1. Limited time only! Sign up for a free 30min personal tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

I Strange Optical Phenomenon (diffraction or something else?) (solved)

  1. Apr 16, 2016 #1
    First of all i will post a picture of it ... you can find it ... somewhere
    sorry this is my first post :I

    I photographed our television with the flash on in order to recreate a strange patter i saw when looking at the suns reflection in the screen.(with my eyes it's not just the camera)
    You can clearly see diagonal stripes and the different wavelengths get split into multiple maxima.
    Over all it reminds me A LOT of light hitting a diffraction grating, which would make sense since the boundaries between the individual (sub-)pixels are dark and reasonably close together and that would make them behave a lot like a grating.

    Why the hell are they DIAGONAL ???
    And why are there two STRIPES ???
    Usually when light hits a diffraction grating you get a line of bright dots (perpendicular to the slits or absorbant stripes) or in the case of a two dimensional grating you get a grid of dots.
    And on another monitor that is exactly what i am observing.
    When shining a laser on it i even see 5 dots (the others are too faint to make out) in exactly the way you'd expect (perpendicular lines) when the reflected light is projected on a surface.

    My first idea was that the pixels might be arranged in some strange not perpendicular pattern, but they are not :(

    Though that you can observe a diffraction pattern at all is a little strange since these dark stripes between the pixels are behind the glass surface off which i thought the light would be reflected.
    That might be important ...

    I would be glad if anyone of you would give this a shot and try to figure out what is going on o:):confused:

    Attached Files:

  2. jcsd
  3. Apr 16, 2016 #2
    Suggestion, can you take another picture with the camera rotated 45 degrees from the last shot. Does the pattern rotate?
  4. Apr 16, 2016 #3
    This might explain it.


    From the above,

    "Most of the light from the flash is reflected straight back, causing the white region in the centre of the image. However some light is reflected back at an angle, causing it to appear displaced from the centre of the image. Furthermore, red wavelengths are deflected outwards to greater angles, causing the light to be spectrally dispersed in the reflection.

    This kind of device is known as a diffraction grating, used extensively in experiments to split a light source up into different frequency components. In this case fine structures in the TV screen are causing the reflected flash light to diffract. The camera lens then maps diffraction angles to positions on the camera CCD."
  5. Apr 16, 2016 #4
    Wow i am really glad someone responded to this :D

    And no the pattern does not rotate :I
    As i had mentioned the phenomenon is equally visible to the human eye and i did a lot of moving around with my head but sadly nothing changed.
    The (observers) distance from the screen also does not effect the picture you see.
    It always fills an equal part of your field of vision much like a rainbow which always is 40 something degrees in radius.
    The picture you posted was very helpful.In that picture you seem to have ALL the things you would expect from a diffraction grating:
    Two "rays" perpendicular to each other and two diagonal ones (which are part of the gridyou should see)

    Here are some pictures of what a normal 2d diffraction grating will do
    see 3.3 especially the picture 2D-GRATING-DARK%20TRIM%20RED.jpg

    But i think my new question is

    If TV screens act as diffraction gratings why are the diagonal maxima SO MUCH more intense than everything else ?
    It is like we are only seeing a part of the picture.
    How are screens different from usual gratings ?

    PS: I think i can explain why the two "rays" in my picture aren't perpenicular:
    The pixel densities in both directions are not the same and so the first 8 maxima dont form a nice square but a normal rectangle.
  6. Apr 17, 2016 #5

    I really dont think saying "it is just a diffraction grating" is a decent explaination for this.
    you can clearly see the dark stripes between the pixels, something that wouldn't happen with a normal grating
    And when closely examining the transition from red to green i believe you can see that the dots of light of different color are not on the same grid.(cant tell in the uploaded version/quality to bad)
    This offset between the centers of the dots makes me believe that what we are seeing is actually light that penetrated the glass somehow, interacted with the subpixels and then reemerges from the screen.
    That would explain why the amounts of cyan and yellow seem to die out in the outer "rainbows".
    The light is simply absorbed by the respective subpixels which only let red or green or blue through.
    Why the light would come back this way and why it is arranged in this pattern is still a mystery to me :L

