Wednesday, April 16, 2008

Window Shade/Rocking Disk Illusion

Here is the "window shade/rocking disk" illusion from the Journal of Vision article I wrote with Justin Charles and Mallory Shear-Heyman in 2005:

In this demonstration, you see a ring that is half black and half white, and a center disk that changes gradually from white to black, and back again. [note: the demonstration does not show up in an RSS feed. ]

What to notice:
1. A veil of brightness appears to drift across the center disk (like a window shade that is pulled up and down). The effect vanishes when you remove the white/black ring (click on the "add/remove surround" button).

2. Click on the "inner ring" button to add a thin gray circle between the disk and the black/white ring. Now the disk appears to rock up and down.

3. Click on the "rotate" button to rotate the ring. The direction of the shading and the rocking shifts in response to the orientation of the ring. The shading appearance can be considered an illusion because the center disk is always physically uniform. That is, the disk is never a combination of colors; the color pixels that make up its surface are all white, or all black, or all an intermediate shade.

Why does the window shade, shade?

Many illusions surprise us (and attract our attention) because they violate our basic assumptions about how objects in the world behave. In the real world, objects appear to be relatively stable: if you roll a white ball from a concrete sidewalk to a green lawn, the ball doesn't suddenly appear red the moment it hits the grass. But many illusions show that context can create dramatic changes in the way something looks (take a look at Adelson’s checker shadow illusion for an example).

In the window shade illusion, the context is provided by a black/white ring. I started using the black/white ring to study the effects of contrast. Contrast refers to the relative difference between lights. Neuroscientists know that most of the information that the eye sends to the brain corresponds more closely to the relative difference between lights in an image than it does to absolute light level.

The window shade illusion is set up so that when the disk is white, the contrast between the disk and the white part of the ring is low, and the contrast between the disk and black part of the ring is high. When the disk is black, the contrast shifts in the opposite direction. The high contrast edge therefore jumps back and forth across the disk.

The window shade illusion is very similar to the "contrast asynchrony illusion" (see the previous post), in which two disks appear to modulate out of phase with each other, but get light and dark at the same time. Both the window shade illusion and contrast asynchrony illusion contain alternating contrast information. The difference between the two is that in the window shade illusion, the contrast alternation occurs within a single disk, whereas in the contrast asynchrony illusion the contrast alternation occurs across two disks.

As a general rule, contrast alternation across a single object creates the appearance of motion. Curiously, the motion tends to shift towards the side of the ring with the lowest contrast (when the disk becomes white, the shade moves towards the white part of the ring, and when the disk becomes black, the shade moves towards the black part of the ring). The motion seems to track the minimum contrast in the scene (for those who like calculus, this is analogous to a change in the sign of the second derivative).

The above explanation for the motion in the window shade illusion leads to a range of questions: Why does the motion shift toward the half of the ring with minimum contrast instead of the half with maximum contrast? Why does the disk appear to rock when an inner ring is added? Why does the shading effect spread across the whole disk instead of just staying at the edges? The answers to these questions help us understand how the brain works.


Anonymous said...

This is a great illusion. It perfectly illustrates a concept that I teach my high school art students about the relativity of values. An area only appears to be dark if it is adjacent to an area that is equally light (and vice versa). I will have to use this site as an example in my classes to illustrate the idea. Thanks!

Anonymous said...

Maybe there's something the matter with me, but I just don't see any of this. It's just a disc with a flashing center...

Anonymous said...

Nice. I looked it and clicked all the buttons, didn't see any illusions. But after reading I can see them.

/*********google analytics*/ /*********google analytics end */