Diffraction?

[SchwinnParamount]

I still have the box and 5.25" diskettes for the first version of Microsoft Office. I don't have an available rig to run it on. On the other side, I've got this Mac and Office 2011 runs wonderfully on it. Unfortunately, this laptop doesn't even have a CD drive. If I want new software for it, I must download it from the cloud. Fortunately, server technology and availability has gotten really really good.

 

[RalphLambrecht]

This page has the diffraction calculator toward the bottom: http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm

Also note the diffraction calculation is independent of the focal length.

if a diffraction calculator asks for the f/stop number (N) then the focal length is baked into that, of course. a rrough calculation of the fiffraction limit for any f/stop is 1,800/N.

 

[RalphLambrecht]

it doesn't take much to show diffraction.Today's high-end DSLRshave sufficient resolution to showa resolution loss from f/11or16 compared to f/22or32. my D800 for example has a resolution limit of 73lp/mm,which makes it diffraction limited at f/22.

 

[RalphLambrecht]

+1

 

[Dan Fromm]

+1

Really, Ralphie, patting yourself on the back.

 

[cowanw]

It also needs to be kept in mind that most DOF calculators (and probably diffraction calculators as well) use what is termed 'manufacturer standard' assumption of human visual acuity, whereas the 20/20 vision standard (in the US) is actually about 3x as good as the manufacturer assumption. For example, if I shoot 8x10 at f/22 and focus at 10', most DOF calculators tell that your DOF zone is from 16" to Infinity; using 20/20 vision, DOF zone is from 38" to Infinity.

How does the fact that prints are viewed as closely as the viewer wishes impact on that?

 

[wiltw]

How does the fact that prints are viewed as closely as the viewer wishes impact on that?

Viewing CLOSER simply means that your eye can detect 'blur' far MORE READILY than any calculator program thinks you will -- even if 20/20 visual standard is used, or the far more tolerant 'manufacturer standard'!

So if any program says 'no visible diffraction at f/16', you would probably see it even at f/8 because you eyes are better AND you are viewing from a closer distance than assumed.

 

[DREW WILEY]

Depends on the film or printing paper too. I cut my teeth on Cibachrome and large format, so what most people consider acceptably sharp I
kinda regard as mush. That doesn't mean prints necessarily have to be that sharp to be esthetically pleasing, but it is a different kind of standard. But there's inherently a lot of BS whenever anyone tries to peg this down to some cut-and-dried formula, just like that "circle of
confusion" and "normal viewing distance" stuff. Diffraction is something to be aware of because it's there. Hope it impacts you and what to do about it all depends. I happen to be someone who needs to keep it in mind rather routinely, but always balanced with other concerns, like desirable depth of field, shutter speed, blah blah.

 

[RalphLambrecht]

Depends on the film or printing paper too. I cut my teeth on Cibachrome and large format, so what most people consider acceptably sharp I
kinda regard as mush. That doesn't mean prints necessarily have to be that sharp to be esthetically pleasing, but it is a different kind of standard. But there's inherently a lot of BS whenever anyone tries to peg this down to some cut-and-dried formula, just like that "circle of
confusion" and "normal viewing distance" stuff. Diffraction is something to be aware of because it's there. Hope it impacts you and what to do about it all depends. I happen to be someone who needs to keep it in mind rather routinely, but always balanced with other concerns, like desirable depth of field, shutter speed, blah blah.

a circle of confusion is a group of photographwrs sitting around a table talking about depyh of fieldand if you want a deeper understanding,you need to give up photography and study mathematics.

 

[Rick A]

a circle of confusion is a group of photographwrs sitting around a table talking about depyh of fieldand if you want a deeper understanding,you need to give up photography and study mathematics.

Bazinga!

 

[L Gebhardt]

What are your views on this topic? For 35mm format, don't stop the lens further than F8? If need be use a filter to cut the light out? Having seen those who do shoot at F11, F16 or even F22 - is it that bad? There also may be times when one wants a deliberate slow shutter for creative effects, do you simply just pop on a filter?

The only answer is to test for your uses. The depth of field and diffraction calculators all make assumptions that are not how I print or look at images. I think for every scene there is going to be a limit on print size dictated by the combined effects of required depth of field, diffraction effects, focal plane curvature, film resolution, camera shake, atmospheric haze, and your viewing requirements. You really need to define your goals to answer your question.

Here's a bit of my experience with testing lenses. With a Nikon D800E (high res digital) and a sharp lens I can see diffraction effects starting a bit before f/8. To see this in a print requires heavy cropping or a large print. Most depth of field calculators (and lens DOF markings) assume an 8x10 print viewed at arms length. In that situation you would be hard pressed to see any diffraction up to f/22 from 35mm. However if you print 16x24 the depth of field calculator will be off by a couple of stops and diffraction will be much more visible.

