number of times it’s already been said. We could keep a running total…

  38. It probably has something to do with the poor child being burned by hot grease whilst pancaking with no shirt.

  39. i think this could be a photoshop, for example look at the tv its at an angle faceing away from the child, so it cant possibly be an actual reflection, the boy would have had to been staring right at the tv itself with the photographer pointing at the tv, what has happend here is that two pictures were taken, with the camera mounted on the tripod the boy was instructed to stand still and close his eyes, then pic 2 was taken of the eyes open, you crop a small portion of pic 2 to give the illusion of a reflection ,then its a matter of fitting the cropped pic 2 in the area of the tv

  40. It was taken with a cell phone camera. It was in a vertical position, that means that the cell phone was rotated to take the picture. My Moto Q takes the picture progressively from top to bottom. The kid just blink normally but the camera is slow to take de full still.

    1. That doesn’t explain it. The reflective surface is not angled so that it is picking up the entire scene–you only see the front of the range skirt, too. If you draw that imaginary line up the left side of the image, you can see that the range hood would not be there, because it ends in the middle cabinet and the reflection only shows the one on the end.

  41. It’s a symmetrical kitchen and the reflection is actually his twin brother, standing a few feet away from him in excatly the same pose.

    If you think this trick is good, you should see them do The Vanishing Man.

  42. You can see the same effect with most of today’s 35mm film SLRs, which use a curtain-style shutter. There’s a famous example in this happening with film in this photo by Jacques Henri Lartigue (he was using a larger format camera, but still with a focal plane shutter): http://bit.ly/ddB6PV

    Because he was panning with the camera, the people seem to be leaning to the left. But the car was going faster than he was panning, so it appears to be leaning to the right.

    There’s a more in-depth explanation at http://www.largeformatphotography.info/forum/showthread.php?t=31903

  43. I just saw some cool pictures a (very non-professional)friend took of propellers during a flight using her I phone. It was like several pictures of the prop completely stationary in different positions and then composited together. It looked pretty cool. I wonder if the same thing is happening here and, is that a bug or a feature?

  44. the camera is in a tripod, took 2 pics of the kid who blinked. being only seconds from one picture to the other it’s a lot easy to overlap them in photoshop and just erase the reflection bit from the picture that is above with very few tweaks. ridiculously easy.

  45. I’m not sure I understand the question. Is this a clever trick that the submitter is daring us to figure out, or is this a mysterious natural phenomenon that the submitter is seeking our help to explain?

  46. Got my pancakes here. With hot applesauce. Thanks, Mark for the lunch suggestion. I need practice with my flipping, though. I wonder if the kid gives lessons.

  47. Any chance that the TV being closer to the camera could have some effect as well? I know it’s a slim chance, but could that compound the rolling shutter effect?

  48. FTR, an eyeblink is about 350ms. Compare that to shutter speeds and the shutter explanation becomes not not just reasonable but inevitable.

    1. An the blink of an eye is 350ms, but from open to closed is only have the blink of an eye and actually a shorter time 175ms. However, this doesn’t support the rolling shutter idea – it contradicts it.

      From open to closed the eyelid takes 0.175 seconds. The slowest shutter speed setting that could be used without blurring is likely around 1/30 or 0.033 seconds. At this speed, the shutter is way faster than they eyelid. This isn’t rolling shutter.

      A number of commentators mention the rolling shutter effect of airplane props. These props are making around 33 complete revolutions in one second. Constant eye blinking taking 350ms a blink would produce 3 blinks in one second. The props are moving much faster than the eyelid.

      This is not rolling shutter. This is not rolling shutter.

    1. That is THE single funniest thing I’ve read in days. So subtle and complex. Understated, yet still managing to say so much.

      I called my roommates around me to explain to them the genius of that.

      Thank you, anon. Go in peace. ICP forever.

  49. ummmmmmmm the eyes aren’t open in the reflection, the darkness/contrast of the eyes make it appear that way. drop the image into photoshop, up the brightness all the way and kill the contrast… and unlike the kid in the pic, you will see

  50. WHAT IS WRONG WITH YOU PEOPLE!?! The kid is making pancakes with a SPOON… that’s the real story here. And all you can think about is camera shutters, pixels and photoshop.

  51. the boy is astronomically large and submerged in a bose-einstein condensate. the photons in the reflection traveled a longer distance, and therefore show an earlier time.