    I would really appretiate your help :nb)o:)
  7. Apr 17, 2016 #6
    We have two Panasonic flat screens that have similar effect but the lines are at right angles and at approximately the 11-5 and 2-8 o'clock positions. We also have a sony flat screen with two primary lines at the 12-6 and 9-3 o'clock positions with two secondary lines (reduced intensity) 45 degrees to these. Computer monitor flat screens in the house show no effect.
  8. Apr 17, 2016 #7
    What sort of screen were they ?
    EDIT I have observed similar effects on mobile phones, monitors and TVs
    (sometimes rough surfaces mess it up)
    Interestingly the angle between the maxima does not seem to have any nice correlation with the pixel densities. (Which it would if that grid acted as a diffraction grating)That in turn suggests that the pattern is most likely NOT formed by the dark bits between the pixels.
    Do you know who we could ask ?
    Apparently we both fail to fully explain the patterns or in any way come up with a theory with which we could
    anticipate wether or not a screen will produce a pattern and which pattern.
    So maybe actively looking for help might not be a bad idea.
    It hink it would be immensely helpful to know how various screens are "made" up and what part could makethe pattern-

    If we have to figure it out on our own then more data cant hurt xD.

    Ps: On observing it: i found it very helpful if you either use a flashlight, camera flash or the suns reflection.
    And if you actually take a dark picture you can make out more details ... usually
    Last edited: Apr 17, 2016
  9. Apr 17, 2016 #8
    Maybe your TV's pixels are arranged like this. Note how vertical rows are not all lined up horizontally but alternate up and down.


    and not this,


    Note the slightly different "crystal" patterns. Gives different diffraction patterns?

    By changing the ratio of the height to the width ratio of the "primary LCD atom" in the first image above you could change the angles in your photograph?
  10. Apr 18, 2016 #9

    Andy Resnick

    User Avatar
    Science Advisor
    Education Advisor

    Interesting pattern- can you provide some details, such as how far the flash was from the screen, etc.?
  11. Apr 20, 2016 #10
    I am somewhat confused by what you mean by "primary LCD atom" and crystal in this context.
    Are you referring to the shape of the subpixels, because they resemble crystals ?
    And that because they aren't perfect rectangles but "crystal-like" you suggest that might influence the diffraction pattern.
    Is that correct ?

    Sorry that i was gone for quite a while...
    I was busy with school.
    Interesting theory you got going there Spinnor :woot:
    I had a similar thought about it, but for some reason i assumed the pixels might be hexagonal.
    When i took a closer look at it and found out they were oriented in rectangles i dumped the theory.
    However your idea looks very promising :).
    The bad thing is : pixels are small
    I can't really tell whether they are arranged in the way you suggest it.
    So after forever trying to get my camera to focus of the screen 3 cm in front of me while simultaneously shining a flashlight on it, this is the result :

    I got to admit was a little disappointed when i saw they were arranged side by side without any offset.
    But then i realized this is by far the best photo we have of this "penomenon".
    PS: I was not shining the flashlight on the colourful part, the (out of focus) reflection of the flashlight can be seen in the lower left corner.
  12. Apr 20, 2016 #11


    User Avatar
    2017 Award

    Staff: Mentor

    Some thoughts on this:
    • Your camera is not a screen. It has its own optics and focal distance. The pixel pattern would suggest that it focuses on the screen. It can resolve individual pixels, but probably not structure within those pixels - you can get one single-slit pattern impression per pixel, and you see light only if the angle fits.
    • The flash and the camera have some offset. This could create an asymmetry.
    • The patterns are quite dim. Had to make the monitor black (screen saver) to see them.
    • I get a bright horizontal line, a weaker vertical line, a weak 45 degrees diagonal line, and some weak substructure between the horizontal line and the diagonals. All things apart from the central spot are green. The flash in my phone camera is about one centimeter left of the camera.
    • Switching a different monitor off completely didn't work, then only the central bright blob appeared, radially symmetric without visible pixels or substructure in general.
    • I really have to clean the monitors
  13. Apr 20, 2016 #12


