Incense

by Bart Arondson

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

There is obviously a limit to what the human body can process, such as frames per second. My question is how many MegaPixels would it take before the human eye can no longer differentiate it from life?

Bonus for including an answer for other species.

share|improve this question
1  
+1 definitely an interesting question. –  Regmi May 22 '13 at 5:22
3  
The 'limits' that you think are obvious aren't even remotely obvious. For one, our eyes and brains do not have frames per second. Secondly, we ultimately see with our brains, not with our eyes, further complicating the issue. –  whatsisname May 22 '13 at 15:56
    
Sometimes you have to provide an analogy to a known quantifiable source to fully understand the question. –  Spechal May 24 '13 at 8:31
1  
@whatsisname We see with our brains? Surely it boils down to OP wanting to know if he hacked a lens onto an eye and the eye to a computer, what the upper limit of the info is that he could get. –  Good Gravy May 27 '13 at 0:45
1  
There are quirks of the human brain that make this question tricky like false colour and things. –  Good Gravy May 27 '13 at 0:46

10 Answers 10

Question about thing like frame rate, resolution or dynamic range of the human eye and how they compare to cameras always have the same problems:

  1. The "picture" you see isn't a "single exposure", the eye is constantly moving and adjusting.

  2. The part of tee brain that handles vision is really good (and pretty big), it constantly combines the "frames" is gets from the eye and fill in the blanks.

Basically, every image you see with your eyes is an HDR panorama that was fixed with content aware fill (and just like with a camera, when you get into HDR panoramas you can make them at any arbitrarily high resolution and DR)

Also, the eye/brain actually only works on the part of the scene you are concentrating on, you get amazingly high resolution for the tiny part of the world you are thinking about right now - for the rest of the scene you don't really "see" it at all, you only really have to notice things if there's anything dangerous heading your way (that is why movement at the sides is so distracting).

share|improve this answer
    
That's basically why what we 'see' is in a much higher resolution than the physical properties of our eyes allow us to really register. The sharpest part is also exactly in the center of our field of view, closer to the edges resolution quickly deteriorates. –  Emiel May 21 '13 at 10:47
2  
@Emiel - not exactly the center - around the exact center, in exactly the center we have a conveniently located blind spot –  Nir May 21 '13 at 12:34
2  
Neither the fovea nor the blind spot is exactly on the optical axis, but the fovea is closer than the blind spot. –  coneslayer May 27 '13 at 0:42
    
Another thing to note: the eyes are a rather loose thing, if you take a look at an eye moving in slow motion, you can notice it jittering around. Using that jitter our brain interpolates the images it gets to increase the resolution seen. –  SinisterMJ Jun 13 '13 at 15:16

If you look at the specifications of the human eye as if it's a camera, you're going to find it's pretty low-specced.

  • Very low resolution in terms of pixels - very few megapixels - with most pixels concentrated in a very small area in the centre. Virtually no ability to distinguish fine detail outside of a small area in the centre of the frame.

  • Horrible extreme chromatic aberration, spherical aberration and noise.

  • Minimum and maximum focusing distance deteriorates with age, and a lot of models have defects from the factory.

However, the reason none of this counts is that measuring the eye as if it's a camera doesn't make sense: The image we see is created by our brain, which flawlessly and continuously stitches together countless images taken by our eyes and processes them.

Whereas the eye only has a very small area in the centre of our vision that has any real ability to discern detail, the brain has a motor mechanism that swivels the eye around in order to take hundreds of quick samples of the image one after the other, then assembles this into one big picture (with three dimensions, and motion!).