  52. I’m surprised nobody has suggested the Pauli Exclusion Principle as an explanation. If the atoms in the monitor screen are spin polarized the mean free path of a particle…
    Wait… Speed of light… Nope, I got nothin’

  53. It is a distortion effect, related to both a slight change in angle, and the darker reflection.

    The reflected image is a slight degree off the line needed for a perfect mirror (the top of the cabinet seems to be angling slightly down and away), so you’re getting a slight bit more of a view from under the eyes (perhaps more lashes visible?). Then, the overall reflected image is slightly darker, and the color registration is different, making darker areas look larger (or more shadowed). Note how the light in the reflected window is more blue than the light from the window itself. It may be the blue wash that seems to saturate the reflection at some level.

  54. There’s a really simple explanation. Obviously, Luke cloned his son and built an elaborate faux-kitchen with a tinted screen designed to look like a television, and then set up the photo we see above.
    Then he realised he didn’t want pictures of his half-naked son (or his clone) on the internet so he photoshopped in a different child (and his clone).

  55. I’m guessing that the image on the television is not a reflection but instead a live feed of the image being captured by the camera. The delay involved with capturing / encoding / transmitting / decoding / displaying is more than enough to account for the delay of a blink.

  56. First off, it doesn’t. they are closed in both. If it did appear in any way that his eyes where open, it’s because the tv is a few inches to the right of the camera lense. This means that the angle from the camera and the tv reflection are off just a bit. Enough that you can see more of his face by a few centimeters. Since you can see more, the eyelashes are more apparent. since it’s a reflection, it’s easy to mistake the eyelashes as the pupils of his eyes. that is, if the pupils are at the bottom of his eye socket, instead of in the middle, like when they would be open.

  57. Could it be his eyes are only half open, such that from a direct angle they look closed but from a slightly skew angle they look open?

    It does appear to be a different angle, since I notice in the reflection his head stick outs several inches from beyond the cabinet, however on the left his ear is flush with the cabinet.

  58. Horizontal axis is compressed in the reflection, and the reflection is darker. Vertical axis is unchanged. So…child looks thinner, more tan, and the closed eye slit on the left appears to be more of a rectangle than a slit on the right, which your brain then interprets as ‘eyes open’. Gotta admit some of the other explanations are a lot more creative though!

  59. It doesn’t look like that’s a reflection at all. The reflected surface (a TV) would have to be in front of the child, not off to the right as it is in this picture.

    1. Glad you said that because I was thinking the same thing — the reflecting surface isn’t in front of the child so how can it be a reflection?

      1. That’s not how optics work. If you look across a lake you can see the reflection of trees on the opposite shore even if they aren’t actually growing out of the water.

        1. Yes, for anyone who has any doubt as to whether Brainspore is correct, I suggest a simple experiment: Next time you’re washing your hands in front of a mirror in a public restroom, look to your right or left. If you can’t see the reflection of the person standing next to you, it doesn’t mean your theory about reflections is correct. It means that person is a vampire and must be destroyed!

      2. Imma let you finish, but I just have to say, the American** education system is the worst education system of all time*!

        * not really
        ** I presume you are American

  60. Photo is rescaled and exif tags stripped. Is either shopped or an artifact of the unknown (but definitely crappy) camera.

  61. Why has no one realised that you cannot get a perfect reflection from a flat-screen monitor by standing to the side of it! The kid would have to be standing directing in front of the monitor to get a non-obscured reflection. The reason for the photo is that the photographer initially took the photo with the intention of creating a transparent screen-look. He photoshoped it to fit the background but he then turned the creen to its side and realised this would make a much more interesting photo. He darkened the image and ‘shopped it so the objects fit (failing to factor in the range hood at the top). When he took his final photo the kid blinked and hence the difference in eyes. Plus, the remote control is sitting in such a contrived position, indicating that this was not an ad hoc photograph, instead it was very much set up!

  62. Who cares about the kids eyes – what I want to know is how come I never knew you could train your kids to make pancakes at 4 years of age?!? Mine are 9 and 10 and I’m still cooking for them! ><

  63. i’m the mum….just became aware of this posting.
    1. It was sausages not pancakes.
    2. He is an identical twin!

  64. I still think rolling shutter’s a possibility, but IANAE. Given the timings of 175ms for half a blink, the fact that it *looks* like he’s about to open his eyes on the left and they’re *almost* open on the right, shaves that 175ms down even further. I don’t think 1/10 shutter speed on a crappy camera is that uncommon, especially on something like a phone camera, but I agree with the others, this needs some more background on the source camera or (much) better, EXIF data.