In addition each lens is different. They all tend to improve as you stop them down, up to the point that diffraction effects start to become dominant. Some lenses are at their best at f/2.8 and others are there at f/11 despite diffraction. I have a Nikon 200mm f/4 AI that improves up to f/16 in some aspects even as it's getting less sharp from diffraction. It's not a great lens optically, but it still makes nice images. I don't worry about using it at f/11, but I know I won't be getting a tack sharp large print no matter what I do. I also have a 300mm f/4 AFS that's razor sharp at f/4. If I need more depth of field I don't hesitate to stop down. I just know I won't be able to print as large (both because of inadequate depth of field and diffraction effects).

When I shoot 35mm film I know I won't enlarge past 16 x 24, and that would be be very rare for me. Most of my prints are 11x14 or 8x10. So I tend to set an upper limit of f/11 and try to shoot at f/5.6 or f/8 unless I see a reason to deviate (more or less DOF usually). I also use filters to cut light so I can stay in that range and get the effect I want.

I'll generally stop down to the following apertures before I even start to consider diffraction:

35mm --> f/8
120 (6x6 or 6x7) --> f/11
4x5 --> f/22
5x7 --> f/32

 

[wiltw]

Most depth of field calculators (and lens DOF markings) assume an 8x10 print viewed at arms length. In that situation you would be hard pressed to see any diffraction up to f/22 from 35mm. However if you print 16x24 the depth of field calculator will be off by a couple of stops and diffraction will be much more visible.

I need to insert a very important additional bit of information to everyone's understanding...

"assume an 8x10 print viewed at arms length" ...by someone with LESS THAN 20/20 visual acuity!

In other words, the average person who has 20/20 vision will see 'this area is blurred!' rather than it appearing to be in 'sharp enough' size of CofC in the enlargement. In this thread I had already brought up the point in Post 22.

 

[L Gebhardt]

I need to insert a very important additional bit of information to everyone's understanding...

"assume an 8x10 print viewed at arms length" ...by someone with LESS THAN 20/20 visual acuity!

In other words, the average person who has 20/20 vision will see 'this area is blurred!' rather than it appearing to be in 'sharp enough' size of CofC in the enlargement. In this thread I had already brought up the point in Post 22.

I saw what you wrote, but I've seen some variation in what people claim the manufacturers use. Some say 8x10 viewed at arms length, some say 8x10 viewed at a foot with someone of less than 20/20 vision. Those may really be the same, and there are probably other claims out there as well. What I haven't seen is a statement from say Nikon on how they calculated the markings. It's all a bit nebulous and really makes no difference to the fact that you can't trust the numbers implicitly.

 

[Sirius Glass]

a circle of confusion is a group of photographwrs sitting around a table talking about depyh of fieldand if you want a deeper understanding,you need to give up photography and study mathematics.

... optics and physics.

 

[cowanw]

The thing is that 20/20 visual acuity is measured at 20 feet and says nothing about what one can see at 2 feet.

 

[RalphLambrecht]

The thing is that 20/20 visual acuity is measured at 20 feet and says nothing about what one can see at 2 feet.

the visual angle of acuity doesn't change, does it?:confused:

 

[cowanw]

Hard to say, because the ability to accommodate up close does. If you can't accommodate then the angle of acuity is a mute point.

 

[wiltw]

Hard to say, because the ability to accommodate up close does. If you can't accommodate then the angle of acuity is a mute point.

The 'standard' of being able discern a character of a certain height at a certain distance defines the ANGULAR acuity that a human should be be capable of achieving. And that is an established value. As you point out, being able to ACHIEVE that at reading distance (or any other distance) is merely physiology, though! In the eye the natural lens grows throughout life, and studies suggest that the enlarging of this lens with age is the cause of its reduced pliability resulting in the inability of the lens to change shape when one looks from far to near objects. A young child - who has the most pliable lens in their eyes -- much better accommodates to seeing things super close, even better than a 20 year old, and both beat the 60 year old -- who has the least piable lens in their eyes - hands down! And that is why reading glasses, to compensate for the declining ability to accommodate for close distance. But the STANDARD of visual acuity does not 'vary'. And while that standard does not vary, the CofC calculation is not a single universally accepted value!

The general statements made about human visual acuity define the angular measure, but often it is restated as 5 lines-per-millimeter, which corresponds to a CofC size for 35mm of 0.025mm. Also, the 'Zeiss formula' defines the circle of confusion as calculated as d/1730 where d is the diagonal measure of the original image (the camera format). For full-frame 35 mm format (24 mm × 36 mm, 43 mm diagonal) this comes out to be 0.025 mm. But a much more widely used CoC is d/1500, or 0.029 mm for full-frame 35 mm format, and values of 0.030 mm and 0.033 mm are also common for full-frame 35 mm format.

