Human Eye FPS

[phongn]

Image recognition in 1/220th of a second is not the same thing as the brain being able to process some 200 "frames" per second.

Anyone have something out of a medical journal or something?

[Admiral Valdemar]

The human eye cannot discern frames like a computer or film reel does, it is a nonsensical question, it's really what frame rate can we notice flicker and so on since the eye and visual cortex don't work that way as the brain does not slice every image feed we get into a framerate.

The real question is beyond what framerate can we not notice a difference, the same as what bitrate can we not hear a difference and resolution and so on.

I'll get back on this though.

[Rye]

The human eye cannot discern frames like a computer or film reel does, it is a nonsensical question, it's really what frame rate can we notice flicker and so on since the eye and visual cortex don't work that way as the brain does not slice every image feed we get into a framerate.

The real question is beyond what framerate can we not notice a difference, the same as what bitrate can we not hear a difference and resolution and so on.

I'll get back on this though.

Well there must be limits to work out a relative framerate. For example if something's framerate is too slow it seems jerky, if something moves too fast for the eye to follow, you get motionblur and so on.

[kojikun]

Well there must be limits to work out a relative framerate. For example if something's framerate is too slow it seems jerky, if something moves too fast for the eye to follow, you get motionblur and so on.

you must also take into consideration the variances in eye frame rate and displayed video frame rate.

Eye ------- Image ------- Effect
Fast--------Fast-----------Smooth motion
Fast--------Slow----------Jerky motion
Slow-------Fast-----------Motion blurring (possibly combined with jerky motion)
Slow-------Slow----------Jerky motion

Also, we must also remember that our perceptions of time affect this as well. Slow eyes and slow image rate doesnt mean smooth motion because we're still perceiving a considerable amount of time between the two pictures and the amount of change theyre undergoing.

[Howedar]

Watch a quickly spinning ceiling fan sometime. Then, close your eyes. You can see for a fraction of a second the image right before you closed your eyes. You can count the blades of the fan, while you can't when watching the streaming video.


I don't know what if any relevance this has, but its cool.

[kojikun]

Watch a quickly spinning ceiling fan sometime. Then, close your eyes. You can see for a fraction of a second the image right before you closed your eyes. You can count the blades of the fan, while you can't when watching the streaming video.


I don't know what if any relevance this has, but its cool.

its creating artificial jerky motion when you do that, or blink your eyes real fast.

[Warspite]

Image recognition in 1/220th of a second is not the same thing as the brain being able to process some 200 "frames" per second.

Anyone have something out of a medical journal or something?

On an extended Google search (about 400 hits for "human eye test frame rate") I've only saw pages about games, cameras, and video compression papers, no medical papers, not even the "famed" Air Force research paper. (hum.... This must mean something...)

[Howedar]

Watch a quickly spinning ceiling fan sometime. Then, close your eyes. You can see for a fraction of a second the image right before you closed your eyes. You can count the blades of the fan, while you can't when watching the streaming video.


I don't know what if any relevance this has, but its cool.

its creating artificial jerky motion when you do that, or blink your eyes real fast.

No, you close your eyes and keep them closed and see the image when its still stored in your brain's cache.

[kojikun]

oh uh.. i dont see that howwie.. i think you should lay off the LSD...

[AdmiralKanos]

The 200fps reference uses a bright picture flashed briefly in a dark room in order to "prove" that the human eye and brain can process 200 frames per second, disregarding the "residual after-image" effect of the way we process visual information.

What they neglect to mention is that if the same picture were inserted into a movie, you would never be able to identify what it was. The only reason people could identify details of the picture flashed in the dark room was that there was nothing else to process, hence the system's slow processing ability only had to work with the one picture (aided by the after-image).

[BoredShirtless]

tests on Air Force pilots have shown their ability to not merely notice, but identify the type of aircraft when shown an image for only 1/220th of a second

That doesn't prove a framerate of 200 fps, however.

Yeah it does. The question was what FPS can a human eye register. If a human can not only see but identity a plan at 1/220th of a second, then the image was registered.

A camera with a rate of 30 fps could pick up an object that appears for less than 1/30th of a second if it is bright enough. Remember, the eye is collecting light over the entire frame. It's not an instantanous snapshot like a frame from a video game.

Red herring, we're talking about eyes, not cameras.

[BoredShirtless]

The 200fps reference uses a bright picture flashed briefly in a dark room in order to "prove" that the human eye and brain can process 200 frames per second, disregarding the "residual after-image" effect of the way we process visual information.

What they neglect to mention is that if the same picture were inserted into a movie, you would never be able to identify what it was. The only reason people could identify details of the picture flashed in the dark room was that there was nothing else to process, hence the system's slow processing ability only had to work with the one picture (aided by the after-image).

