About diffraction and its real impact

[Juan Valdenebro]

In past years there was a thread here on RFF I remember because I felt surprised when a forum member stated "no one should photograph beyond f/4 or f/5.6..." (my words to say what he said).
I remember I wrote my opinion, a different one, about f/8 and f/11 being perfectly usable... I talk about 35mm photography.
As I had done some tests about it with Tri-X in Rodinal (big grain) I thought maybe that was the reason I could see no sharpness difference in the focused subject/center of the image...
Then I asked for some help from other forum member using an M9, and he kindly did a test with the CV 28 3.5, and he found no sharpness loss at f/11.
Maybe other lenses (1.4 ones?) behave differently...
Internet and pixel geeks tend to say after f/4 sharpness suffers... I wonder if they refer to photography or to microscope peeping...
I know optic laws, but for real 35mm handheld photography, isn't all this a bit of a myth?
Recently I did a new test: instead of Tri-X and Rodinal I went for a "sharper" film and developer set, and I did it with TMax100 and Microdol-X...
I won't upload images because they're good for nothing: the five of them have the same sharpness: f/4 1/2000, f/5.6 1/1000, f/8 1/500, f/11 1/250 and f/16 1/125. The lens was the CV 35 2.5, M-mount.
Yesterday I read a couple of articles and both referred to f/22 as the first f-stop to care about... That seems more real to me...
I scanned at 4800 dpi, but that doesn't mean a lot, because some scanners are cleaner than others... Anyway, after checking the five negatives with a 22x loupe without perceiving different sharpness, I guess
no difference would be appreciated in 8x10 prints, or even bigger, and that's sharp enough when it's sharp...
In any case f/11 is the smallest f-stop I use, but it was nice to see f/16 was fine too.
I started this thread to share other forum members' opinions, perhaps related to other/faster lenses... It's always good to know...
Cheers,
Juan

 

[View Range]

There are several considerations. Photos are visual media, and the eye/brain interpret the apparent sharpness; depth of field can more than balance for loss of lines of resolution in the exact focal plane. Next, the optical performance of the lens may improve with smaller aperture because you are narrowing the diameter of glass that actually forms the image. Then there is the issue of photographers technique; higher shutter speeds (larger aperture) may reduce the effect of camera movement.

I use any of the f-stops available on the lens, from min to max, depending on the lighting situation and the desired image. Except for very exacting copy situations, I would not worry about diffraction at all.

 

[Roger Hicks]

Dear Juan,

Diffraction limitations to resolution are for the most part trivial next to almost every other consideration. Under ideal conditions, shooting a test target with the camera on a tripod, you may be able to see the difference below f/8. I have myself seen the difference at f/8>f/11, trivial, and f/11>f/16, noticeable, but only under ideal conditions, shooting a test target with the camera on a tripod.

If the picture is any good at all, the importance of diffraction limitations at f/8 is likely to be non-existent; at f/11, trivial; and at f/16, little more than detectable. But there's a good reason why lenses for 24x36 don't stop down to f/22 and f/32, except for applications where DoF is more important than ultimate resolution, e.g. Micro-Nikkor.

Cheers,

R.

 

[peterm1]

In a practical sense you are quite right, diffraction effects do not really have a noticeable effect on an image till the lens aperture in use is quite small. In 35mm photography this usually means at least f16 or even smaller depending on the lens design. However, in a lens for a small format digital camera this can happen at f8 or even less as these lenses are physically smaller than an equivalent lens for a 35mm camera. So you will often notice, if you check, that small format digital cameras simply do not have apertures smaller than f8, maybe even f5.6. Because anything smaller will simply have too much image degradation from diffraction, and manufacturers want to avoid this.

Why? Because f8 for example, is physically much smaller on a lens designed for a small format camera than f8 is on a lens designed for a 35mm camera. I suspect this may be what internet "pixel geeks" are referring to when they say diffraction can begin degrading image quality at f4. Maybe it can on a small format digital camera where everything is scaled down. If people do not understand that camera format matters, then they could be generalizing from this statement and drawing the wrong conclusion by assuming that it applies equally to all cameras.

