Rubén's Article on Sensor Resolution

[wparsonsgisnet]

Advertisement (gone after registration)

There is an article by our very own Rubén Osana and a collaborator on Luminous Lanscape.

 

Do Sensors “Outresolve” Lenses?

[ho_co]

And it's excellently done. He and Efraín García have gathered data from many sources and succeeded in pushing many old film-oriented ideas out of the way.

 

Clearly written and explained, derived from sources such as Puts, Myhrvold, Zeiss et al, and referencing Atkins, Koren, van Walree etc.

 

Superb compilation at a very readable level.

 

Thanks, Bill, for pointing it out; and thanks, Rubén, for the work!

 

BTW--The fancy quotes confuse the forum's parser. The article title should look more like:

 

Do Sensors "Outresolve" Lenses?

[rosuna]

Thank you for the post and comments!

[lars_bergquist]

The conclusions of the article are largely based on an error of assumptions. This is that 'typical print sizes' are relevant.

 

This is true as long as you assume a viewing distance of 25–30 cm, and naked eye viewing. The first part of this does hold but only for small print sizes. The limiting factor is the angle we can take in while still considering the picture as an image. This angle is about equal to the diagonal of the picture. Small pictures (diagonal =/< 30 cm) are all viewed at minimum focus range of the eye. This means that small prints are more forgiving than larger ones. Beyond that, you start increasing the viewing distance. It is easy to see that anywhere where pictures are displayed hung on walls, like an art museum. People tend to stand at a distance equal to the diagonal of the picture, if barriers, furniture or other people do not restrict them.

 

This means that in general photography, A4 or similar sizes are the test. If these look sharp to the eye, they are sharp, and more edge contrast or more detail will not make them look any sharper.

 

Now you can of course choose to regard the print as a technical or scientific specimen and not as a picture. Then there is no limit. You can make a 2x3 meter print and view it piece by piece through a magnifier, or why not a microscope. AND you could do that already in 1928, or 1938, or 1948 ... In this case you come up against what the microscopists (who are indeed viewing technical or scientific specimens) call 'empty magnification', i.e. magnification which does not add further detail. But microscopy and general photography are different matters, and applying the criteria of one to the other is meaningless.

 

Sharpness is the mental hobgoblin of failed photographers.

 

The old man from the Age of the Darkroom

[Telecentric]

The conclusions of the article are largely based on an error of assumptions. This is that 'typical print sizes' are relevant.

 

...

 

Sharpness is the mental hobgoblin of failed photographers.

 

The old man from the Age of the Darkroom

 

+1 on this

 

Take a look at an excellent exposition on this at DSLR Resolution and Noise Comparison

 

To quote from this article "Resolution and noise have almost nothing to do with real image quality, unless you look at three-foot (1m) wide images from a few inches away."

[lars_bergquist]

+1 on this

 

Take a look at an excellent exposition on this at DSLR Resolution and Noise Comparison

 

To quote from this article "Resolution and noise have almost nothing to do with real image quality, unless you look at three-foot (1m) wide images from a few inches away."

Indeed, though that argument was about noise, not definition. My central point is that we subconsciously try to adjust our viewing distance to correspond as closely as we can to the diagonal of the picture. The close focusing limit of the young and healthy human eye is about 25–30 cm (10–12"). So when looking at pictures with a smaller diagonal, we must be content with holding them as close as we can, i.e. at the eye's close focus. Reading glasses for us old people are chosen to give us this reading distance. Type in octavo format books is chosen accordingly (7–10 points size). And good modern gear designers do understand the 'diagonal rule'. I am writing this on a laptop with a 15" screen. And what is the reading distance that the layout of the keyboard and the rest gives me? Your guess is correct ...

 

Magnifiers, microscopes and the rest are for allowing closer looks. A 5x magnifier reduces your 'reading distance' from 25 to 5 cm. Now you can check on the reproduction of single characters on the page. But – that does not make it easier to read the text! And 'reading the text', reading the subject, is what general photography is for.

 

Now, even if we are concerned only with the technical aspects of the image, there is much else than pixels to think about. Things like image geometry (i.e. distortion), vignetting, tonality, colour rendition ... many of these things can to some extent be manipulated in post-processing. But I suspect that there is a line, though perhaps not a totally sharp one, between photography and PhotoShopography.

 

And as long as technical concerns do not prevent us from successfully 'reading' the picture, what makes the picture interesting *as a picture* and not as a specimen, are things like the subject and how the photographer does visually relate to it. Because this is what helps US to relate visually to the subject.

 

My final remark can be translated as 'if you can't make an interesting picture, you can at least brag about how 'sharp' it is'.

