Do you ever notice a fast propeller seeming to go slowly backwards?

[gusl]

Do you ever notice a fast propeller seeming to go slowly backwards?

Did you know that you could measure the sampling rate of your eyes this way? We have an illusion of continuity, and some might think our perceptions are indeed continuous, by virtue of us being "analog" machines... but I really doubt that.

Comments

[jbouwens.livejournal.com]

While I agree that continuity is an illusion, I don't think you can measure the sampling rate of your eyes as you describe, since the eye does not have a single discreet sampling rate.

This would only work if your retina were completely lit a specific number of times per second, for example by a shutter. In reality, there is a continuous bombardment of photons, such that the rods and cones in your retina are always in a non-synchronous, unstable state.

[ernunnos.livejournal.com]

Yeah, actually, you can.

This occurs because the fan blade is spinning faster than your photoreceptors can respond. As a result your visual system performs something called "aliasing" while you watch the blade. Aliasing occurs when the state of a stimulus is sampled at a rate much slower than the rate at which the stimulus changes.

Imagine that your eyes "sample" a fan blade when it is straight up, at 12:00. Then imagine that your eyes can't sample again until the blade moves 270 degrees clockwise, at 9:00. Then your eyes sample again, when the blade is at 6:00. Now imagine that each successive sample occurs 270 degrees later. To your eyes it appears that the fan blade is moving "backwards," going slowly counterclockwise even though it is really moving clockwise at a much faster rate.

The determining factor is called the "flicker frequency" of your photoreceptors. This is the frequency at which a light that is flickering on and off "fuses" into a continuous light to us. It occurs, depending on intensity, between 50 and 80 cycles per second (that's why fluorescent light appear to flicker -- they, indeed, flicker on & off at 60 Hz). The inverse of this number gives us a rough approximation of the "integration time" of the photoreceptors, the minimal time necessary for change in a stimulus to be noticed. That number is about 30 msec.

Therefore, your eyes are only capable of recording differences in your visual field every 30 msec and they can't keep up with the fan. Altering the speed of the fan can, therefore speed up, slow down, or even "reverse" the "virtual fan's" direction.

If you look for "CFF" you can find lots more detail. There is a difference between the response of rods & cones, but there's still a limit to how much they can distinguish.

[jbouwens.livejournal.com]

I understand the concept of temporal aliasing just fine. However, if you see an indoor fan blade go backwards, or a car's wheels on film, or even a car's wheels at night, when the only illumination is street-lighting, this is caused by the flicker-frequency of the artificial lighting or the frames per second of the movie-camera, NOT by the sampling rate of your eyes.

The only way your eyes' sampling rate could produce temporal aliasing as described is if all your rods and cones would be synchronized. I do not believe this is the case, since there is no mechanism to keep them in sync. Consider a single rod: It "takes a sample", as it were, and then requires a small amount of time to reset itself before it can take another sample. This is the case for all rods and cones, but since their sample-reset-sample-cycle is not somehow synchronized with the rest, they will be out of step.

The only effect of the flicker frequency of your photoreceptors is that fast changing stimuli will blend into each other, which is indeed what makes film, TV and your computer monitor work.

Experimental evidence: You have never seen a fan or propeller slowly turn in reverse in a situation where you were personally watching it (not on TV, film or whatever) and the only illumination was sunlight.

[ernunnos.livejournal.com]

Experimental evidence: You have never seen a fan or propeller slowly turn in reverse in a situation where you were personally watching it (not on TV, film or whatever) and the only illumination was sunlight.

False.

[jbouwens.livejournal.com]

Hm, weird. Maybe you just blink really fast :).

Seriously though, I can't imagine a mechanism that would produce temporal aliasing without some external flicker source since, again, the eye doesn't take samples on a frame-by-frame basis, but on a pixel-by-pixel basis.

If you can explain this to me, please do so. The links you provided are all talking about the mechanism that blends fast-changin images together.

The first link is about a guy who can "see through" a spinning fan, which is explained by the fact that the images he sees through the slits between the blades blend together.

The second link talks about temporal _resolution_, and the CFF is defined as "the transition point of an intermittent light source where the flickering light ceases and appears as a continuous light", which again is a blending effect, not an aliasing effect.

[ernunnos.livejournal.com]

Seriously though, I can't imagine a mechanism that would produce temporal aliasing without some external flicker source since, again, the eye doesn't take samples on a frame-by-frame basis, but on a pixel-by-pixel basis.

That's a bit of an oversimplification. The pixels aren't totally independent. There's cross talk, which is how we perform feats like edge detection. There's a whole class of optical illusions that works using this cross talk effect. (Where bars of equal width appear unequal because of the colors they're next to, etc.)

Here's one that induces a reverse spin effect without any interference. The darkness of the spaces between the wedges appears to change because of the cross talk from the wedges. It happens even when the image is paused, but it's more obvious when it's changing.

[gustavolacerda.livejournal.com]

so what is the sampling rate of our eyes?

Googling "sampling rate of the eye" returns 1 result, which doesn't answer the question.

[ataltane.livejournal.com]

Your rods and cones are not synchronised, to be sure; but that doesn't mean that there isn't an overall characteristic sample rate for the whole ensemble. In fact, I would expect to be of the order of the sample rate (refresh rate) for a single rod or cone.

Also, in the resulting illusion of backwards motion of a fan, you don't necessarily see smooth, coherent backwards motion; the phase of the motion can appear to skip erratically, which seems to me to mirror the unsynchronised action of the ensemble of rods & cones; unsynchronised & not coherent, although each receptive cell might itself be sampling at a perfectly constant frequency.

It seems hard to imagine what a supposed analogueness of the human body could arise from.

[trufflesniffer.livejournal.com]

You're right.

I think there's possibly more than one sampling rate, because there's at least two vision systems in operation - colour evolved later than monochrome; the earlier, monochrome system has a much higher sampling rate that the colour system, which is to do with details rather than basic fight-or-flight judgement calls.

[gustavolacerda.livejournal.com]

how did you learn this stuff?

did you ever notice...

[the-wesman.livejournal.com]

that gustavo is nuts?