How many PIXELS is the Human Eyeball?

[Shannon Rawls]

Anybody know? *smile*

- Shannon W. Rawls

[Michael Struthers]

None. Unless you jab yourself with a sharp jpg.

[Michael Struthers]

But seriously folks...

I've heard the number 8k bandied about. 35mm is "about" 4k, is it not?

[Michael Struthers]

I meant 8k resolution equivalent.

[Daniel Deck]

About 576 Megapixels according to Clarkvision Photography. Go to:

http://clarkvision.com/imagedetail/eye-resolution.html

[Steven White]

I read through that page, and the "numerical equivalents" are stupid.

The eye is actually a very very bad camera, with the luminosity information provided by the rods, and the colour information provided by the cones (or is it the other way around?).

While it is quite sensitive to low light (basically the eye can count single photons) the aperture is rediculously small, making it quite inefficient at light gathering. The lenses are pretty low quality.

However, thanks to the amazing software provided by the brain, the eyes scan all over the place and "pixel-shift" like crazy to sample edges, resolve fine detail and increase the perceived resolution. I've seen a video of the rods and cones on the retina trying to resolve a light pattern projected by a laser in the last physics conference I went to. It's amazing how erratic the eye movement is. Imagine if you were to take a jackhammer to your video camera - that's pretty much what your eye does.

In addition, while the FOV of the eye is very large (~nearly 180 thanks to peripheral vision), you can only focus and resolve detail in a very small area at any one time. Your brain fools you into believing the rest of the world is that sharp, by resolving it when you choose to look there.

Personally, I think the best answer is that resolution is meaningless in the context of the eye. How long would it take you to resolve 1 line pair per mm at a distance of 1 foot? If it takes more than 1/60th of a second, how can you even compare this to a video or a still camera?

[Rhett Allen]

I remember reading somewhere that to reproduce a similar image of a high quality 35mm film camera you would need an 80 MegaPixel digital camera! I will see if I can find it again but it was a pretty old article, probably written during the early stages of digital photography.

[Luis Caffesse]

"The eye is actually a very very bad camera...the aperture is rediculously small....the lenses are pretty low quality."

And for the life of me I can't figure out how to turn off the 'auto white balance' feature!

:)

[Chris Hurd]

Yeah, but you've gotta love the built-in 3D feature, not to mention native 16:9.

[Steve Crisdale]

<<<-- Originally posted by Chris Hurd : Yeah, but you've gotta love the built-in 3D feature, not to mention native 16:9. -->>>

and the perfect colour balance.... though any defects of manufacture in hardware or software aren't covered by a warrantee of any sort!!

[Filip Kovcin]

and how about updates? any news?

... and - on which version we are currently working?


and interested one - the software is region free - no PAL, SECAM or NTSC or different frequency rates...

who is the owner of the copyrights?


filip

[Pete Bauer]

I'm sure most everyone has seen the various round-n-round discussions that people get into about the equivalent pixel resolution of film. Well, the visual system is so much different than either film or video that those kind of debates would never be resolved (BTW, that's an optics pun).

The clarkvision article linked above does provide some interesting information, but goes too far in terms of equating visual acuities to pixel counts; some of those confounding points have already been mentioned, such as the super "pixel shift" of visual scanning and the very limited area of sharp central vision (just a couple of degrees wide). The old demo of that is to hold a pen at arm's length at fixate on one end. You cannot read the writing on the shaft of the pen just 5 or 6 cm away without diverting your gaze.

Standard Snellen eye charts are designed such that the 20/20 line represents a visual resolution of one arc-minute (the width of the lines in the letters is one arc-minute, so the "E" is 5 arc-minutes in total height -- 3 "arms," 2 spaces high). Some people can see as much as twice that sharp, 20/10, or resolving power down to 30 arc-minutes. If you accept that as optimum performance, you can determine an approximate pixel count for foveal (central) visual acuity.

Still there will be all kinds of caveats because of the complex way the human visual system works. For example, high contrast linear targets (like distant power lines) can often be resolved at much finer resolutions -- presumably a real neural pixel shift. Another interesting phenomenon is that although your peripheral vision is terrible in terms of resolution, it is exquisitely sensitive to motion; computer geek example: ever notice how you can see a CRT screen flicker in your peripheral vision, but when you look at it, the flicker disappears? And then there's the effects of retinal after-images, the "CPU" involved...etc, etc.

Anyway, I guess one could approximate an answer to the original question by calculating a pixel density for the sharp central 1.5 or 2 degrees of vision, based on pixels that are half an arc-minute in size. It might actually be more practical to convert an image of known pixel resolution, size, and distance from the eye into arc-minute resolution to get a comparison. That might even be useful for deciding how big of an HDTV you "need" for your family room! Either way, I'm too lazy to do the math this morning, though... ;-)

[Jesse Bekas]

<<<- While it is quite sensitive to low light (basically the eye can count single photons) the aperture is rediculously small, making it quite inefficient at light gathering. The lenses are pretty low quality.-->>>

I rather retain the depth of field that our eyes possess, than have better night vision.

[Carl Merritt]

No images pixel count will ever convey more information than an eye can perceive - not to mention it's an analog system - so I would say the human eye has infinite pixels.

[Graeme Nattress]

Analgoue != infinite

Does analogue video tape have infinte resolution - no, it's noise and bandwidth limited, as is analogue audio tape. Noise is very equivalent to digital bit depth, and bandwidth to resolution.

Also, the eye is made op of discrete light sensing cells, and if you count them, you'll get the resolution of the eye, although, as has been pointed out, the software that interperets these pixels is phenominal at perceiving more detail than there is sensed at the eyeball by integrating information over movement and time.

Graeme