 

The old man from the Age of the Darkroom

[sean_reid]

Advertisement (gone after registration)

I want to congratulate Rubén and Efraín García for a thoughtful and rigorous piece. In fact, I just logged into the forum specifically to post a thread about this new article - but then found one in place. Like so many aspects of photography, one can debate the assumptions and conclusions in any article (as Lars is doing in his posts above) but Rubén and Efraín made an effort to write something that was both rigorous and literate.

 

The theory they discuss in this article also supports what many of us have noticed in the course of doing normal photography. And that is that the resolution of some digital cameras can be lens-limited. I would include the M8 in this group. That camera is certainly able to show us differences in resolution among various RF lenses. The article also emphasizes the role diffraction plays in resolution and that's important to be aware of as well.

 

I think photographers who are interested in this subject should certainly read this article and give it some thought. Ruben, as many may know, is an economist and professor and so its natural that he tries to approach this subject with rigor, supporting references, etc. That's not to say that everyone will agree with all aspects of the article but its great food for thought.

 

Lastly, I don't see scientific/theoretical discussions of technical topics in photography as something to be viewed in opposition to the understanding of the medium as a visual art. They simply present one aspect of photography that may be of interest. But it doesn't follow that such discussions are proposing that the technical aspects of photography *are* photography. It's somewhat a case of giving to Caesar what is Caesar's.

 

Cheers,

 

Sean

[Guest jimmy pro]

Lastly, I don't see scientific/theoretical discussions of technical topics in photography as something to be viewed in opposition to the understanding of the medium as a visual art. They simply present one aspect of photography that may be of interest. But it doesn't follow that such discussions are proposing that the technical aspects of photography *are* photography.

 

You'd think that would be obvious, and most everybody probably knows it's true, deep down. But you can also understand why people trying to support there own emotionally oriented opinion, such as that film (including film scanned to digital) is still "superior" to digital capture, would like to censor scientific discussion, since they can't disproove it.

[jaapv]

+1 on this

 

Take a look at an excellent exposition on this at DSLR Resolution and Noise Comparison

 

To quote from this article "Resolution and noise have almost nothing to do with real image quality, unless you look at three-foot (1m) wide images from a few inches away."

 

Hmm, I think you did not read the article thouroughly enough. That point is made quite early on, reason why it moves on to frequency response and MTF curves, limited as those are too.

 

Resolving power and acutance aren’t good measures of image quality if we take them separate.

 

Anyway, I found it an excellent and concise article which summarizes much of the thinking on the subject.

[Bob Ross]

Thank you Bill and thank you Ruben. That is the second good article on digital imaging that I've read this week. The other one is on digital image noise at

Noise, Dynamic Range and Bit Depth in Digital SLRs

Bob

 

PS: Ruben, since MR who hosted your article is into pocket digicams, you might want to do a tiny pixel addendum, say down to 1.7 µm pixels......

[lars_bergquist]

You'd think that would be obvious, and most everybody probably knows it's true, deep down. But you can also understand why people trying to support there own emotionally oriented opinion, such as that film (including film scanned to digital) is still "superior" to digital capture, would like to censor scientific discussion, since they can't disproove it.

Well, I simply wanted a more realistic perspective, one that takes into consideration why we make photographic pictures. The criteria have to be suited to the purpose.

 

And I am not a hibernating film nostalgic. I have not taken one single picture on film since I received my M8 early last spring. The article was interesting, I just found some of the conclusions unwarranted.

 

In the National Museum in Stockholm hangs Rembrandt's 'Claudius Civilis', a painting that is literally wall-sized. Last time I saw it there was no barrier in front of it. But I have yet to see anybody trying to view it from closer than about four meters, and most seem to prefer five. Why should you stick your nose into it, unless you are a conservator? The size of Rembrandt's brushes is irrelevant.

 

The old man from the Age of Analog

[Guest jimmy pro]

Why should you stick your nose into it, unless you are a conservator?

 

Ha! Reminds me of what a guy I knew said who was a fashion photog in Manhattan, he said a photo of a girl only has to look good from a distance, even if the girl has to look good up close;)

[rosuna]

Thank you to everyone.

 

Lars' comments about optimum viewing distance are correct, but I see photographs from different points, even very close. The optimal distance for a natural angle of view of the photograph (the same AoV from which the photograph was taken) is a criterion, but the optimal viewing distance for maximum detail is a different one. Even more, the evaluation of the maximum resolving power of a camera and lens doesn't assume a particular print size or viewing distance.

 

Bob, you are right. Pocket digicams need much higher sampling frequencies, and much smaller "pixels". It is easier to correct aberrations too, but manufacturers are addict to super-zooms.

 

Sean, thank you very much. I agree these technical subjects are interesting and complete the knowledge of any photographer. Many great photographers were good at a technical level. It reminds me the case of Standley Kubrick, great photographer and director and a great and very cultured "technician". It isn't necessary, but it helps.