Theory vs practice

[Bollygum]

First I should introduce myself. I'm Steve, a photographer (still and timelapse) from Australia and I specialise in fungi photography (what you could call a very small pond). I have only been photographing seriously for the last dozen or so years and I have always used digital. I have now had my photographs printed throughout the world in magazines like Nat Geo and had time lapse shown on programs like Planet Earth II.
I recently joined in another thread here and had in interesting conversation with a couple of your regulars (Kodiak and sk66). The subject was macro, and the interesting part (for me) was the difference between theory and practice. The theory said that with a full frame camera, that using f16 would give no better resolution than 12 or 14MP camera, yet in practice this is clearly wrong. Has anyone else stumbled across areas where there is such a stark difference between theory and practice, and why do you think this is so?

 

[Kodiak Qc]

•

Hi Steve,

In theory, theory and practice are the same. In practice, however, they are not!

I read that somewhere… a long time ago!

 

[Bollygum]

Sums it up. It just surprised me how wrong the diffraction theory was. Surely others have noticed this but nobody has bothered to adjust the theory. Odd! Clearly the manufacturers understand what is happening, else why make a lens for modern ff cameras that goes to f22?

 

[Kodiak Qc]

just surprised me how wrong the diffraction theory was

The theory is not wrong, for sure, but — in practice — new
sensors with higher MS are reacting differently and offering
other possibilities.

bollygum.87017… this may help:

View: https://www.youtube.com/watch?v=u1rINl-NXCs
​

 

[Brazo]

Which lens(s) are you using? A modern lens may perform superbly at f22 but ‘far’ better at f5.6.

Have you looked at any Imatest charts and have you compared real world results at different apertures? Theory and practice there!

 

[chris malcolm]

Sums it up. It just surprised me how wrong the diffraction theory was. Surely others have noticed this but nobody has bothered to adjust the theory. Odd! Clearly the manufacturers understand what is happening, else why make a lens for modern ff cameras that goes to f22?

People misunderstand rules of thumb and simplifications for theory. Diffraction softening of an image is present at every aperture, including the widest. It just increases as you stop the lens down. The diffraction occurs as light passes the edge of the aperture. The image is formed from the non-edge rays of light passing through the lens. Since the size of the edge of the aperture varies with the diameter of the aperture, and amount of the non-edge rays varies with the square of the diameter, the proportion of diffracted rays to non-diffracted rays increases as the size of the aperture is reduced.

So theoretically a perfect lens would be sharpest with least diffraction softening at its widest aperture. In practice none of us can afford a perfect lens, if indeed anyone knows how to make one. Our imperfect lenses suffer from a variety of imperfections which lessen as the lens is stopped down. With wide open apertures the imperfections cause worse image degradation than the amount of diffraction softening. As the lens is stopped down the lessening lens imperfections improve the image more than the increasing diffraction softening worsens it. Until we reach the optimum aperture beyond which diffraction softening worsens the image more than the decreasing imperfections lessen it.

People often speak of that aperture as the aperture at which diffraction softening starts. It's not. It's the aperture at which diffraction softening starts obviously worsening the image. It's specific to that particular lens. There is a tendency -- but only a tendency, not a rule -- for better lenses to stop masking diffraction softening with improving imperfections at wider apertures. I have at least one lens whose optimum aperture is f4. I have another where its either f8 or f16, depending on whether you want best central sharpness or best edge sharpness.

So why make a lens which stops dpwn to f22? Because depth of filed increases as the lens is stopped down, and sometimes you want more depth of field, more spread of sharpness, even though that reduces the sharpness of the best focused parts of the subject.

 

[Bollygum]

People misunderstand rules of thumb and simplifications for theory. Diffraction softening of an image is present at every aperture, including the widest. It just increases as you stop the lens down. The diffraction occurs as light passes the edge of the aperture. The image is formed from the non-edge rays of light passing through the lens. Since the size of the edge of the aperture varies with the diameter of the aperture, and amount of the non-edge rays varies with the square of the diameter, the proportion of diffracted rays to non-diffracted rays increases as the size of the aperture is reduced.