  65. Well I’m guessing it wasn’t taken with a nicer DSLR…

    -It’s high unlikely a picture on the TV. How the hell would that even work? It couldn’t be a CRT/tube TV, or else you’d see the scan lines. LCD nope, there’d be off axis discoloration. Perhaps plasma with an ass load of tweaking to get the colors just so. Perhaps…but I’m not betting on it.

    It’d guess it’s something like a rolling shutter effect. Perhaps the device taking the picture is accessing the sensor serially and is very very slow? Hell it might not even be a digital camera in the normal sense. It might be a crappy old camera phone, or a nanny/surveillance camera…

  66. To those who say that this can’t be caused by rolling shutter because the shutter would have to been open for at most 1/30 of a second:

    You’re confusing a mechanical shutter with a digital scan.

    It’s true that a mechanical shutter would need to take no more than about 1/30th of a second or so in order that the image not be blurry. This is because at longer shutter speeds, the film will have been exposed to light long enough to pick up motion.

    However, a digital scan of the image does not have the same limitation, because by the time the bottom sensors are receiving light, the top sensors are no longer receiving light and thus will not be subject to blur.

    Instead of blur, the left and right halves of the picture will have been taken at different times, and thus we get the rolling effects everyone is talking about. Each part of the image, though, will have been exposed for much less than 1/30th of a second.

    I couldn’t get any concrete information about the speed of the digital rolling shutter, but this post sugests that the iPhone’s rolling shutter is about 200 ms. This then fits perfectly with the speed of a blink and the rolling shutter theory for this photograph.

    1. I retract my arguments that this couldn’t be a rolling shutter effect. SamSam has written a clear logical response that doesn’t involve demons or complicated trickery.

      My problem is that I shoot pictures with a Pentax K1000 made 5 years before I was even born. I wasn’t clear on how a digital camera produces an image. But it does make sense that while each small area may only be exposed for 1/30 of a second, the whole picture may take much longer to produce.

      Kudos to PrettyBoyTim in post #133 demanding and using some actual science to address the question. Logical explanations are far, few and repetitive in this comment section.

      I don’t even know what to say about those who said this couldn’t be reflection because the boy isn’t directly in front of the mirror. That’s just a poor conceptual model of optics. As long as the boy is in front of the plane of the mirror, it is possible for his image to be reflected. Hang a half-body length mirror on a wall so that the top of the mirror is about 2m from the ground. You can see your feet even though your feet are not directly in front of the mirror. Does that mean you have demon toes? Probably.

      1. Motion blur appears at any exposure time, it just depends on how fast something is moving, and mechanical and digital cameras function in ways that are similar enough to provide a useful analogy to understand the ‘rolling shutter’ effect in this digital photo.

        The exposure time only means how long one part of the camera’s photo receptor or film is being ‘scanned’ by the camera – the total time the camera is taking the picture can be much longer. If you have a mechanical SLR, you can open the back and see that at higher shutter speed, the shutter curtains form a slit that travels across the film plane, and that becomes narrower as the speed setting increases, although the total time the shutter is operating remains the same. A digital camera can have an mechanical shutter, but usually it works by ‘scanning’ across the photo receptor electronically.
        The 150 millisecond figure for eye closing comes from dividing the 300 to 400 millisecond in Wikipedia for a blink, but as mentioned before, this does not take into account the dwell time when the eyes are closed. The closing time is given as 50 to 120 milliseconds here: http://www.ehow.com/about_5372151_fast-eye-blink.html
        So the explanation of ‘rolling shutter’ works quite well.

  67. And I agree, differences in the actual image and the reflection are due to parallax. I.E. Orange Juice gut fully visible on the left, partially obscured by the stove on the right.

  68. I think we’re all tip-toeing around the obvious here:

    The kid on the right is trapped in the Phantom Zone.

    1. It’s not going to take a nuclear explosion to get him out, is it? Can we just use the microwave somehow?

  69. This is easy. By your use of the word embiggen, you are obviously a Farker, and your camera tricksy friend too, by extension. Farkers are indefatigable Photoshoppers (and Fappers).

  70. “An the blink of an eye is 350ms, but from open to closed is only have the blink of an eye and actually a shorter time 175ms.”