You would need hundreds of megapixels of resolution and a virtually flawless lens to replicate the composite image that the brain assembles, even though the eye in isolation is nowhere near capable of anything this good.

share|improve this answer
    
Where did you get <1MP from? –  Good Gravy May 28 '13 at 19:02
    
Just a guesstimate - I was mainly just trying to illustrate a point, that our eye really can't see much detail at all except for in a tiny area in the centre of its vision. –  thomasrutter May 29 '13 at 1:24
2  
Our eye has about 6-7 Millions cones in its eyes, so that would be the absolute maximum resolution. Since the outer regions are rather sparse (see hyperphysics.phy-astr.gsu.edu/hbase/vision/rodcone.html), I think a good guess would be around 6 million cones -> mega pixels. The pixels on the camera sensor are masked by a bayer filter just like our eyes' cones are only resceptible to certain colors. I feel the <1M is too little. –  SinisterMJ Jun 13 '13 at 15:19
    
flawlessly might be too strong. It seems flawless because the brain also gets to tell the story. –  William Shakespeare Nov 11 at 20:10

How many "pixels" the human eye captures does not really answer the question. It only equates when, say, the picture you've taken with a camera is blown up to be big enough to consume the viewer's entire visual field. At that size, the original photo would have needed to be approximately 576 Mp.

Detail for a picture is usually measured in DPI (dots per inch), and, even then, the size and distance from the viewer have to be fixed in order to determine how dense the dots have to be for the human eye to no longer be able to tell they are dots.

High quality print made for the average reading distance (18-24 in.) is on the order of 5-10K DPI. For a 1 inch square picture (@10K) that is 100 Mp right there ... for a 1x1 inch picture.

The problem is that even though a general scene may only need 576 Mp, when the eye actually focuses on a specific region, all of its acuity comes to bear on that region. Thus a 1x1 inch picture needs to be much higher density to "fool" the eye.

To make a picture big enough, and yet detailed enough to be focused in on, well, the number of MegaPixels is huge. That's why you see glasses being used. The screen is much closer to the eye, thus making the picture denser and yet appear larger.

Say you have a 5 MP camera. That's roughly 2,200 x 2,200 pixels. If the sensor (CCD) is roughly 1 in x 1 in, that is ... you guessed it 2,200 DPI.

Now blow that up to an 8 in x 8 in photo, and it is only 275 DPI. Nowhere near the 5000 DPI you need for a high quality print. (however, if you look at it from 8 times as far away ... )

To be honest, 2K DPI is passable for a standard print (@ reading distance), and when viewing a photo on a small screen (or print) it looks much more "real".

To get a 4x5 @ 5K DPI you'd need 500 Mp. @ 2K you still would need 80 Mp. Roughly speaking, a 24 Mp (CCD) camera is equivalent to 35mm film quality.

Of course, there are a lot of enhancement techniques you can use to "fill in" the missing density when you have a digital image.

But if you need big pictures, old fashion film can be made in much larger sizes than CCDs can ( 8in X 10in film for example: http://answers.yahoo.com/question )/index?qid=20061123192628AANDiGx )

share|improve this answer
5  
Source on that 576 mp number? –  Fake Name May 22 '13 at 8:25
    
I really have to QUASH the use of DPI in situations such as this. DPI stands for DOTS per inch. A Dot is an arbitrary measure...and it means different things in different contexts. The only appropriate term is PPI, pixels per inch, as that is the smallest complete element of the PICTURE that as a photographer you are capable of working with. –  jrista May 22 '13 at 21:24
    
I also need to ask that the postulations in this answer be referenced. I know where the 576mp number comes from, however it is an unreferenced number for any casual reader. Some of your numbers are just flat out wrong, unless you are using mixed meanings for the term "DPI" (which I've already noted is a terrible term to use in this context.) The FACT here is that as viewing distance INCREASES, minimum required spatial print resolution DECREASES. Assuming slightly higher than average visual acuity (say 20/18), a print at 18" would only need 230PPI. Assuming someone with high visual acuity... –  jrista May 22 '13 at 22:03
    
...such as 20/10, one would need 280ppi. At 24", that number drops to 210ppi! (In the case of most ink jet printers, DPI is a fixed factor...default dots per inch in a Canon is always 4800x2400; in an Epson it is 5760x1440...regardless of the actual printed image resolution.) So, I'm sorry, but your facts are skewed, confused, and in some cases just wrong. Please see my answer here. –  jrista May 22 '13 at 22:05

