2015-02-28

Color Perception

So a picture of a dress freaked out the nettubes Thursday.  My favorite response:

That anyone saw white and gold is very strange to me and somewhat disturbing, though polls indicate at least 2/3rds of people see white and old instead of reality.  Even when other pictures of the dress were posted, including from the manufacturer, some have refused to concede, claiming conspiracy.  (Meanwhile, one blue-black camp member replied to the proof of blue-black with "THERE ARE FOUR LIGHTS!", much to my nerdy pleasure.)

Of course we have probably all seen assorted optical illusions, some of which deal with color perception like the orange/brown shadowed object on a chessboard that is actually the same color (courtesy of Wikipedia):



And I presume you're all aware of the blind spot a few degrees outboard of your center of vision that is normally not perceived.

The brain engages in many such 'post-production' camera-correction efforts.  What makes this one strange is that it is such a stark contrast with no obvious solution like the orange/brown thing, and has to do with perception in such a basic way.

See, I have long used color as an example when discussing perception and objective reality.  Despite eye contruction variations and visual cortex differences, generally we can all expect that we can all agree that light of such and such nanometer wavelength is red.  That is to say, even if the output of your eye and brain, piped into mine, looked green or yellow or ultraviolet, we both know that everyone knows it as red and all agree on that.

It is distressing to learn that this example is not necessarily correct at such a scale.  It disturbs people because it is a basic issue of perception.  It's one thing to show the same facts to a commie bastard and a Constitutionalist American and have them disagree on the correct conclusion, but we would commonly expect them to at least concur on the basic color of an item.  I know sane and truthful people who inexplicably (to me) saw it as white and gold.

Of course, I'd be lying if I said I wasn't happy to have been firmly in the blue-black camp since the first moment.  Apparently, this suggests I have superior color discernment, in addition to my being able to pick up on spatial orientation cues that others miss.

I am led to wonder if perhaps there haven't been a few Vs. Debate events where folks are similarly color-blind, or at least color-illusory.  I thought of the last objection on this page, the old "band of brightness" debate, and some others where folks seemed to me to be actively refusing to see the obvious.  Perhaps they were, or perhaps they really are blind, by comparison.

Either way, I don't feel bad about responding to their flames with equal fire.  If you're blind or daft, don't flame ... in the debate of the blind, the one-eyed nerd is king.

4 comments:

  1. I was one of the rare individuals who saw neither white and gold or blue and black, but rather blue and gold. From this, I realised that there's more than just an optical illusion going on here. There's a side-by-side comparison of the controversial photo vs a manufacturer photo of the dress here:

    http://www.telegraph.co.uk/news/science/science-news/11440142/Dressgate-the-science-of-why-THAT-blue-dress-looks-white.html

    But even side-by-side with the manufcaturer photo (which is very apparently blue and black), the original still looks blue and gold to me. And it turns out, even using a tool as simple as the colour sampler from MS Paint, the blue in the original photo is in fact a significantly lighter blue than in the manufacturer photo (significantly different RGB value), and the lace does indeed appear to be gold in the original photo. Specific RGB values follow:

    Manufacturer lace (lighter part in mid-chest area): ~ 50 Red, 45 Green, 40 Blue
    Manufacturer lace (darker part around collar): ~ 35 Red, 30 Green, 30 Blue

    Original lace (med-chest): ~ 125 Red, 115 Green, 70 Blue
    Original lace (darkest part in shadow I could find): ~ 75 Red, 60 Green, 35 Blue


    Manufacturer dress body: ~ 25 Red, 50 Green, 125 Blue

    Original dress body: ~130 Red, 145 Green, 190 Blue


    So take heart, it's not just our eyes. The camera itself was bamboozled by the unusual design and nasty lighting conditions too, which was then further compounded by the optical illusion factor.

    ReplyDelete
  2. Oh, and just so we have a control test just to prove the above isn't me being incompetant, I also did the same thing with the sample optical illusion you provided, and according to Paint, both brown/orange spots have identical RGB values of 209 Red, 134 Green, 0 Blue.

    ReplyDelete
  3. I should also point out that when I say, "I see it as gold", I'm using the word "gold" very loosely. It's a very dull gold at best, and indeed on the basis of RGB, the dress lace in the original photo has much lower Red and much higher Blue than even dull shades of Goldenrod, so technically it isn't gold at all. Indeed, the indicated RGB is closer to Coyote Brown. So more accurately, then, the camera is actually seeing the dress as air force blue* and army brown. XD

    *It's not quite Air Superiority Blue (which is RGB 114, 160, 193), but surprisingly close.

    ReplyDelete
  4. Sometimes I see it as Blue/black sometimes as White/gold. It's been a few years since my last eye exam, perhaps I should schedule one.

    ReplyDelete