    That would be from “mid-blink” to “open”. I’m fairly certain that an eyelid wouldn’t reach the bottom of travel and instantly change direction. That suggests there is a pause at the bottom of travel, and the actual “transit time” of an eyelid would be notably less than 175ms.

    Note also – the whole point of the “rolling shutter” is that the exposure time of an given pixel might be only 1/30 second, but it’s not the same 1/30 second for every pixel. I should make the time go go read the Rolling Shutter site, but I would hope that it documents that the elapsed time of an expusure might actually be 100ms even though exposure is only 3.3ms.

    A neat “rolling shutter” test would be a solid colour display that shifts through the spectrum of colours over the course of about 1/10th of second. Then each individual pixel would be coloured according to its absolute capture time.

  71. Either intentionally or unintentionally, the photo contains information from two different points in time – perhaps not that far apart, since it wouldn’t take that long to blink.

    Intentional Theory:

    Photographer took two pictures in rapid succession with the camera sitting on the counter or mounted on a tripod, and simply caught the child with eyes open and with eyes closed. All that would be required is a very simple operation to paste the left half of one image over the right half of the other image since the only information that changes is the position of the child’s eyelids. Not very complicated at all and while photoshop could be used to do that, any image editing tool (ms paint) could be used to pull that off in a few minutes.

    Unintentional Theory:

    The photograph actually shows that the time it takes for the camera’s imaging electronics to record an image and the time that it takes a child to blink are on comparable time scales. Meaning the camera is scanning pixels vertically starting at the left or right edge and working it’s way across the image, and that the time that passes between the time when the camera’s digital image capturing electronics captured the child on the left and the right is long enough for his eyes to open or close during a blink.

    – Todd

  72. I love the way this was clearly answered in the 3rd post yet people don’t even read that far and keep giving the same answer, or trying to say it’s faked.

  73. I love it – 102 comments and only three main themes…rolling shutter, photoshopped/faked, or vampire.

    I’m going with vampire.

  74. Enough of this. It’s time for some SCIENCE. I’m mighty suspicious of this 300-400ms blink time that’s been bandied around, and also of this idea that it takes half the time of a blink to shut your eyes.

    Therefore I’ve done an experiment. The set up: one PS3 and one PSEye camera.

    I shot some footage using the camera of myself blinking, close up, at 30fps. Examining the footage frame by frame I found that my eyes can go from 80-90% open to closed between one frame and the next. At 30fps that’s 33ms between each frame. The total number of frames taken to blink was 4 – 5, suggesting a total blink time of around 150ms.

    As my five year old is in bed I can’t compare his blinks until tomorrow. If someone has both a young child and one of those cameras that does slow motion we’ll be able to get some even better data.

    Also, we need some people to measure the ‘rolling shutter lag’. I’d try videoing myself dropping a rod against a wall with a tape measure on it, but my Wife is watching TV in the same room and I don’t want to disturb her. Anyone else up for the challenge?

    Finally, we should try and replicate this photographic effect ;-)

  75. I would guess the monitor is showing a photograph that is very similar to the live shot. If you notice, the child’s right wrist is slightly more bent in the monitor.

  76. Nevermind the spoon–why is a four year old making pancakes with a hot skillet on a hot stove at all. Barechested, no less, for maximum burnage.

  77. I don’t think it works. the tv is at the wrong place to get that reflection. The angle just doesn’t play out in my mind.

  78. What about the phosphor coating on the inside of the Tube, as it appears it might be a CRT, rather than LCD screen. This would retain both open and closed states of light info, granted the reflection we’re seeing is on the front surface, there might be some light lag from the phosphor as well… combined with rolling shutter lag/photon gating…

  79. “Can you explain why his eyes are closed on the left but open in the reflection?”


  80. Answer:
    Camera sweeps from top to bottom.
    Camera is on its side.

    I would further assume that the camera is sideways with the flash or top facing the tv. Therefore when the picture was taken it actually swept from tv to child, and it may even be possible that a flash was the reason for the instant blink.

    They call me Einstein.

  81. I first thought of a rolling shutter. Otherwise it could be an image displayed by the television (depending on the type of the screen) and the boy tried to keep the same position (not the live feed, otherwise one would see the television on the image).

  82. I know the text said pancakes and I said bacon and eggs, but that’s exactly my point. How could the boy be making pancakes AND bacon and eggs?

    We’re through the looking glass, people.

  83. M. Night Shyamalan’s the Sixth Sausage.

    “He thought he was cooking links, but he was linking… to a terrible other place of mystery!”