Or in other words, the FPS our brain processes images is not the same as the FPS our eyes can register images.

[Dahak]

The 200fps reference uses a bright picture flashed briefly in a dark room in order to "prove" that the human eye and brain can process 200 frames per second, disregarding the "residual after-image" effect of the way we process visual information.

What they neglect to mention is that if the same picture were inserted into a movie, you would never be able to identify what it was. The only reason people could identify details of the picture flashed in the dark room was that there was nothing else to process, hence the system's slow processing ability only had to work with the one picture (aided by the after-image).

Or in other words, the FPS our brain processes images is not the same as the FPS our eyes can register images.

The brain is no GPU, and doesn't work in "fps" or somesuch.

[BoredShirtless]

The 200fps reference uses a bright picture flashed briefly in a dark room in order to "prove" that the human eye and brain can process 200 frames per second, disregarding the "residual after-image" effect of the way we process visual information.

What they neglect to mention is that if the same picture were inserted into a movie, you would never be able to identify what it was. The only reason people could identify details of the picture flashed in the dark room was that there was nothing else to process, hence the system's slow processing ability only had to work with the one picture (aided by the after-image).

Or in other words, the FPS our brain processes images is not the same as the FPS our eyes can register images.

The brain is no GPU, and doesn't work in "fps" or somesuch.

Right, the brain doesn't work like a GPU. But that's irrelevant to the question, which is what speed can the eye register images. Which is again different to the speed our brain can process images.

The Air Force experiment shows our eyes can register images at 1/220th of a second. But can our brains process that fast? The answer is no. The pilots managed to process the image because no images preceeded or followed it. The brain was clear to concentrate on that one image. The experiment would have failed if the image was inserted between other images.

[Mad]

That doesn't prove a framerate of 200 fps, however.

Yeah it does. The question was what FPS can a human eye register. If a human can not only see but identity a plan at 1/220th of a second, then the image was registered.

A 30 fps camera could theoretically register an image that appears for 1/220th of a second, that does not mean the camera can make out 220 separate images per second.

Red herring, we're talking about eyes, not cameras.

The same basic principles are involved, although they are implimented differently. Light stimulates a surface for a short period of time, creating a picture. Anything that moves in that period of time will have a motion blur effect associated with it.

[BoredShirtless]

That doesn't prove a framerate of 200 fps, however.

Yeah it does. The question was what FPS can a human eye register. If a human can not only see but identity a plan at 1/220th of a second, then the image was registered.

A 30 fps camera could theoretically register an image that appears for 1/220th of a second, that does not mean the camera can make out 220 separate images per second.

There are fundamental differences between eyes and cameras, which makes it invalid to take the limitations of a camera to apply to the eye.

For one, information is continuously streamed to our eyes; a camera is a discrete device [unless you leave the shutter open]. The fact we receive data continuously means our eyes don't need to be in synch like a 30fps camera needs to do to faithfully register a 200fps image. In other words, unless you take the shot and release the frame at exactly the same time, a 30fps camera needs to get lucky to register a 220fps image. But our eyes will ALWAYS register an image at that speed [proof: Air Force experiment]. Remember, the question was what FPS can our eyes register.

So can a 30fps camera register 220 frames in one second? Of course not. But can our eyes? Yes. But can our brains consciously process that many images in one second? Probably not [depending on the images complexity and training. I'm loath to say it's impossible].

[kojikun]

For one, information is continuously streamed to our eyes;

Not entirely. Our eyes are kind of like video cameras -- they output a signal at a given rate (the neuron recharge time), but the input to the sensor is constant. The main difference between our eyes and a shutter camera is that our eyes don't go to black but rather to the next input period, while a shutter camera does go to black.

a camera is a discrete device [unless you leave the shutter open]. The fact we receive data continuously means our eyes don't need to be in synch like a 30fps camera needs to do to faithfully register a 200fps image.

But you're missing the point. Our eyes can detect anything going at any rate of appearance, as could a shutter camera. The only difference is that a shutter camera has a period of blackness between input periods. A video camera does not have this problem and so any analogy between the human eye and a digital camera of any kind is correct.

In other words, unless you take the shot and release the frame at exactly the same time, a 30fps camera needs to get lucky to register a 220fps image. But our eyes will ALWAYS register an image at that speed

This also is assuming that the rods and cones don't have a very short period of inactivity. There are atleast some instances during vision, called saccades (those jerking movements of your eyes when reading, etc), during which your vision shuts down entirely for a very short amount of time, but it goes unnoticable. I'm not sure if there are similar periods between rod-cone recharge during which input is ignored, but if there is then there ARE synchronisation times.

[proof: Air Force experiment]. Remember, the question was what FPS can our eyes register.

So can a 30fps camera register 220 frames in one second? Of course not.

Sure it can. They'll just blur into one another.

But can our eyes? Yes.

Nope. The input will all hit the rods and cones during a single input cycle and will blur exactly the same way it would in the camera.

But can our brains consciously process that many images in one second? Probably not [depending on the images complexity and training. I'm loath to say it's impossible].

Why is impossible bad to say when you've been making unconnected claims of possibility during this entire arguement?

[Mad]

In other words, unless you take the shot and release the frame at exactly the same time, a 30fps camera needs to get lucky to register a 220fps image.

Take a dark room, with a video camera at 30 fps. Start the video camera. Flash a bright image for 1/220th of a second. The image should show up on one frame of the video, because it was the only light that was registered for that frame. (Unless the flash happens to occur in between the collection periods for frames; I'm not sure how long that delay is.)

But our eyes will ALWAYS register an image at that speed [proof: Air Force experiment]. Remember, the question was what FPS can our eyes register.

That's only because there's no delay period where no data is being taken in. The only reason that image showed up was because the image was bright enough to register in a fraction of a second.

If the image was brighter and shown for an even shorter period of time, the same thing should occur: enough photons would register with the rods and cones in the eye to produce an image. (There's no other light source to interfere with and overwrite the image.) It's a relationship between intensity and time, not "frames per second."

So can a 30fps camera register 220 frames in one second? Of course not. But can our eyes? Yes. But can our brains consciously process that many images in one second? Probably not [depending on the images complexity and training. I'm loath to say it's impossible].

You're applying the experiment incorrectly. The eye can only notice an image flashed for 1/220th of a second if it is bright enough to not be interferred with by the rest of the "frame" in a period of time between refreshes.

[BoredShirtless]

For one, information is continuously streamed to our eyes;

Not entirely. Our eyes are kind of like video cameras -- they output a signal at a given rate (the neuron recharge time), but the input to the sensor is constant.

Did you read what I said? Information is continuously streamed to our eyes IS entirely true, and you even agree with me "but the input to the sensor is constant". No where in my statement do I discuss output.

The main difference between our eyes and a shutter camera is that our eyes don't go to black but rather to the next input period, while a shutter camera does go to black.

More or less.

a camera is a discrete device [unless you leave the shutter open]. The fact we receive data continuously means our eyes don't need to be in synch like a 30fps camera needs to do to faithfully register a 200fps image.

But you're missing the point. Our eyes can detect anything going at any rate of appearance,

That's our eyes response time, not its FPS rate. FPS requires full saturation of the visual cortex, which won't occur if the object is moving fast enough.

And it's ridiculas to say our eyes have a response time of 0. Not that anything can, but if something COULD move fast enough, we won't detect it.

as could a shutter camera.

Thanks for repeating one of my points.

The only difference is that a shutter camera has a period of blackness between input periods. A video camera does not have this problem and so any analogy between the human eye and a digital camera of any kind is correct.

Yeah, we can create analogies up to our throats. But analogies aren't very important when it comes to debating science.

In other words, unless you take the shot and release the frame at exactly the same time, a 30fps camera needs to get lucky to register a 220fps image. But our eyes will ALWAYS register an image at that speed

This also is assuming that the rods and cones don't have a very short period of inactivity.

The response time of the retina [i.e. rods and cones] is about a picosecond. The odds of enough rods and cones being "inactive" to make a difference during an image displayed for 1/220th of a second, are astronomical. Or: it ain't gonna happen.

There are atleast some instances during vision, called saccades (those jerking movements of your eyes when reading, etc), during which your vision shuts down entirely for a very short amount of time, but it goes unnoticable. I'm not sure if there are similar periods between rod-cone recharge during which input is ignored, but if there is then there ARE synchronisation times.

Synchronisation is not an issue for 1/220th of a second.

[proof: Air Force experiment]. Remember, the question was what FPS can our eyes register.

So can a 30fps camera register 220 frames in one second? Of course not.

Sure it can. They'll just blur into one another.

My fault, should have been clearer: a 30fps camera can NOT register 220 frames without superimposing frames on top of one another.

But can our eyes? Yes.

Nope. The input will all hit the rods and cones during a single input cycle and will blur exactly the same way it would in the camera.

What a crock of shit. You're seriously underestimating the speed of our visual system. Like I said above, our retina has a response time of about a picosecond. Add time taken to travel along the optic nerve to our visual cortex, and unlike a 30fps camera, our eyes WILL capture each of those 220 images without blur.