    User Avatar
    2017 Award

    Staff: Mentor

    That is a result of the LCD technique. There is a polarizing filter, then a liquid crystal, then another polarizing filter. If the screen is dark, then the light that passes through one filter arrives at the other filter at an orthogonal orientation, so it gets reflected back (some part gets absorbed). This works for outgoing and incoming light in the same way.

    In case you missed it, see my previous post, we posted nearly at the same time.
  14. Apr 20, 2016 #13
    The part about the dirty monitors made me laugh quite hard :biggrin:

    But i thought the polarizing filters for optical wavelengths absorbed exclusively without reflecing ?

    PS: I have to admit i didn't really understand your "single slit" explaination.
    What acts as a slit ? The dark or the bright part ?

    And how would you explain the strange angle at wich they are oriented ?

    PPS:I appreciate your help.
    Last edited: Apr 20, 2016
  15. Apr 20, 2016 #14


    User Avatar
    2017 Award

    Staff: Mentor

    The bright part. Not sure if it is relevant, but it could be.
    Depends on the filter I guess, but in general there is always some reflection.
    I was thinking about the displacement between flash and camera, but with flash+eye I could see that it does not matter. The angles I see here differ from those you get, so it is monitor-dependent.
    A laptop monitor shows a nice pattern, while the PC monitor does not show any pattern. The central bright spot looks much larger and blurry on the PC monitor.
  16. Apr 20, 2016 #15
    Do you have a magnifying eyepiece? With mine I was able to understand the odd set of diffraction lines on our Panasonic flat screens (plasma I think). With the eyepiece I could focus on the pixels and as I was moving the eyepiece away from the TV and the pixels moved out of focus I saw a square-rectangular grid of thin black lines with approximately the dimensions of the pixels BUT! they were not running north-south, east-west but the grid was rotated from such an arrangement so that the grid ran approximately in the 11-5 and 2-8 o'clock positions.

    The image of the alternating basis of pixels I posted I now think was from a old style TV with a cathode ray tube, my theory bits the dust I guess.

    You never did say what type of TV or camera you were using, we need some facts!
  17. Apr 21, 2016 #16
    Here are your facts:

    The TV is
    • LCD
    • 1080p
    • 1.2m * 0.7m approx 54 inch diagonal
    • All pictures were taken with the TV turned off (but i think it also happens if it is turned on)
    • it is made by Toshiba
    The camera is a little harder because i have used multiple ones:
    • My sis' compact camera wich just happened to be lying around (first picture) (no idea about focal length or aperture size)
    • My Nikon d3300 with a (pretty crappy) kit lens "18-55mm focal length" at f/3.5-5.6 using the biggest Aperture
      (This one produced the more high res pictures)(i think the forum somehow downscaled or compressed the images ... so i might add a dropbox link)
    I also observed the phenomenon with
    • my eyes
    • a shitty mobile phone camera
    So you see the observer does not matter !
  18. Apr 21, 2016 #17
    THANK YOU VERY MUCH Spinnor :biggrin::smile:
    You solved the mystery.
    Your first idea was correct !!!
    It has something to do with the shape of the pixels.
    I was actually shocked that the solution was so simple.(somewhat underwhelming)
    In the first TV screen each pixel is made of two parallelograms at an angle.The boundaries of these ARE perpendicular to the "diffraction line".
    And in the second TV screen something similar happens except that the "crystal" (as you called it) upper and lower ends of the pixels form the patterns.
    Although i find it a little weird that such a discontinuous pattern can create regular old diffraction like
    a set of continuous lines.
    I will post the dropbox link with some pictures here https://www.dropbox.com/sh/rjlerz83qaelhuq/AACQ5ew_S7AYFLfO5ZFscA0Ea?dl=0
    I suppose from the pictures you will be able to piece everything together.