The number 576MP, which is derived at Roger Clark's site here, is an EXTREMELY ROUGH APPROXIMATION. For one, it is a conservative estimate given a 120º FOV, when human vision is closer to 180º (which actually clocks in at 1.3 GIGAPIXELS!!!) It also ignores the fact that we have a 2º "foveal spot" near the center of our eyes where our acuity is highest, and a wider 10º region where our vision is decent, but not really "good" and certainly not excellent (as a quick test...see how much of the text in this answer is actually fully clear, and how much is actually indistinct and unreadable when looking at exactly the same spot for a period of time...you might be surprised at how much of your screen you cannot actually analyze in any real meaningful detail.) At the periphery of our vision, acuity is fairly low, lacks color fidelity, etc.

In my opinion, I don't believe it is valid to even describe human vision in terms of megapixels. I have great respect for Roger Clark, however his article needs to be taken in the right light: It assumes maximum visual acuity across the entire field of vision! The critical fact here is that our maximum visual acuity only affects a small region of the central portion of our vision. A region that probably doesn't even cover a single 8x10" print viewed a foot away...which requires less than 9 megapixels (3330x2664 pixels) to print at 333ppi (the required resolution for a one-foot viewing distance.)

One would theoretically require fewer and fewer megapixels to continue printing rings of 8x10" prints circling a central one to fill the entire human field of view. From a real-world acuity standpoint, one probably needs 1/3rd fewer megapixels per "ring" of prints (rough guess), and maybe four rings of prints to entirely fill the field of view "corner to corner". That clocks in at less than 85 megapixels all told!

That said...I still do not believe it is accurate or useful to try and describe human visual acuity in terms of megapixels. We have varying acuity from the center to the edge of our visual field of view, with rapid falloff outside of perhaps a 4-5º central high acuity region.

share|improve this answer

Overview.

A very difficult, but interesting question. There is one key thing before we get started. The brain instantly deletes unneeded info amongst other super intensive processing and focuses on stuff thats worth remembering. What you 'see' is not accurate of the eye's technical ability. But as for its technical ability; there are a range of estimates, from 5 to over 500 megapixels.

Note: None of these calculations are scientifically accepted.

Human Eyes.

A human with 20/20 vision, is able to resolve the equivalent of about 52 megapixel camera (assuming a 60° angle of view). This is based on each rod and cone cell being able to represent a megapixel. There are about 7 million cones (need high light levels and provide colour) and 120 million rods (work in low light, don't output colour, not always activated). Together these work to create somewhere between 50-500MP. (REALLY appoximately!). Less conservative estimates claim 500+ million megapixels.

None of these articles have been peer reviewed, so there is no scientific viability to any of these ideas. The 567MP estimate does not assume a still image. It takes into account tiny angular vibrations that the eyes do to gather more information. The estimate also takes into account a wider field of view (120˚)(hence it has more MP than there are photoreceptors).

This article disputes those high estimates and says "such calculations are misleading". Amongst things like low light and there not being a shutter speed, the most notable difference in a picture vs your eyesight is derived from the way your eyes focus on something.

Only the central vision is 20/20. The overall image is pretty pants away from the center. Just 20° off-center our eyes resolve only one-tenth as much detail. At the periphery, we only detect large-scale contrast and minimal color. Based on this a single glance of the eyes is therefore only capable of perceiving detail comparable to a 5-15 megapixel camera (depending on one's eyesight). The eye needs to therefor glance several times and even then only the memorable textures, colours and shapes will be remembered.

Other animals.