  84. I’m surprised nobody has mentioned the obvious distortion in the reflection: the curve of the TV screen makes the reflection much narrower. This compresses the eyes horizontally, making long thin ovals (mostly-closed eyes) more circular.

    This may not explain the whole effect (rolling shutter sounds like it’s a major contributor), but it’s certainly a factor.

  85. when you click on and enlarge the photo it’s obvious his eyes are open on the TV screen.

    The 666 on his head is a little trickier to see, however.

  86. That’s a Samsung DLP TV, isn’t it?
    If it is, then the screen is actually a coated kind of multi-layer plastic, and the TV image is internally projected to reflect off a mirror mounted in the back of the cabinet, facing the screen.

    I’m not sure if that means anything: just that the reflection may not be off glass.
    And, such screens are flat: no curvature.

    A friend had one of those DLP Samsungs, and this set looks similar to it. FWIW.
    Perhaps its multi-layer nature of the screen alters the speed of the light bouncing off of it by a hair, at this angle.

  87. I still don’t see how the TV screen, at that angle, even picks up his reflection. Looks like a combo shot to me. But I have a tin eye.

    1. Try imagining any flat reflective surface as the edge of a pool table and each photon as a cue ball.

      If you’re banking photons off the TV from the kid to the camera lens then the TV would have to be somewhere between the two objects, not directly in front of the kid.

  88. Several people have already pointed out the scanning direction, assuming it was taken with a digital camera. If this was taken with any film camera with a sliding focal plane shutter, it would be easy to explain. I don’t know much about how digital cameras acquire images, but it makes sense to me they might do it progressively from one side to the other. This old photograph by Henri Lartigue demonstrates an extreme version of it. http://www.shorpy.com/images/photos/lart.jpg

    A google image search for “focal plane shutter distortion” will turn up numerous other strange effects.

  89. Never mind the photo. What I want to know is how GrumpySteen and fataltourist managed to post the same link at virtually the same time!

    1. Why is the sky blue?

      Short answer: nitrogen and quantum mechanics.

      Entertaining answer: See Cliff Stoll’s ‘The Cuckoo’s Egg’.

      Long answer: Obtain a PhD in physics. See you in 7 years.

      Why is the sky blue *in the picture*?

      Fill-in flash allows for an interior exposure that also (reasonably) accurately exposes the sky through the window.

  90. The monitor seems pretty far away from the child (as well as on the same side as him) to be picking his reflection up and at that angle.

    1. Cut the picture vertically with a line passing through the vanishing point. Flip the kid over and he’ll be right on top of his double except for a little distortion from the curve of the monitor.

  91. It’s an HDR? The dark area of the screen has more pixel artifacts, so I guess it took longer exposure to get a reflection of a quite dark TV-surface.

  92. This child’s reflection looks like he’s dropped some weight.
    He’s skinnier in the reflection.

  93. …and I yet suspect that this screen, as a DLP screen, is made not of glass, but of a multi-layer translucent plastic, which has different optical properties than glass.

  94. I believe recent Sony point-and-shoot cameras (the HX5V for example) have the ability to take multiple rapid exposures and then stitch them together, choosing the sharpest areas from each image to make a single composite result.

    So the pixels on the left and those on the right could have come from exposures made at slightly different times, and this could all happen automagically from a single press of the shutter button.

    You guys are not considering the possibility that cameras aren’t as simple as they used to be.


  95. Night setting on the digital camera — image captured brightly illuminated kid on the left with eyes closed in mid-blink and shutter stays open longer to capture dark areas like the reflection off the TV during which time the kid had completed blinking and now has his eyes open. The human eye would have seen the TV screen as dark when viewing the entire scene, the camera kept its shutter open longer to capture more light from the dark areas.

  96. That’s easy.


    Note how the secondary reflections (the reflection in the monitor of the reflection in the stove) are not quite positioned the same. That is physically impossible. An image of an image has to be located at the corresponding point.

  97. I have read every comment and have come to the conclusion that the only plausible possibility is the evil twin who lives in the tv.

    1. I have read all 181 comments preceding mine, so as to be a thoughtful commenter, and have now completely forgotten what I signed in to say.

      I’m not going to let that stop me though. Look Ma! I’m on Boing Boing!