    Also a thank you to mfb for clearing up some of my misunderstandings about polarizers and LCD's in general.
    I am still intrigued by two facts :
    1. That your monitor made a radially symetric pattern when turned off.(thats pretty lame/why no diffraction pattern?)
    2. I think it behaves more like a grid than a single slit.(but whatever)
    (And thanks for making me laugh.Just imaging the look an your face as you came to the conclusion: "I really have to clean the monitors"o0):DD)

    PS: I would like to edit this somehow so that it is clear the issue has been resolved and you won't read through the rest of my confused rambling if you newly arive in this chat
    Last edited: Apr 21, 2016
  19. Apr 21, 2016 #18
    Cool! I was going to ask you for the model number and send Toshiba customer support an email and ask for help. Good work!
  20. Apr 22, 2016 #19


    User Avatar
    2017 Award

    Staff: Mentor

    Well, one of them. I don't know why.
    A grid would show much more pronounced maxima, and it would not show the pixel structure.

    You don't see the dirt under regular light conditions:

  21. Apr 22, 2016 #20
    I am gonna try to convince you that it mostly behaves like a grating.Ok ? :)

    It does show pronounced maxima.At least the screens in my/our household.
    As proof i would have you take a look at the "diffuse laser long exposure" (the name was something like that) picture.
    The reason you see multiple dots where you would just exect one maxima
    is that i moved the laser around a bit (during the long exposure) by accident and these are the same maxima in the pattern.
    Just made by differently postitioned light sources.
    That the maxima are still a little diffuse is likely due to the "diffraction grating" having huge holes.
    See this unspecified.jpg I tried to draw in the "pixel boundaries"
    I know it is quite the opposite of well made °~°.
    So it doesn't really act like a grating but it is the model that comes closest.
    Actually the part that acts as a grating in the screen form the "laser long exposure" is even patchier.

    To further proove that it behaves like a grid i photographed the reflection of the laser after being reflected of the screens.(projected on a surface)
    I put them in the dropboxfolder.https://www.dropbox.com/sh/rjlerz83qaelhuq/AACQ5ew_S7AYFLfO5ZFscA0Ea?dl=0
    They all have Laser in the name.

    I think that is half-true.The colour filters of the subpixels will interact with the light no matter what...
    You cant block that effect.
    The lines of the grating however schouldn't be visible wich is probably what you meant.
    But wounld't that be exactly the same for single slits ?
    The slits themself dont cast any shadow ...

    The vertical shadows between the pixels are very thick (on screen1) and therefore dont produce any diffraction pattern.
    Only the horizontal ones do. And they don't cast any shadow as far as i can tell.
    (even if they would cast a shadow the resolution would not be good enough to see it.The bright parts would simply "bleed" light into the shadow)

    I know the pattern in your photo looks a lot like a single slit
    but think about it ... wouldn't the frequencies get seperated more ?
    I think the grating maxima are incredibly close together.
    and because the maxima of different colours and different "diffraction orders" overlap the whole thing looks pretty white.
    Additionally in every grating you can still see some single slit effects:

    I hope i could convince you :D
    Let me know if i did :wink:

    Actually in the laser long exposure picture there is a strange horizontal pattern that
    looks like it was made by either either a single slit or just a few slits.
    So i might be wrong after all :eek:
  22. May 2, 2016 #21


    User Avatar
    2017 Award

    Staff: Mentor

    Sorry, forgot about that thread.

    If we would see a grid diffraction pattern, the pixels could not be visible at all. You cannot have interference between adjacent pixels if they are resolved in the image.
    You see light if two conditions are satisfied: (a) the direction is right for light coming from the pixel (it is a camera/eye observing the pattern) and (b) the angle is right for positive interference.

    The last picture is not from an extended grating, it looks like a double-slit or maybe triple-slit pattern. Otherwise the maxima would be much smaller and sharper.