Some say 8x10 viewed at arms length, some say 8x10 viewed at a foot with someone of less than 20/20 vision.

We have a range even for 'normal' reading distance! Depending upon your source you hear values of 15-25". Commonly, the assumption often is about 16". Reading vision selftest charts that you find state that you should find the right reading glasses by viewing the 5" chart at a distance of 14".. The Grafco Jaeger Eye Test Chart assumes 14". The "arm's length" or about 25" is probably defined by the 40 year old fighting the need for reading glasses.
Then we get to the definition of 'in focus' and the ability of the eye to perceive a Circle of Confusion as a blur circle vs. as a virtual point, with the values of 0.025 - 0.033mm as stated above!
Lastly, we have the definition used by a given DOF calculator for what distance...and I have heard 8x10" print at 10", or at 12" distance (about the diagonal).

No wonder there is NOT COMMON AGREEMENT about any of these three different values!

 

[blockend]

In the mid-1980s I shot a lot of landscapes at f16 on 35mm, mostly on Nikon lenses. I printed 16 x 12" carefully, with a Nikon enlarging lens. Looking at them recently, all showed clear diffraction issues.

Something that intrigues me is how many modern lenses show their best performance close to wide open, sometimes within half a stop of maximum aperture. I don't know what's changed since the days of manual optics which peaked around f5.6 -f8, perhaps computer design has enabled manufacturers to tweak their lenses for portraiture at the expense of resolution further down the dial.

 

[E. von Hoegh]

In the mid-1980s I shot a lot of landscapes at f16 on 35mm, mostly on Nikon lenses. I printed 16 x 12" carefully, with a Nikon enlarging lens. Looking at them recently, all showed clear diffraction issues.

Something that intrigues me is how many modern lenses show their best performance close to wide open, sometimes within half a stop of maximum aperture. I don't know what's changed since the days of manual optics which peaked around f5.6 -f8, perhaps computer design has enabled manufacturers to tweak their lenses for portraiture at the expense of resolution further down the dial.

Many, if not all, modern 35mm lenses are well into diffraction limits at f:16, quite a few at f:11. When I want the maximum detail I try to stay around f:5.6, depending on the lens and film combo.

 

[blockend]

Many, if not all, modern 35mm lenses are well into diffraction limits at f:16, quite a few at f:11. When I want the maximum detail I try to stay around f:5.6, depending on the lens and film combo.

I agree, but think I the point may have been misunderstood. I'm saying that on test some modern lenses are at their sharpest wide open. The recent Canon 40mm 2.8 and 10-18 zoom are actually at their sharpest at the widest aperture, and sharpness drops as the aperture is closed. That was never previously the case. As you say, two or three stops down was usually the sweet spot, with diffraction kicking in after f8.

 

[Dan Fromm]

Um, sharpest edge-to-edge or sharpest in the center? I ask because received doctrine is that stopping down reduces off-axis aberrations that are sensitive to aperture. For this reason as the lens is stopped down the circle that's truly diffraction limited grows.

Years ago Brian Caldwell and I discussed this with reference to the 55/2.8 MicroNikkor. Turns out that at f/4 the center 8 mm circle is diffraction limited. IIRC, may not have the number right, Brian considered a Strehl ratio > 0.85 to mean "diffraction limited."

With this in mind, I'd be astonished if the Canon 40/2.8 really is diffraction limited at the edge of a 43 mm circle.

People who say "diffraction kicks in at f/whatever" are annoying. Diffraction is always there. It doesn't "kick in" at an aperture, at some aperture it becomes the dominant influence on the blur circle. And this has nothing to do with what's capturing the image.

 

[Sirius Glass]

People who say "diffraction kicks in at f/whatever" are annoying. Diffraction is always there. It doesn't "kick in" at an aperture, at some aperture it becomes the dominant influence on the blur circle. And this has nothing to do with what's capturing the image.

As Dan said: Diffraction is always there for any aperture. The diffraction grows until it becomes the dominant blur circle influence. Manufacturers usually design their lenses to stop reducing the aperture before that point. See my earlier posts on this thread.

 

[Alan Klein]

I calculate (or view using the manual stop down lever) the f stop based on the DOF I need. Then I stop down one additional stop to be sure. Is this a bad practice?

 

[LJH]

Cropping changes nothing in the image. Now do you understand why your unsubstantiated statement is wrong?

I didn't write that cropping changed things. I wrote that it is impossible to put a predetermined criteria on a lens if you don't know its intended application. Now do you understand why your unsubstantiated statement is wrong?