  98. The subject blinked. This could have been taken using a camera with a slit scan shutter, like an iPhone, or even a conventional focal plane shutter found in digital SLRs. The shutter gap travels vertically in an SLR. The camera was held portrait style, therefore the gap was traveling horizontally.

    It would be interesting to see the EXIF data. The complete absence of IPTC and EXIF data is either a normal BB procedure, or done deliberately by the submitter.

    The sky through the window is blown out, but the tree is visible, the tree is mostly in focus, the interior of the house is in focus from the closest buttons on the remote to the window. The photographer was using a small aperture. I don’t see a lot of noise in the shot either indicating a lower ISO number was used (there is apparent noise and noise-reduction in the deeper midtones).

    The average duration of a blink is 300 to 400 milliseconds, so a 1 second exposure give the subject enough time to blink, given that the subject and his reflection are at opposite sides of the image. Given 1 second, there is roughly half the frame between the two sets of eyes. That’s 500 milliseconds or less, depending on the shutter gap.

    On the other hand, the shutter speed might have been faster than one second, with the effect developed through practice. There is some movement with the mouth in the right half of the frame. I would expect blurring and more subject motion with slower shutter speeds. There are some subtle differences in posture in the two images of the subject, indicating some time has passed.

    The small aperture and portrait format leads me to think this photo was staged to demonstrate this effect.

    In any case, a shutter isn’t all or nothing, it’s almost always a mechanical progression across the
    focal plane (except in the case of leaf and various high-speed and special purpose shutters).

  99. He’s being recorded. The video camera is tied into the TV, and there is a delay between what is happening and what is showing up on the screen. So it’s not a reflection. The tv’s on.

  100. I hope this kid doesn’t grow up to be neurotic because of 180+ remarks about his eyes. It would have done it for me if I wasn’t already on my way by that age.
    Larry Niven said the only way to fry bacon was in the nude. That way you don’t overheat it and make it spatter. I have not tested this theory, but go ahead, and let us know what happens.

  101. The photographer probably had the camera cables, plugged into the AV ports on the TV, while taking the picture. The image on the TV would lag, as the camera was capturing a photo.

  102. because iphone cameras process a picture from left to right in straight vertical lines. over a second a bunch of them line up. good timing, this picture is awesome!

  103. I like the rolling shutter/CMOS-scan theory. The picture may have been cropped to change the aspect ratio (or maybe CMOS scans are perpendicular to mechanical shutter sweeps).


    Flash-fill was clearly employed, and the flash was on when it was at the T.V. (see the bevel around the screen).

    So, if it is shutter, then it started at the T.V. (right side) and swept left. The kid’s eyes were still open but his reaction to the flash was complete within the added few millisecond it took the shutter to make it to his eyes.

    -NetlabLoligo (who tried to sign up but couldn’t)

  104. Two cameras set up next to eachother. One Videocam hooked up directly to the tv, with real time mirroring or possibly just shooting at a mirror. The other camera is the camera that takes the picture. Difference in face; image processing lag from the Videocam and tv.

    If it was a reflection I don’t think he would look thinner in the television. It does look lie the kind that has a flat screen.

    Sorry if I am not first with this theory, but there are just too many comments to waste your time reading them.

  105. Upon reflection, I think the most likely explanation is a mechanical one, relating to the way the camera used “builds the image”.

  106. Relatively slow shutter speed with early or late (second) shutter fill flash – you can see the flash in the dishes to the left of the boy on the left side of the photo. Nice picture!

  107. Or maybe it’s a cleverly shaped mirror resembling us of a TV ( http://blog.makezine.com/archive/2010/07/cleverly-shaped_mirror_opens_a_door.html ), and the boy in the reflection has his eyes closed, and the right one has his eyes open? As the real world (right side) is already heavily pixelated, the reflection was much worse, so the photographer decided to smooth it out with some photoshop plugin for salt and pepper, which actually mistook the open eyes of the boy for pepper.

  108. Can you explain why his eyes are closed on the left but open in the reflection?

    first part: young boys aren’t used to being photographed while cooking half naked, so he blinked at the wrong time.

    second part: the TV is part of an entertainment center, the whole purpose of this device is to entertain human beings. The screen is part of the entertainment center, and is designed for showing humans entertaining pictures. A shot of a boy with opened eyes is deemed a better picture than a boy with closed eyes. TVs don’t reflect reality. Humans assume that what we see on TV and looks convincing enough, probably is real. Therefore we see that the reflection of the boy in the TV has open eyes.

