Visual Perception: Investigating the Speed of Seeing

[Chen]

"Speed" of seeing?

Let's say a tennis ball is crossing your field of sight at a great speed. How fast does it have to travel for you not to notice it at all? I guess what I'm asking is, how fast do "pictures" in the human mind update? What is our inner FPS? (And can we improve it with a better graphics card? )

I realize that this changes for every person, but is there some kind of average value?

 

[pallidin]

That's a great question. Maybe someone more knowledgeable than I can answer.

I do know that it varies with distance and size of the moving object.
For example, if I place my hand about 1-foot away from my eyes and move it sideways at about 3-5 mph, that my eyes see a translucent blur.
In contrast, I can see a passenger jet 30,000 ft above moving at 600 mph with no blurring at all.

 

[Chen]

Thank you pallidin. I think we should distinguish between two concepts of speed: "Apparent speed" and "relative speed". Relative speed, in this case, is obviously the speed of the object relative to your eye, whereas apparent speed depends on the distance of the object from you and the time it takes to cross your whole field of sight.

 

[wimms]

If spaceship passed your view in less than 20ms, you wouldn't become aware of this, no matter how close of far from you it happens. On similar manner, if object accelerates fast enough to leave your view in 20ms, to you it would appear as if it just vanishes. Quite simple to try.

Our inner FPS isn't clearly fixed. In one sense, you could say 25fps, where we stop noticing discrete frames, then we could say around 100fps, where we stop noticing flicker. But it goes beyond, some say 400fps isn't limit to noticing changes in the stage. And it doesn't depend only on person, but very much on the visual material and its dynamics.

 

[chroot]

From what I understand about the neurophysiology of sight (i.e. not too much), the brain takes "pictures" roughly 1000 times a second. That does not mean that you'll generate enough of a neural response to recognize something that flashes past your eyes in a millisecond, however.

- Warren

 

[wimms]

1000? not 100?

 

[chroot]

From what I've heard, physiologically speaking only, the eye does 1k frames per second. But I'm no expert.

- Warren

 

[Scott Sieger]

This is interesting to me in that I was under the impression that perception was continuous and not in "frames" like a video camera.

Out of curiocity Chroot, do you know how they have the managed to get a figure like 1000 frames per second?

 

[jdavel]

This is interesting to me in that I was under the impression that perception was continuous and not in "frames" like a video camera.

Out of curiocity Chroot, do you know how they have the managed to get a figure like 1000 frames per second?

That number (1ms/frame) must be something like the minimum time for the signal coming from a single rod or a cone to change by the minimum amount that the brain can detect. As chroot said, I don't think the time something needs to remain in the visual field for us to be able to tell it was there is as short as a millisecond. Try wathcing those numbers go from 1 to 100 in one second on a digital stop watch. You couldn't exactly say whether one of them was missing or not, and that's .01 sec!

Also, when it comes to the brain, I'd be careful about thinking too much in terms of anything as neat and tidy as a "frame".

 

[Scott Sieger]

Also, when it comes to the brain, I'd be careful about thinking too much in terms of anything as neat and tidy as a "frame".

Actually this was just my question...Maybe Chroot was making an analoguous statement rather than fact, and I guess I have to assume this...
I thought possibly it had been proven that the brain see's in "frames" and this proof would be interesting to see...

 

[Gara]

i think its like this:

you can't tell if it always misses the number 0.005.

but you'd know if something moved. if something was on 0.005 for awhile, then went to 0.006 then back to 0.005, in 0.001 seconds, you might go "...wait, did that just move?"

 

[wimms]

Of course "frames" is just analogy. Brain processes input from eyes in about 100fps. Thats why I wondered 100 vs 1000. But yes, eyes are doing some of processing before passing on to the brain, so 1000fps may be more correct. But this also is only in some restricted view, as eyes don't have uniform "precision" not "speed" across all of the view field.
This 100fps ability from brain is imho responsible for fact that if whole image is changed in less than 10ms and back, brain has not enough data to detect anything. Yet we can notice smaller changes in a manner that "something happened" without being able to realize what exactly.

 

[Nereid]

There's also a magnitude effect - a very large signal will leave an 'echo' in the rods, cones, neural processing circuits, etc, even if the signal only 'lasts' <1ms. Or, in other words, the overall visual system has an 'integration' capability (which, IIRC, is put to good use by keen visual amateur astronomers!)

 

[mrdex]

Watch a car overtake you on a motorway. Quite often the wheels of the car appear to be going backwards or even to stop. I suspect experiments of this nature could be used to find how quickly we see things.