Hawk. This is probably what people will be most familiar with as an eagle eyed bird of prey. They have around 5 times greater density of photoreceptors than we do, so lets say they have a quarter of a gigapixel (250 MP-5.5GP). What's better about these guys than us is that they have way more nerves going to the brain than we do. There is no sure way of saying that indicates a better resolution, but it indicates more information is being relayed to the brain from their eyes.

http://en.wikipedia.org/wiki/Hawk#Eyesight

Mantis Shrimp. We have 3 types of colour photoreceptors (cone cells). Scientists have identified 16 colour receptors in mantis shrimp. Obviously this is beyond our minds comprehension, but a quick trip to Reddit's S***ty HDR find the most over the top HDR and then overlay the image with rainbows, explosions and dreams and you might get somewhere close to what the crustaceans see. Obviously this has nothing to do with resolution, but the colour depth those guys have is awesome.

share|improve this answer

From what I read, I believe when discussing the ultimate resolving power of the eye, you need to consider that fovea is the only part of the retina able to distinguish the fine detail. The size of this region on our retina is quite small requiring us to constantly adjust our eyes to allow the “subject” to fall on this area. In fact it is so small that even when concentrating on a small object, we must scan over it, we can’t resolve the details of even a small object at one time. How large an area can we resolve with maximum clarity without darting our eyes about? That area has a diameter of about the distance between the two dots of a colon read at a normal reading distance.

Regarding frames per second, I believe the equivalence for humans is 1/10 of a second. Try an experiment – while stopped at a light, notice how the details of alloy wheels on the cars crossing your path are a blur. While following one with your eyes, tap (not whack) the side of your head at the temple. This will jar your eyes and sometimes, for the briefest moment, your eyes will “pan” with a portion of the wheel which will reveal its details.

share|improve this answer

Simple answer to this question would be 2 megapixels. I mean it. Here is a scientific explanation to that MindLabs.

Human eye does not see well at all. When focusing close we are really selective that it may equal to f1. The 99% of the scene is too blurred.

We also have a blind spot which is explained in the link above.

Also we can't freeze any scene, which can't be comparable to even a cheapest camera.

To sum it up, our eye suck, but our brains compensate too well that we all believe we are better that every camera out n the market.

share|improve this answer
1  
Or to put the summary a different way: we are better than every camera out in the market, because of our amazing post-processing. –  mattdm May 26 '13 at 14:02

You should probably not ask about megapixels, human eye is a complex system, not only the "matrix". You should better ask about range of angular resolutions.

Look for it here:

http://en.wikipedia.org/wiki/Naked_eye

http://en.wikipedia.org/wiki/Angular_resolution

Angular resolution: about 4 arcminutes, or approximately 0.07°,[1] which corresponds to 1.2 m at a 1 km distance.

share|improve this answer
    
+1 for pointing out angular resolution –  fortran Aug 30 '13 at 8:55

576 megapixels - that is according to an article by scientist and photographer Roger Clark, which also tells more about human eye and its equivalences to digital technology...

share|improve this answer
1  
The article contains the right answer but please summarize the critical parts in your answer. –  Itai May 21 '13 at 13:06
2  
The question specifically asked "How many MegaPixels". This answer provides a simple, concise answer to that question and cites a reference. The critical part is the number - 576 MP. –  Michael Clark May 21 '13 at 14:47
1  
Yes, but I would like there to be an explanations of why those megapixels do not compare directly to that of a digital image. –  Itai May 22 '13 at 21:00

There are about 120 million rods and about 6 million cones, so the maximum theoretical human eye resolution (considering perfect optical light transmission in the retina) should be around the figure of 2 megapixels (it takes 3 cones for an RGB triplet) with a really high dynamic range in the peripheral areas (that's what the rods are for).

share|improve this answer
    
Are you suggesting that the rods don't contribute to our resolving power at all? –  mattdm Aug 30 '13 at 11:51
    
And for that matter, that our brains don't do interpolation/integration similar to Bayer demosaicing? –  mattdm Aug 30 '13 at 11:52
    
Well, take a look on how are they distributed and what they are for (in the periphery of the eye, monochrome night vision)... I wouldn't dare to say they contribute much for things that require a precise sight like reading (but they are of course useful for seeing predators approaching!). And the real spatial resolution for a demosaiced Bayern pattern is less than the MPx reported (ask the Foveon guys :p). –  fortran Aug 30 '13 at 12:28

protected by mattdm Jun 13 '13 at 10:49

Thank you for your interest in this question. Because it has attracted low-quality answers, posting an answer now requires 10 reputation on this site.

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.