Of course the other thing that affects aperture (physical) size on a lens is the focal length of that lens. Longer lenses in each format have the iris / aperture placed further away from the film plane, so the camera's iris has a larger physical size at each aperture setting. This is because an f stop effectively measures light hitting the film (sensor) plane. For the same amount of light to hit the film (sensor) on a long focal length lens compared to a short focal length lens, each aperture needs to be physically bigger in size. Diffraction occurs when light "grazes" the edge of the metal aperture blade and as a result its path changes very slightly -it is diffracted away form its expected course of travel, so it hits the film (sensor) in the "wrong" location, thus blurring the image. When the aperture is wide open and the hole is physically large, most of the light hitting the film (sensor) comes through the middle of the hole and is unaffected by this phenomenon. But when the aperture hole is small, proportionately more of the light passing to the film (sensor) grazes the edge of the aperture and is affected by this phenomenon and after it has been through the aperture it is traveling in a slightly different direction. Hence the diffraction effect is more noticeable on a (physically) smaller aperture than a larger one. Some diffraction is present when the aperture is larger - it is just less noticeable as less of the image is affected by it. If you doubt any of this, just get two lenses for the same camera. Say a 28mm wide angle and a 135mm tele lens. Set both to f5.6 and compare the size of those apertures. On the longer lens it is noticeably much larger.

Here is the one bit I science I can put into this response. The area of a circle (i.e the aperture hole in this case) is calculated by the formula [pi times radius squared]. The circumference of the a circle (i.e. aperture) is calculated by the formula [2 times pi times the radius of the circle]. Thus as the size of the circle (the aperture) increases, its area gets bigger much more quickly than its circumference - its area grows 'exponentially". As diffraction is affected by the light touching the circumference of the aperture, as the aperture is enlarged (e.g. as you open up from f16 to say f5.6) the diffraction effect quickly becomes much less as proportionately, much less of the light passing through the aperture has been affected by the circumference. Phew!

So, what matters with diffraction is the physical size of the aperture that the light is passing through. This is why Ansell Adams (who shot with large format plate cameras) could often use apertures as small as f64. He did it because it maximized depth of field. He could get away with it because on a large format camera f64 is physically a lot bigger than it would be on say a 35mm camera and hence there is less diffraction even with an aperture that is nominally f64.

That's it. You have cleaned me out. :^) That is about all I know. But while some geeks will talk about "circle of confusion" and blah blah blah none of which I really understand or want to, my explanation is still pretty accurate as I understand it.

 

[Richard G]

I avoid f22. But with digital, hand held, you can detect some motion blur from camera movement below 1/1000s. I consider a tripod more and more.

 

[willie_901]

From what I've read, one would have to use a specialty film designed for ultra, high-resolution applications to observe significant diffraction degradation with 24 X 36 mm media.

I also agree with the statement that handheld use often reduces resolution more than diffraction.

 

[YYV_146]

For handheld photography, diffraction is usually not a concern. With film the issue is usually non existent. The limitations are caused by processing and scanning, not by the lens.

But there are high-density digital sensors. The D800E's sensor can show differeaction at F7.1-8, and is best used at larger apertures. Smaller sensors with high pixel counts may incur diffraction even earlier.

Anyways, I would not shoot my full frame bodies at f22 or smaller, or my NEX-7 beyond f13. That's when diffraction becomes obvious.

 

[Juan Valdenebro]

F1.4 - 80% of a photo is out of focus, and the 20% which is in focus is 5% sharper.

F22 - 100% of the photo is in focus, and the 100% which is in focus is 5% softer.

Which is sharper? 😀

None of them...
Most lenses are softer at 1.4 than at f/8...
Thanks for the comments, everybody!
Sorry about the "it's" instead of "its" in the title...
Cheers,
Juan

 

[michaelwj]

Hi Jaun,

As Roger says, its far down the list of considerations. If you need he depth of field, then stop down. I think Roger @ Lens Rentals did some tests and found that a smaller aperture, whilst being softer, sharpens up to better across the whole depth of field or something like that. Either way, he concluded that if you need the depth of field, use the aperture you need.

For some science - diffraction happens at every aperture. Essentially, it's a 2D Fourier transform of the hole, so a smaller hole gives diffraction at larger spatial frequencies. Other things come into play too, like the propagation distance from the aperture to to the imaging plane, but we can ignore it for practical purposes. The problem arises when the diffracted light from the aperture spills into neighbouring imaging areas (grain/pixel, whatever). Therefore, at a given aperture, smaller pixels are impacted more. But, for a given print size it doesn't matter.

Cut back to reality. See first paragraph and set aperture for depth of field.

Cheers,
Michael

 

[Juan Valdenebro]

Depends on the lens... but you rather missed the joke, if there was one. 🙂

As Michael says: If you need the depth of field, then stop down.

Yes, I missed it but got it a second after I wrote... Thanks!

