Shutter Speed Of the Human Eye ?

[Diogenes]

There is a blurb in the February issue of Discover (https://w1.buysub.com/pubs/AG/DCR/DCR_New_subpage_324647.jsp?cds_page_id=15191&cds_mag_code=DCR&id=1200668455594&lsid=80180900555035165&vid=1&cds_response_key=I8JANG&cds_mag_code=DCR) magazine about
subliminal advertising..

It recounts, that in a 1957 showing of Picnic, the messages :

“ buy Popcorn “ and “ buy Cola “

were flashed on the screen for 1/3,000 of a second, every 5 seconds …


Can the human eye quantify an image in 1/3,000 of a second ?

I come up with ~60fps as the frame rate the eye is capable of
resolving , and I suppose there is some margin of error; however
I suspect it is not a factor of 50 ….

I could be wrong ...

[PixyMisa]

I think they're making stuff up. Standard projectors run at 24 frames per second, even today. 3000 frames per second is high-speed science-type stuff, and no way would they be using that to show a feature film. (They'd need several hundred reels of film, for a start.)

Not sure the human eye has a shutter speed as such; as far as I know, there's nothing that synchronises the firing of the photoreceptors. You could certainly register a flash of light that lasted 1/3000 of a second, but actually reading a sign at that speed is exteremely implausible.

[bokonon]

You could probably find out more by searching "persistence of vision" or something similar on the web. I'm interested in the topic, but a bit too short of time at the moment to research it myself.

You wouldn't need a whole movie projected at 3000 frames/second to do the subliminal ads; I suspect a second projector with a strobe or (more likely) very fast opening and closing shutter would do it. ETA: You could even do it with a slide projector, and a revolving circle with a lens-sized slit in front of it. How fast can a power drill spin a disk?

I suspect that the answer is that there is no lower limit per se that would correspond to a "shutter speed" for the eye. There are chemical reactions and signal processing layers that might make it doubtful that an image would find its way to a useful decision-making portion of the brain.

Astronauts report seeing "flashes" with their eyes closed, which purportedly correspond to single cosmic ray events. A cosmic ray is more energetic than a photon; I think I recall that it takes 10 photons before a "flash" will register, but it may be less. In any case, these interactions are presumably happening in the amount of time it takes a particle traveling at the speed of light to cross the space of a single cell, which I suspect is several orders of magnitude smaller than your 1/3000th of a second. That's what makes me say that the retina itself is probably not a factor in any hypothetical "shutter speed" for the eye.

[Diogenes]

That grabbed me at first also, since movie film is about 24fps..

So I allowed for some type of shuttered device that superimposed an image on the screen for 1/3,000 of a second.. ( Could a mechanical shutter operate that fast ? )

That said, would it register on the retina ?


edit:

I see bokonon has addressed some of my concerns while I composed.
Thanks

Thanks ..

[lomiller]

I think they're making stuff up. Standard projectors run at 24 frames per second, even today. 3000 frames per second is high-speed science-type stuff, and no way would they be using that to show a feature film. (They'd need several hundred reels of film, for a start.)

Not sure the human eye has a shutter speed as such; as far as I know, there's nothing that synchronises the firing of the photoreceptors. You could certainly register a flash of light that lasted 1/3000 of a second, but actually reading a sign at that speed is exteremely implausible.

They actually double flash each frame so what you see on a movie/TV screen is refreshed every 1/48 of a second. In addition the screen itself has persistence so it actually displays the image longer then that.

Even 1/48 of a second only works due to the skill of the cameraman. That’s enough for a stationary object, but not enough to fully quality an object that is moving across your field of vision. In the case of a movie, the camera tracks an object in the foreground that corresponds to what the director wants you to watch so it isn’t moving in your field of vision. The background, however, is moving relative to what your eye is tracking but the camera gives this a natural motion blur so your brain accepts it as “normal”.

In a video game you don’t get this natural motion blur. There is constant talk of adding it in, but there is a fundamental problem in that you don’t know what the person playing the game is going to be watching. For example, if another stationary player shoots a rocket, are you watching the rocket or the other player? If you are watching the moving rocket, everything in the background should blur, if you are watching the stationary player, the rocket should blur.

At higher levels of play there are people who claim noticeable differences in framerates even at 120 per second or higher (which is at the edge of what most monitors can display). Mathematically you need an infinite framerate too accurately display any moveing image on a stationary screen because the screen is discreet while real world motion is analogue, so you always get some amount of sampling error. For computer graphics I’ve seen anecdotal claims that framrates of 1000+ per second may be needed for perfect realism without motion blur in very fast moving action. To me this sounds to high, but it’s difficult to test with current computer hardware.

Obviously you reach a point where the brain will simply fill in the gaps and ignore discontinuities. What happens to this information that is dropped is more in the realm of people who study the brain, and no doubt there are studies on it.