    I tested it with some more monitors. Some show nice patterns with variable angles, some do not show a pattern at all.
  23. May 3, 2016 #22
    The reflection of the Laser does have extremely pronounced maxima.(take another look at the pictures)
    It couldn't have been produced by a single slit or double slit.
    I trust the observations that much.
    I think we agree on that right ?
    But that seems like a big problem:
    I finally get why that is now, although it took a while.
    Thanks for beeing insistent after i falsely declared this problem completely solved. :smile::wink:

    I ran the numbers for the grid spacing you would need to have for this pattern.
    Since i still think it looks like a grid.
    ## n \lambda = \Delta s ## for constructive interference with a grid
    ## \Delta s = sin(\alpha) l ## l is the grid spacing
    alpha the angle at wich the light exits

    I first thought we would have to consider the optical density of the material in wich this happens.
    (The angle of outgoing light in the material is different due to the shift in wavelength and then the angle changes when leaving the material due to refraction.)
    Now i think both changes cancel each other out.

    After solving for l and plugging in approximate values for lambda and sin (alpha) i get the VERY rough estimate of
    ## l=2*10^(-5) m ##
    The reality could easily be twice or half that.
    PS:What is actually the correct way to write such an extreme uncertanty? usually you do it with plus/minus some value or percentage but that makes no sense here.

    The previous result is far from the real ## l=\frac{1.2m}{1920*3}=2*10^(-4) ##

    It is off by a whole order of magnitude !!!!:))
    So my initial assumption is completely off the table.
    And to make it worse it is not small enough to make a grid between the pixels....
    I thought maybe the dark lines between the subpixels consist of multiple lines.
    For example, one in the middle grounding and two on the sides which each control one subpixel voltage.
    Somewhat like the thin wires in this under_microscope.png
    But i don't think that is true.

    I can see up to about the sixth maximum.
    That has been confirmed by an experiment with a laser pointer.(disclaimer i did actually shine the laser in my face with a one of the "other maxima"o0)
    but my eyes are all good the laser is only p<5mW)

    BTW I am talking about the screen from the beginning.The one which had diagonal pixels.
    Do you know of any structure that is somewhere around ten times smaller than the subpixels (but parallel to them)?
  24. May 3, 2016 #23


    User Avatar
    2017 Award

    Staff: Mentor

    The reflection of a laser can be different thing, the maxima are on a screen.
    I don't know. The polarization filter has to have even smaller structures to work.
  25. May 4, 2016 #24
    I think they are the same ...
    Just imagine you put one eye where one of the secondary maxima is projected.
    You would then see a bright dot on the screen,where the laser hits the screen ALTHOUGH the beam is not directly reflected into your eye.
    That would be true for all the positions where some not-primary maximum hits your eye.
    So when shining a really big laser at the screen you would only see the points which would project some maxima into your eye.
    The patten which emerges form this is very similar to the projection you get when only hitting one part of the screen.

    It is similar to the fact that a single raindrop will project a whole rainbow around it, but very many raindrops all over the sky project white light back.
    But a camera or human can see a rainbow in the sky similar to the projection of one drop.

    I don't know whether i explained that very well. :sorry:
    Thats really cool btw... and it might take a pen and paper to really get it if you never thought about it that way before.

    An interesting observation:
    In the reflection patterns there is usually one very bright maximum (the reflection off the glass surface)
    And then there is a grad pattern for most screens.
    The interesting thing is that the center maximum of the grid reflection is slightly offset compared to the glass reflection.
    That seems to confirm that the laser enters the screen get reflected off the back polarizer and exits again(after interactin with something).
    That extra distance in the screen produces the offset.
  26. May 4, 2016 #25


    User Avatar
    2017 Award

    Staff: Mentor

    An eye has a lens and its "screen" is behind its optics. An eye without a lens would see a completely different world (most notably: extremely blurry).
    A camera/eye can. A screen cannot (unless you add optics).
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted