Max res of the human eye?

[RadicalEd]

Sorry for being slightly OT, but just out of curiosity, does anybody know the approximate resolution that a 20:20 eye maxxes out on and can no longer differentiate? I imagine sumday we'll have extremely high def TV at that resolution :cool:

[Emp3r0r]

that's an odd question but I'm assuming it would be when a white-black checkerboard pattern looks gray.... i think @ max rez on a 21inch trinitron is like this.

[ulfschack]

I heard the eye had a resolution upwards 120 Mpixels, an estimate made from counting staves and cones I guess.

Happy tech-hunting :)

cheers

[WorldBook]

Interesting question :)

When tv reaches the same quality as that of our eyes, it'll be a great time :D

[-h]

From http://newton.dep.anl.gov/askasci/comp99/CS041.htm .

Extrapolating from the dpi value, I'd say for viewing distances over 12 inches from your face, 4000x3000 would be adequate for photorealism.

Might need some work on the bitdepth though, RGB24 wouldn't cut it.

-h

[TheXung]

According to a "big data in computer graphics" lecture I sat thru, the point where you can stop anti-aliasing images when sitting 1.5 feet away is around 4000x3000 in a 15 inch monitor. Again the density is based on how close you sit.

[redcane]

it's a resolution over your field of view, or more accurately an angle between pixels corresponding to the average angle between receptors in the eye when measured from the "cross point" of the eyes focussing mechanism.

HDTVs today viewed from a normal viewing distance are close enough to the full res of the eye as far as I can tell. Hell your average monitor at it's best res, viewed from across the room is pretty well fine....

RGB24 may be a limiting factor, but since the eye tends to percieve colours with relativity it's less obvious.

Get some really high quality photos and crack them on your monitor at good res, sit back a bit and marvel at the detail you can see in them... I would be surprised if most people can discern any pixels from a bit more than a short distance away... The brain has a way of filling in data.

[-h]

RGB24 may be a limiting factor, but since the eye tends to percieve colours with relativity it's less obvious.

Oh it's just banding that I worry about. My cousin is a professional photographer, and he's shown me on many occasions how much better some of his shots look than some 2400dpi scans he ordered, particularly how badly the greys banded due to 8-bit precision.

Medical purposes usually mandate 12-bit precision for such cases.

-h

[redcane]

heh, well there'd be nothing wrong with RGB48.....

I agree there can be a difference, but generally if you have your monitor set right (Who does?) it is pretty hard to notice any banding. I guess there's no reason why we can't move to 48 though...

[AlainDelon]

I've read that the resolution on 35mm film is 4000 lines so the human eye should be higher.

[trbarry]

Maybe see Understanding image sharpness part 1:
Resolution and MTF curves in film and lenses (http://www.normankoren.com/Tutorials/MTF.html).

This is a introduction to the Modulation Transfer Function, a good measure of detail.

Somewhat technical but very interesting.

- Tom

[((( atom )))]

i wonder how much data per second the brain actually has to deal with, although /this/ surely is nearly impossible to put into bits and bytes.

still its fun to think about what we are actually doing there!

[EDIT]
...the above link is broken, could you repost it, please?

[mpucoder]

One massively parrallel processor. But, the speed is known. It's amazingly slow, around 50-100ms, or about 10 to 20 Hz. That's about .000000001 the speed of current CPUs, but it still does more (with respect to image processing) in less space, with less energy. Maybe "wetware" is the way to go. (apologies to officer Alex J. Murphy)

[redcane]

Originally posted by AlainDelon
I've read that the resolution on 35mm film is 4000 lines so the human eye should be higher.

Hmmm, 35mm film is a chemical wash, I don't know how you would work out how many lines it is.... Unless your using the wavelengths of light to figure how many distinct lines you could fit....


Same as your eye, you can only average across it's surface the number of receptors.....

Is 35mm lower resolution than the eye? if so not much, I think I can just see the lack of detail, but some photos it's damn hard. Possibly just because your brain fills in the extra,.

[robz]

@mpucoder:
The speed of the human brain is known, you say... And you even put a number on in.... 50 to 100ms. (wonder where you got this from)

More over.... 50 - 100 ms for doing what exactly??? Calulating 1+1? Thinking: "therefore I am"? Expanding the lungs just enough to let their volume grow 1 ml? To spark one electric signal from one neuron to the next? (They're called neurons aren't they??)

And 10 to 20 Hz?!? So it does... well, some basic brain-thing... at some number of times per second? What happens when the required rate is higher?? Does the brain cache the extra requests for its attention? And how does it do this since it is already operating at full capacity...

Typing this...it occurs to me that my last question is a silly one. The reason this is so incidently also explains why I think the comparison to a CPU is a flawed one...

Unlike a CPU the brain has no Ilde time. It does it's ticking *all* the time. Plus, the brain does not just tick at a certain rate idly until it gets something to do, or does it??? (well.... most people's brains I guess :D )

But then perhaps neither do CPU's?

I don't know... I just wonder...

[mpucoder]

I got the number from an article in Scientific American some years ago. And, no, the brain does not tick like a digital computer, nor does it process information serially. The figure is the propagation delay of the neural network. Stimulus arrives at one end, and a reaction occurs at the other. Conscious thought was not discussed in the article, as there was no way to measure it at the time.
For those wondering how on earth can a 10 to 20 Hz processor perform so marvelously - it is a massively parallel process. Take the visual cortex, here stimulus from all the receptors in your eyes arrive simultaneously, and are processed simultaneously in a continous flow, resulting in an image in you head about 100ms later.
And since we're on vision, the eyes preprocess the stimulus in the retina, performing adaptation and edge detection via its own neural network that compares neighboring receptors.

[Acaila]

The speed of the human brain is known, you say... And you even put a number on in.... 50 to 100ms. (wonder where you got this from)It takes about that time for a neuron to become sensitive to a new electrical signal again after having been stimulated and 'fired' as a result of the previous one. But since the brain is indeed a parallel processor this "lagtime" doesn't result in very much speed loss, and it's actually needed to function correctly.

And 10 to 20 Hz?!? So it does... well, some basic brain-thing... at some number of times per second? What happens when the required rate is higher?? Does the brain cache the extra requests for its attention? And how does it do this since it is already operating at full capacity...The average human uses about 3% of his total brain capacity. So there's no need for caching anything because apparently the brain doesn't even need to use all of it's "processor power".
Most input however is filtered. Signals that don't cross a certain threshold don't get processed at all. Most of this happens subconsciously, but if you actively ignore something it also won't get processed. Something you ignore remains in sensoric memory for a short period of time. If for whatever reason there's no need to store it away then it will be lost.

And yes the brain has idle time, but for certain areas not the entire brain at once. If no signals come in it won't have anything to do, it's just that there's always something to do you just don't notice it.

The brain (or actually the memory part) also has a nifty optimization routine. Association. With this even such an extremely slow processor can beat a super-computer's brute force method at chess :D. Whoever build the human brain sure did a nice job ;).

[robz]

And yes the brain has idle time, but for certain areas not the entire brain at once. If no signals come in it won't have anything to do, it's just that there's always something to do you just don't notice it.

That's just what I said isn't it? No idle time... otherwise one would be braindead :D
(parts of it may do little at times but that's... just parts ey?)

I don't know very much more the inner workings of the brain than I know of the way a CPU does whatever it is that is does, which is as you've probably guessed by now, not particularly much ;)