 

[Juan Valdenebro]

From the beginning I was wondering if maybe some forum members had examples of 35mm twin images, one from 1.4 to 5.6, and the other one (same image) at f/8 or f/11 showing a considerable sharpness loss...
Cheers,
Juan

 

[Monochrom]

Diffraction is real and also real is the fact that many lenses focused at infinite must be closed down to f11-16-f22 in order to get sharp enough corners.

So it´s subjective balance. everyone chooses what he needs.

 

[Ko.Fe.]

I like it at 2.8-8. I like it to be sharp where the subject is. I don't mind 16, but it is boring and unnatural. I don't mind it to be OOF as well.

Then you look at something only this something is in focus and the rest isn't.
To me more pictures where it is sharp from 1 meter to infinity are boring comparing to brownie, holga style shots. Those are on the art side to me.

And I see zero reason for scans at 4800. Waste of time and disk space.

 

[lukitas]

I have a 80mm f2.8 nikkor for enlarger, thread-mounted.
There was a distinct drop in sharpness from f8 onwards.

Then again, when I was a little boy, my dad told me to aim for f11, if at all possible, because everything would be sharp (at the hyper focal).
In those days, bokeh was shunned, photography was about representing reality exactly and sharply. At the most some fuzziness about the ears was accepted as inevitable.

And Ansel and friends went for f64...

 

[Roger Hicks]

In a practical sense you are quite right, diffraction effects do not really have a noticeable effect on an image till the lens aperture in use is quite small. In 35mm photography this usually means at least f16 or even smaller depending on the lens design. However, in a lens for a small format digital camera this can happen at f8 or even less . . .

Dear Peter,

You are of course absolutely right. I was thinking (lazily) only of 35mm/full frame digi. The point about larger formats is why minimum apertures for lenses for 35mm typically stop at f/16 or f/22; for medium format, at f/22 or f/32; and for LF at f/45 or more.

Cheers,

R.

 

[Ko.Fe.]

...

And Ansel and friends went for f64...

On LF, right?

 

[bigeye]

Distortion differs by lens. Numbers can quantify the difference, but to the eye it matters little for most photography.

The Summicron 50mm as an example (see MTF chart midway down the page): http://www.photozone.de/leicam/678-summicron50?start=1
Here's the Nikon 85 1.4G: http://www.photozone.de/nikon_ff/606-nikkorafs8514ff?start=1

I prefer f11 and larger, but would not hesitate to go smaller if necessary in 35mm.

(I think this all matters most for small sensor compact/P&S cameras, as noted above. Diffraction can crowd you toward the widest apertures, particularly since you have lower quality resolution to begin with.)

 

[EdwardKaraa]

It is said that the sharper the lens, the faster diffraction kicks in. Zeiss were proud to mention that their lenses are affected by diffraction as early as f/5.6. In practical terms, I don't think it makes much difference, even on a high pixel count digital sensor. I use very often f/11, and sometimes smaller if needed.

 

[Roger Hicks]

It is said that the sharper the lens, the faster diffraction kicks in. Zeiss were proud to mention that their lenses are affected by diffraction as early as f/5.6. In practical terms, I don't think it makes much difference, even on a high pixel count digital sensor. I use very often f/11, and sometimes smaller if needed.

Dear Edward,

Well... Sort of...

There is an absolute theoretical resolution limit at any aperture, so the better the lens is (by almost any criteria) the bigger the aperture at which diffraction limitation kicks in. The two biggest problems are that diffraction limits depend on wavelength, so assumptions/approximations have to be made there, and that resolution is a question of contrast, so assumptions need to be made as to what is a meaningful level of contrast.

That's before you start on Nyquist frequencies. Do not look for more precision in a system (or equation) than it can deliver.

Cheers,

R.

 

[EdwardKaraa]

Dear Edward,

Well... Sort of...

There is an absolute theoretical resolution limit at any aperture, so the better the lens is (by almost any criteria) the bigger the aperture at which diffraction limitation kicks in. The two biggest problems are that diffraction limits depend on wavelength, so assumptions/approximations have to be made there, and that resolution is a question of contrast, so assumptions need to be made as to what is a meaningful level of contrast.

That's before you start on Nyquist frequencies. Do not look for more precision in a system (or equation) than it can deliver.

Cheers,

R.

Agreed on all points Roger.

I would like to add that diffraction can act as an excellent soft focus filter at very small apertures (16 and smaller) and it removes moire too (especially on non AA filter sensors such as the M9 and M(240)).

While I do not totally disregard diffraction effects when I'm shooting, I try to use the required f/stop creatively without too much concern either.