So theoretically a perfect lens would be sharpest with least diffraction softening at its widest aperture. In practice none of us can afford a perfect lens, if indeed anyone knows how to make one. Our imperfect lenses suffer from a variety of imperfections which lessen as the lens is stopped down. With wide open apertures the imperfections cause worse image degradation than the amount of diffraction softening. As the lens is stopped down the lessening lens imperfections improve the image more than the increasing diffraction softening worsens it. Until we reach the optimum aperture beyond which diffraction softening worsens the image more than the decreasing imperfections lessen it.

People often speak of that aperture as the aperture at which diffraction softening starts. It's not. It's the aperture at which diffraction softening starts obviously worsening the image. It's specific to that particular lens. There is a tendency -- but only a tendency, not a rule -- for better lenses to stop masking diffraction softening with improving imperfections at wider apertures. I have at least one lens whose optimum aperture is f4. I have another where its either f8 or f16, depending on whether you want best central sharpness or best edge sharpness.

So why make a lens which stops dpwn to f22? Because depth of filed increases as the lens is stopped down, and sometimes you want more depth of field, more spread of sharpness, even though that reduces the sharpness of the best focused parts of the subject.

Makes sense. I know that diffraction does effect image quality and I probably push those limits more than most, but I work in work in forests with variable light and time constraints, so using f16 is much better than using something like f8 for focus stacks - and the difference is barely noticeable, even at 100% magnification. I rarely get a failure with a stack and I can get a lot of photos in a day. I think 150 stacks in a day in a forest would be the most I have done and that would be impossible using f8. Many lens reviews will specify real world results, but most theory articles imply something far worse, without any qualifiers. I get the impression that a lot of people who write these articles have never actually tried this in practice. If they did they wouldn't write such rubbish. Theory is useless if it never applies to the real world.

Thank you for this response as it does explain things quite well.

 

[Bollygum]

Which lens(s) are you using? A modern lens may perform superbly at f22 but ‘far’ better at f5.6.

Have you looked at any Imatest charts and have you compared real world results at different apertures? Theory and practice there!

A Sony 90mm f2.8 G series lens. I have looked at some good reviews on lenses, but theory is usually a good reference too - just not in this case. Ok, it's complicated to include all the relevant factors, but what they (as in them and us) do is just lazy.

 

[Phil V]

@Bollygum
The 'data' is simple to show, using test charts you can see the lpmm which is measurable and repeatable and bears out the 'theory' which is something that a 'real world' image cant as the subject matter introduces too many variables and subjectivity.

So whilst your 'real world' examples in your view are proving something, I'd say scientifically they prove nothing other than you're happy enough with your images (which is important rather than flippant) Too many photographers spend too much time on trying to achieve 'technically' good images - it's only of use to them.

What ever happened to the term 'measurebators'? Did it fall out of favour when photography forums filled with people with no interest in actual creativity?

 

[Bollygum]

@Bollygum
The 'data' is simple to show, using test charts you can see the lpmm which is measurable and repeatable and bears out the 'theory' which is something that a 'real world' image cant as the subject matter introduces too many variables and subjectivity.

So whilst your 'real world' examples in your view are proving something, I'd say scientifically they prove nothing other than you're happy enough with your images (which is important rather than flippant) Too many photographers spend too much time on trying to achieve 'technically' good images - it's only of use to them.

What ever happened to the term 'measurebators'? Did it fall out of favour when photography forums filled with people with no interest in actual creativity?

. I had worked out long ago what worked in practice, but I could never understand why few photography forum people took much notice when I said use f16 for focus stacks. After all, I do it all the time and I am known for high quality images. I now understand the reasons a little better and can explain why to interested photographers, which sometimes I am asked to do. But - I did find it frustrating not having theory align with reality. Now I am happy that I understand.