But can our brains consciously process that many images in one second? Probably not [depending on the images complexity and training. I'm loath to say it's impossible].

Why is impossible bad to say when you've been making unconnected claims of possibility during this entire arguement?

Don't be a smart arse.

[BoredShirtless]

In other words, unless you take the shot and release the frame at exactly the same time, a 30fps camera needs to get lucky to register a 220fps image.

Take a dark room, with a video camera at 30 fps. Start the video camera. Flash a bright image for 1/220th of a second. The image should show up on one frame of the video, because it was the only light that was registered for that frame. (Unless the flash happens to occur in between the collection periods for frames; I'm not sure how long that delay is.)

Exactly what I said, it needs to get lucky.

But our eyes will ALWAYS register an image at that speed [proof: Air Force experiment]. Remember, the question was what FPS can our eyes register.

That's only because there's no delay period where no data is being taken in. The only reason that image showed up was because the image was bright enough to register in a fraction of a second.

Nope. As I pointed out to kojikun, we can register images much faster then 1/220th of a second; the response time of our complete visual system is roughly a picosecond plus time along optic nerve to visual cortex.

But can we "consciously process" at that speed [NOT visual processing: analytical]? Or put in another way: could those pilots identify each plane in 220 pictures, all in the space of one second? No way.

If the image was brighter and shown for an even shorter period of time, the same thing should occur: enough photons would register with the rods and cones in the eye to produce an image. (There's no other light source to interfere with and overwrite the image.) It's a relationship between intensity and time, not "frames per second."

Yes intensity of light naturally affects our eyes. And yes, duration too. So yeah, I agree. But don't you see that fps is ALSO explicitly related to time, and implicitly to intensity? Your relationship with time and intensity is right, but doesn't preclude an fps analysis.

So can a 30fps camera register 220 frames in one second? Of course not. But can our eyes? Yes. But can our brains consciously process that many images in one second? Probably not [depending on the images complexity and training. I'm loath to say it's impossible].

You're applying the experiment incorrectly. The eye can only notice an image flashed for 1/220th of a second if it is bright enough to not be interferred with by the rest of the "frame" in a period of time between refreshes.

Wrong. See above.

[Mad]

Exactly what I said, it needs to get lucky.

As kojikun noted, however, a digital video camera would not have this limitation, and thus would not need to "get lucky."

Nope. As I pointed out to kojikun, we can register images much faster then 1/220th of a second; the response time of our complete visual system is roughly a picosecond plus time along optic nerve to visual cortex.

This would only happen if enough light enters the eye in that time period. And, like I said, an experiment with brighter objects and even less time should be possible, thus leading to you thinking even higher "framerates" as possible.

Yes intensity of light naturally affects our eyes. And yes, duration too. So yeah, I agree. But don't you see that fps is ALSO explicitly related to time, and implicitly to intensity? Your relationship with time and intensity is right, but doesn't preclude an fps analysis.

But given that a digital video camera with a framerate of 30 fps could give the same results that test gave, any conclusion resulting in "220 fps" is in error unless the eye is accomplishing something a digital video camera at 30 fps cannot, but at 220 fps could.

[kojikun]

Bored: Human eyes do not run at 220fps. Period. You cannot SEE 220fps, you never will be able to. You can perceive things that occure for only 1/220th of a second, but if two frames appeared 1/220th of a second appart from one another and lasted the same length, they would appear to be ONE FRAME because out eyes out output to our brains once every 30th of a second or so. Its the same for videocameras and film cameras: during a single input/output cycle, anything regardless of exposure time will appear in the image, the problem arises when it vanishes and is washed over by 220 times the light from something else. Human eyes cannot perceive motion accuring during 1/220th a second but it cant detect the existance of an object at those speeds.

[Sam Or I]

Anyone see Fight Club?

In the begining where they flash Brad Pitt for only 1 frame, you can easily make him out running at 30 FPS even though everything else was moving smoothly in the picture, with background lighting.

[Darth Wong]

Anyone see Fight Club?

In the begining where they flash Brad Pitt for only 1 frame, you can easily make him out running at 30 FPS even though everything else was moving smoothly in the picture, with background lighting.

30fps? We're arguing about a preposterously high 200fps number which is more than 6 times that.

[Darth Wong]

Nope. As I pointed out to kojikun, we can register images much faster then 1/220th of a second; the response time of our complete visual system is roughly a picosecond plus time along optic nerve to visual cortex.

Don't be ridiculous; you are acting as though the chain is as strong as its strongest link, not its weakest link. Do we measure video camera frames per second by the time required for electricity to travel through its fucking wires?