    That is why, from our point of view, his eyes are closed on the left but open in the reflection.

  109. It’s simply because the reflection is darker and squashed a bit horizontally, like a fun house mirror.

  110. The same picture appeared in Gizmodo. From reading the comments, it appears that the Boing-Boing readers found a possible solution much faster (CMOS rolling shutter) and had more inteligent comments overall.

  111. Yes, camera is set to “night scene” mode. Shutter opens for a long time to capture the dark areas, then at the end of the same exposure fires the flash to fill in the close-up details. Subject blinked at the time of the flash. Because of the angle, all of the flash light bounces off the glass screen away from the camera, so you do not see any of the light from the flash in the glass screen – just the darker image from the earlier time before the flash.

  112. I’m not sure if its already been mentioned.

    The boy in the photo may have an Identical twin. Both strike the same pose in the kitchen to create a mirrored effect. One twin with his eyes open and the other closed.

    The TV set has a replaced screen and no backing case, set up to appear that it is reflexing the one side of the kitchen (for this the kitchen would have to be larger than it appears to be). The TV remote could by a diversion.

    Or the Television set has some kind of photo of the boy opening his eyes glued to the screen. Taken at angle that would appear its mirroring the kitchen.

  113. I don’t think it is too far fetched, that this could be caused by in-camera HDR. It has become increasingly popular in P&S cameras and there would be a short delay between each image being stacked.

  114. how sad is it the first comment is “shopped”. come on you pixelpeepers, not everything you can explain with your limited knowledge can be explained away by photoshop.

  115. If anyone has gotten this far down the comments, then, like me, they must be amused at the creativity of some of these answers. in my opinion, from worst to best:

    rolling shutter: sure, that could be it. but that’s so boring.

    photoshopped: posting a photoshopped image on boingboing and asking ‘how they did it’ is just illogical.

    delay caused by the extra distance the light travels bouncing off the TV screen: no, it’s safe to say that distance is negligible.

    not a reflection in the TV but an image feed from a camera: if so, then all the horizontal lines in the TV image wouldn’t be horizontal, they would be sharply distorted, approaching a vanishing point just to the left of the TV.

    i like the idea that the phosphors in the TV screen are “grabbing” the image of the kid and causing a delay, but no, in that case the light from the kid would be caught by the phosphor and then scattered in all directions. The result would be a soft glow from the screen, not a discernible image.

    the image is definitely a reflection. and for those who claim that the TV would have to be in front of the kid, google “angle of incidence equals angle of reflectance”

    the most convincing post for me is that HDR technology is somehow at play: using a longer exposure for the TV reflection than the actual image of the kid.

    but I’m still working on a theory that involves the light getting polarized off of the TV screen… nah, that ain’t right either…

    sure is a creepy photo though

    1. Anon #216 has a good analysis there.

      However, I looked at the mixer bowl in the non-reflected part of the image. You can see some windows in the background, so that’s the source of the illumination on the vertical edge of the TV screen – not a flash. Also, the boy’s chest is in shadow while his back is lit, so again, no flash.

      So it looks like an HDR artefact – until you realise the boy *is not* reflected in the mixing bowl, while he is in the TV!

      This proves the ‘identical twin brother in the TV’ theory, except the twin outside the TV must be a vampire, and therefore does not appear in reflections.

  116. I think this is the same phenomenon that led people to believe they saw one of the 9/11 planes explode before it hit the building. What they were really seeing was a video frame that took longer to produce than it took the plane to cover the remaining few feet to the building and begin exploding.

  117. camera is connected to the TV when picture was taking, so its the delay between the live moment and the tv display moment. ( the time that takes for the camera to display it on TV)

  118. HDR? Hight Dynamic Range
    Technique qui permet un rendu accentué des pixels, dans les zones a fort contraste.
    Les images se font en 2 temps, la première pour les hautes lumières, la seconde les basses lumières, avec un interval tres court.. 1/4 de seconde.
    Ensuite les 2 images se superposent en une seule.
    Au final une seule photo, deux instants.

  119. He is cooking meth and only one of them is high. Guess which?
    Kids this is you.
    This is you on drugs.
    Any questions?

  120. Eu acredito que a pessoa tirou a foto do menino com os olhos fechados e editou no fotoshop para ganhar publicidade e/ou varias pessoas tentando descobrir como isso poderia acontecer, para mim teorias sobre esquerda e direita significam que vocês são burros ‘-‘

Comments are closed.