Diffraction limited pixels... Really?

We hear way too much about NOT using f/16 or f/22 because of diffraction limits due to our sensors pixel size.

Sure, f/8 is a great general plan, do it when you can. Nevertheless, there are obviously many times when f/16 or f/22, and maybe f/32, can heroically and drastically help our depth of field issues. Especially macro work, but any time greater depth of field is needed.

It is a compromise, but this is true any time that greater depth of field helps more than the greater diffraction hurts.

If your notions have been formed to simply always avoid f/16, you're probably missing out on a very good thing.

So maybe read here: Diffraction limited images?

There's times trying to get more DOF is a good idea but there's times it is an exercise in futility. At least that's how I experience it in macro.

Earlier today I was working at my five and three quarter-legged spider shot when I noticed I used f/8. I wondered if it had made much difference if I had closed down. It's shot at about 1:1 which implies I have a DOF of around 0.5mm at f/8. At f/32 it would be around 2.0mm. While that's 4 times as much, in reality it is less impressive since it's but a difference of 7.5% depth assuming the spider to be around 20mm lengthwise.

Would that 7.5% have made a better shot? Maybe it would but maybe not. Is that 7.5% worth closing down the little light there was? Certainly not.

Now admittedly when the subject-sensor size is different, stopping down might be the best approach. The smaller, the larger the effect of the same DOF. When decreasing the X-factor the same is true but I prefer to do that only when really needed. I didn't spend all this money to make a 1:1 do what a 100$ 1:2 lens can do too.

About diffraction I care less. I experience my image breaking down when I up the ISO while closing down the lens. We're talking 800+ here. Then the image becomes fuzzy. The cause could be me too. Whatever it is, that range no longer works for me.

Yes I know, flash and such but me -angry smilie- flash.

J-see said:

Earlier today I was working at my five and three quarter-legged spider shot when I noticed I used f/8. I wondered if it had made much difference if I had closed down. It's shot at about 1:1 which implies I have a DOF of around 0.5mm at f/8. At f/32 it would be around 2.0mm. While that's 4 times as much, in reality it is less impressive since it's but a difference of 7.5% depth assuming the spider to be around 20mm lengthwise.

Would that 7.5% have made a better shot? Maybe it would but maybe not. Is that 7.5% worth closing down the little light there was? Certainly not.

Click to expand...

Sorry, no sympathy. Photography is about solving the problems, not about complaining about them.

See my link above, the last picture on the page. It uses f/40 (and 4 stops more flash power), but it certainly seems improved. A great picture? Certainly not. But improved? Obviously.

I'm not complaining, I'm ignoring problems.

Sure it works in some cases but in many it is irrelevant. We're worrying too much about numbers.

The way I see it, the cam is but a part in a creative process. It's a piece of plastic, as is the lens these days; tools. What's ironic is that while they're but a part, it's about everything we talk about. All the time we're absorbed in focal length here, a millimeter there, diaphragm, IQ, pixels pixels and pixels. The most important part however is talked about least; the actual creative process. Mind you, I'm in it too.

Would painters have done the same about the size and length of the hair on their brushes, the Sistine Chapel would have been plastered white.

J-see said:

I'm not complaining, I'm ignoring problems.

Sure it works in some cases but in many it is irrelevant. We're worrying too much about numbers.

The way I see it, the cam is but a part in a creative process. It's a piece of plastic, as is the lens these days; tools. What's ironic is that while they're but a part, it's about everything we talk about. All the time we're absorbed in focal length here, a millimeter there, diaphragm, IQ, pixels pixels and pixels. The most important part however is talked about least; the actual creative process. Mind you, I'm in it too.

Would painters have done the same about the size and length of the hair on their brushes, the Sistine Chapel would have been plastered white.

Click to expand...

? My guess is Michelangelo paid considerable attention to his brushes, and to every detail. I doubt he ignored ANY problem in the Sistine Chapel.

WayneF said:

? My guess is Michelangelo paid considerable attention to his brushes, and to every detail. I doubt he ignored ANY problem in the Sistine Chapel.

Click to expand...

And yet, nobody talks about those brushes.

We talk about our brushes all the time and because of our numbers game, start to value everything in relation to these numbers too.

Suit yourself then. Me? I prefer to try to pay attention to what I'm doing. Wish I could do that better.

It's not different for me.

I surely hope I'll get better but even while I see nothing wrong with knowing the basics or technical side, I'm fully prepared to screw that all if that leads to something reflecting what I have in mind.

Diffraction limited images?


Edit there, added second page of samples, intended to minimize fears of using f/22 when it is needed.

I find diffraction to be more of an issue when focusing on very distant subjects. Shot a waterfall a few years back at f22 and the fall wall of the canyon looked terrible compared to shots at f13.

There are no constants. f/22 is hard on short lenses, but is a breeze for longer lenses. A maximum fstop number of focal length / 4 is suggested. My point is that this subject is really not about sensor pixels.

See the second page there, several samples, FX and DX.

I was reading some about it yesterday:

https://fstoppers.com/studio/fstopp...raction-and-when-does-diffraction-happen-6022


There's a row of images that shows the gradual decline but it's pretty much about details. Macro might be subject to it most but like I said, the decline I myself noticed might have had more to do with ISO.

Interesting topic. Gives me something to think about. Might keep me out of trouble for a few days.

I agree on using the tools at our disposal. The difference between a biped-pod (human holding a camera) and a photographer is the photographer makes choices about the pros and cons of a variety of settings and influences on the final image.

WayneF said:

There are no constants. f/22 is hard on short lenses, but is a breeze for longer lenses. A maximum fstop number of focal length / 4 is suggested. My point is that this subject is really not about sensor pixels.

