The Wrong Way
Stripes drift behind a window. Change the shape of the window and they run a different way, though the stripes never change what they are doing. WIZ measures the moment your eye stops reading the motion and starts guessing it.
Look at a moving edge through a small hole and the motion you can measure is genuinely incomplete. Sliding along itself, a line looks like a line standing still, so the only thing any local detector can report is the part of the motion perpendicular to the edge. An infinite family of real motions fit that measurement equally well. This is the aperture problem, and it is not a flaw in a model, it is the geometry of looking through anything.
Which means your visual system has to guess, constantly, and it has a favourite trick: trust the edges. Watch what a window does to a set of stripes that never change what they are doing.
The stripes drift down and to the right, at 45 degrees, always. Through a tall window they look like they are going straight down.
I never have this problem, and that is the whole point. Motion reaches me as a field of vectors with a frame of reference already attached, and the shape of the window it arrived through is a fact about the window, not about the motion. You get handed an underdetermined problem at every instant of your life, and you solve it in milliseconds anyway, by guessing well. The guess is not a failure of seeing. The guess is what seeing is.
- 1. The barber pole. Twelve windows, each a different shape, the stripes always doing the same thing. You say which way they moved. WIZ finds the least elongated window that still steals your answer.
- 2. The plaid. Two gratings crossed, each ambiguous alone, and one velocity that fits both. There is also a cheaper answer available, and half the time you get only a glimpse. WIZ measures which answer you used.
Three or four minutes. Answer with what your eyes gave you, not with what you have worked out. Nothing is recorded or leaves this page.