[edd]

I don't think there's exactly a shutter speed. You might measure it for example by flashing something up very briefly and figuring out how long it needs to be there - but is it a matter of it being brief or just being up long enough to put out enough light that's important? (A CCD for instance might see something very brief if it's bright enough - it'll still register and get stored until read out)
You can come up with something like a frame speed by looking at how fast some movement needs to be to look continuous, and get the 'critical fusion frequency' which would be in the 30-100Hz ballpark, but the exact nature of the flicker again has some influence.
The 1/3000 s thing seems very implausible. Not only does it seem too fast to read but it'd require a specialist high speed projector. Wouldn't be worth doing just for an advert - projectors can't be cheap as is, and putting out a whole bunch of state of the art ones just for one or two adverts with dubious effect seems unlikely.

[patnray]

I believe the original story was a hoax
Subliminal Advertising (http://www.snopes.com/business/hidden/popcorn.asp)

[Diogenes]

Your link says the published results were fraudulent. Not that the experiment was not performed..

[patnray]

From the Urban Legends linK:
You see, Vicary lied about the results of his experiment. When he was challenged to repeat the test by the president of the Psychological Corporation, Dr. Henry Link, Vicary's duplication of his original experiment produced no significant increase in popcorn or Coca-Cola sales. Eventually Vicary confessed that he had falsified the data from his first experiments, and some critics have since expressed doubts that he actually conducted his infamous experiment at all..
The link also explains how images were supposedly flashed for 1/3000 of a second.

[Dancing David]

I think they're making stuff up. Standard projectors run at 24 frames per second, even today. 3000 frames per second is high-speed science-type stuff, and no way would they be using that to show a feature film. (They'd need several hundred reels of film, for a start.)

Not sure the human eye has a shutter speed as such; as far as I know, there's nothing that synchronises the firing of the photoreceptors. You could certainly register a flash of light that lasted 1/3000 of a second, but actually reading a sign at that speed is exteremely implausible.


The flicker rate is created by the plate with holes in it in the projector. I think that there are two flickers per frame so 48/sec.

That is above the threshold for most people, but there is that one percent that would want the flicker rate to be above 60/sec.

But recognition of something at 1/3,000, not very likely the rate of cycles in the power grid is set at 60 cps because then almost all people will not perceive the flicker.

[bokonon]

It isn't necessarily a question of perceiving flicker, is it?

Let's say you were in a dark room. A small white dot is projected on a white wall, and you're told to focus on it.

While you're focused on it, a strobe light flashes for 1/3000 of a second, illuminating a word immediately above the dot.

If the word was known to you, would you be able to recognize what was?

I think you probably would, even if the light only flashed for 1/3,000,000 of a second. If it was bright enough, and not too bright, there would be an after-image on your retina which you'd be able to "read" even after the room returned to darkness, just as you "see spots" after a flash bulb goes off in your face.

[bjb]

A camera flash lasts about 1/1,000 of a second or so and is certainly visible to the human eye:

http://en.wikipedia.org/wiki/Flash_(photo)

[Furious Coder]

I'm not a biologist, but I think what you have to take into consideration is not the "shutter speed" of the eye, but the sensory update rate of the brain. I'd have to go looking for the figures, but the numbers I've seen show a range of 50Hz-70Hz for how often our brains "poll" our sensory organs and build up a new picture of the environment.

This can be demonstrated with some interesting tests, and yes, it shows that we can "see" things that happen ultra fast, but it must be under certain conditions. For example, humans can hear audio events that last around 2ms, but we have trouble differentiating between events that are less than around 10ms apart. Likewise, a bright flash that is very very brief will register with our eyes and be visible, but if there were multiple bright flashes within around 10ms, we'd see them as a single flash.

There's some "persistance of vision", the biochemical actions of our cones and rods, plus the mechanical vibrations within our ear that take time to die down, that are responsible for the ability to percieve extremely short term events. A brief bright flash will stimulate our retnas, which have an extinction period before which they will stop sending signals related to that event. The event will still register and cause neurons to fire, which will let us percieve the event.

However, it is the "update rate" of our brains that is the final determant of how fast we are able to percieve separate events placed closely together.

[Furious Coder]

Double Post.

[Furious Coder]

Sorry, browser wasn't updating correctly.

[Dancing David]

It isn't necessarily a question of perceiving flicker, is it?

Let's say you were in a dark room. A small white dot is projected on a white wall, and you're told to focus on it.

While you're focused on it, a strobe light flashes for 1/3000 of a second, illuminating a word immediately above the dot.

If the word was known to you, would you be able to recognize what was?

I think you probably would, even if the light only flashed for 1/3,000,000 of a second. If it was bright enough, and not too bright, there would be an after-image on your retina which you'd be able to "read" even after the room returned to darkness, just as you "see spots" after a flash bulb goes off in your face.

Yes, the photropins do take time to recharge.

Um, I was answering the vague shutter speed question. The persistence of vision threshold varies from person to person, 99.99% will experience a continuous visual field at 60Hz. So I was discussing the threshold where the discontinuous visual field becomes a continuous visual field.