Tricking the brain
Our visual system does not simply measure the light coming from a surface, it interprets in the context of its surroundings
Many years ago I published a puzzle — an optical illusion. It was for a chess news page, so I took the original image that was making the rounds and inserted chess pieces. This is what I presented my readers:
My questions at the time were:
- Which of the two knights is the white knight (i.e. which is the lighter of the two)? Is it the one on the middle of the chessboard, or the one on the top?
- Which square is lighter, the knight square in the middle or the knight square at the top?
The incredible answer is: both knights, and both squares on which they stand, are exactly the same shade of grey (RGB = 107, 107, 107) with identical light shading of the pieces.
I know this is difficult to believe. Even enlarging the image and examining it closely fails to convince. For instance in this image would you believe that the knight on the top is standing on a square that has exactly the same grey value as the square on which the lower knight is standing? Even if you stare at the picture for a long time and concentrate intensely your eyes and brain will tell you that the bottom one is clearly much lighter.
One way to convince yourself is to punch two small holes in a paper and hold it up to the screen, so that only the knights and the squares on which they stand are visible. You will see that the are exactly the same. I have done this electronically in the following image:
Another way is to copy the image into a paint program and use the colour pipette to determine the colours. Even better: cut out a small section of knight and square and insert it as a floating element in the picture. Move this over the original place and then over the other knight. You will see that it blends in perfectly.
The original picture was created by Edward H. Adelson, Professor of Vision Science at MIT, and had no knights on it. You can find a description and a clean sharp JPG here.
If anything Adelson’s version is even more incredible. Can you believe that the squares A and B in in the picture on the left have exactly the same colour, an exactly identical shade of grey? Not the letters, the squares on which they stand. Once again closer scrutiny does not help.
We can try the experiment in the real world. Here’s a picture of a chessboard with a shadow cast over the left side:
Which square is lighter, A or B? The correct answer is B! If you cut out a bit of the B square and move it over A you will see that it is actually slightly lighter than A. I’ll do it for you:
A description of the illusion was given at Prof. Adelson’s web site. Essentially it said that our visual system does not simply measure the light coming from a surface, but interprets the luminescence in the context of its surroundings. The brain knows that a light surface in a shadow must reflect less light than a dark surface in full light. Adelson explained how the brain determines where the shadows are and how to compensate for them in order to determine the shade of grey that belongs to the surface.
Addendum: here’s a very impressive video demonstrating the effect.
When my son Martin saw the above illusion he said: “My eyes and brain have a bug, I’m sending them in for an upgrade!” But seriously: it is not a defect in our visual system, but rather the result of very sophisticated algorithms used to process the data that it receives. The visual system is not very good as a precise physical light meter, its purpose is to interpret the information and help understand the nature of the objects it perceives.
Ready for an even more outrageous illusion?
Take a careful look at the above image and compare the square in the middle of the upper surface, which is clearly deep brown, with the one in the middle of the front face of the cube, which is bright orange. In front of the cube there is another brown square, right?
Will you believe if I tell you that all three have the exact same colour and brightness? Once again I have cut out a piece of the “orange” square and pasted it next to the “brown”. See?
Finally I have a special treat for you: copy the following image and print it out:
Now cut out the large images and the two strips, and punch holes where the black dots are. Show your friends each picture and ask them about the brightness and colour of the critical bits. When they are convinced that what you are saying is impossible give them the strips to place over the parts they think are completely different. It will blow their minds.
Here’s another variant of the cube illusion by Beau Lotto [click to enlarge]:
Would you believe that the blue squares on the top surface of the cube on the left are exactly the same colour as the yellow ones on the right cube? I have copied the top surface of the one on the left corner and inserted it to the right of the right cube. Very hard to believe, don’t you think?
In this video we see Neil DeGrasse Tyson talking about “optical illusions”, renaming them “brain failures”. In the rest of the video you can watch him and others freak out at David Blaine’s ice pick “trick”, which is not really a magical deception but Blain actually puncturing body parts (something Indian fakirs have been doing for centuries). If you have the nerve watch him horrify Ricky Gervais with a long needle through his bicep. Blain’s a maniac.
Addendum (7 April 2017)
This is probably the best colour illusion I have encountered:
How does the area within the red square on the left compare to the corresponding area on the right? Left is clearly blue and brown, on the right it is yellow and white, right? Absolutely no way we can cut out the square on the left and move it to the right, and discover that the colours are exactly the same. Right? Wrong!
In the above image I have cut out the red square and moved it to the right, to compare the colours. The brown/blue areas on in the left image exactly match the yellow/white areas on the right. Still don’t believe me? Then take a look at the following GIF animation that was published on Imgur:
Hard to believe, but perfectly true.