See the second page there, several samples, FX and DX.

Click to expand...

Too true - diffraction is an optical property and has little or nothing to do with the pixels. However, since the advent of digital cameras, many more people are concerned with diffraction mainly due to pixel-peeping. Back in the days of film, people rarely enlarged images to the degree that a 100% pixel-peep represents. As a result, diffraction effects were harder to see, and there was less concern with them.

As far as lens focal length - it doesn't figure into the diffraction calculation - diffraction depends only on f-value and wavelength of light. i.e. a 20mm lens at f/22 has the same sized airy disk as a 300mm lens at f/22.

Airy disk - Wikipedia, the free encyclopedia

J-see said:

There's a row of images that shows the gradual decline but it's pretty much about details. Macro might be subject to it most but like I said, the decline I myself noticed might have had more to do with ISO.

Click to expand...

? All images involve depth of field. But f/22 is a bigger story than that. Yes, it is pretty much about resolution. It is also pretty much about depth of field. And pretty much about focal length.

fstop number = focal length / aperture diameter.

f/22 on a 20mm lens has an aperture diameter of 20/22 = 0.9 mm. That is a pretty tiny hole, which causes trouble. f/4 is sharper.
f/22 on a 60mm lens has an aperture diameter of 60/22 = 2.7 mm. This is borderline small, but certainly bearable when it helps DOF.
f/22 on a 105mm lens has an aperture diameter of 105/22 = 4.8 mm, much more reasonable.

f/22 might not always be a good plan for a short lens, but certainly a longer lens can usually benefit greatly. It's one advantage of the 105 and 200 mm macro lenses. But image SIZE is also a factor, and it can be the same.

But f/22 is not necessarily always bad for short lenses. Let's consider 20mm and 100 mm, for round numbers.

If you can think of your subject object image size as the count of so many line pairs, the subject object at 20mm covers 1/5 the size at 5x the resolution, and the subject object at 100mm covers 5x the size at 1/5 the resolution. This is the same number of line pairs (same detail) over the size extent of the subject object, and the detail in the objects can appear similar to us.

So standing up up closer with the short lens, to magnify the subject image, will help f/22 at 20mm. It is not necessarily bad. But some situations are.

dslater said:

As far as lens focal length - it doesn't figure into the diffraction calculation - diffraction depends only on f-value and wavelength of light. i.e. a 20mm lens at f/22 has the same sized airy disk as a 300mm lens at f/22.

Airy disk - Wikipedia, the free encyclopedia

Click to expand...

Read the third equation down. Focal length magnifies it. The resolution factor depends directly on fstop number.

WayneF said:

?


Read the third equation down. Focal length magnifies it. The resolution factor depends directly on fstop number.

Click to expand...

Read the same equation and the sentence that immediately follows it - increasing the diameter of the lens reduces the airy disk - the result is that the airy disk size depends only on F-value (i.e. f/d in the equation) & wavelength of light used.

dslater said:

Read the same equation and the sentence that immediately follows it - increasing the diameter of the lens reduces the airy disk - the result is that the airy disk size depends only on F-value (i.e. f/d in the equation) & wavelength of light used.

Click to expand...

Actually, third says minimum resolvable separation of two Airy disks (resolution) depends on fstop number.
1/minimum separation resolved is resolution. It is a very big major concept.

OK, but wavelength is not always easy for us to control. For the given wavelength we discover at the scene, the theoretical maximum for resolving separation depends directly on fstop number. Resolution is reciprocal, inversely.

Airy did this in 1834... so it has always been known. It is NOT new information. He just did not know then that focal length / diameter was going to be called fstop.

In one of the reviews of the Zeiss 135mm APO, the author stated that diffraction magnifies the CA. At very fine detail the CA will make the image fuzzy at say f11. With APO, where the CA is less than the pixel size, the lense could be used at f11 or even f16 without showing any fuzziness due to diffraction.

So I have to use higher F stops for say macros, I would search for a good APO lense. That is why some of the industrial lenses (which are true APO) cost so much. They are corrected for geometric as well as chromatic aberrations, much below the pixel pitch of the current sensors.

aroy said:

In one of the reviews of the Zeiss 135mm APO, the author stated that diffraction magnifies the CA. At very fine detail the CA will make the image fuzzy at say f11. With APO, where the CA is less than the pixel size, the lense could be used at f11 or even f16 without showing any fuzziness due to diffraction.

So I have to use higher F stops for say macros, I would search for a good APO lense. That is why some of the industrial lenses (which are true APO) cost so much. They are corrected for geometric as well as chromatic aberrations, much below the pixel pitch of the current sensors.

Click to expand...

Diffraction is simply NOT about pixel size. It occurred the same in film lenses. The sensor pixels merely try to reproduce the analog lens image, the best they can. The diffraction has already occurred.

Larger pixels, which is less digital resolution, may show diffraction less, but it also shows all detail less. It does not make the diffraction go away.

WayneF said:

Diffraction is simply NOT about pixel size. It occurred the same in film lenses. The sensor pixels merely try to reproduce the analog lens image, the best they can. The diffraction has already occurred.

Larger pixels, which is less digital resolution, may show diffraction less, but it also shows all detail less. It does not make the diffraction go away.

Click to expand...

I think I was not able to explain properly. What was said is that when there is CA, it adds to the edge fuzziness so that the effect of diffraction is visible as a softer image at f stops much lower than if there was no CA. I have verified this with comparison of my 70-300 AF at 300mm and the kit lense at 55mm. Even when the image at 300mm is downsized the CA makes the image much softer than that taken at 55mm with kit lense (It was a bird and both images were made the same size)