PDA

View Full Version : Stopping down and "Diffraction"



Cletus
16-May-2013, 05:01
So, I've always been told, or read somewhere, never to stop down past, say, errrr, about f32-45 'cause anything at, or smaller than about f45 is getting into "diffraction limits" of the lens, or some such hokum.

Now, I know there is indeed such a thing as diffraction. I know it's something you need to be aware of, especially when printing using an enlarger with a lens with an aperture that can stop down past f45, or f32, or whatever the number is.

So here's my main question(s):

1. It seems like I can get a little more rise, or shift, etc. on my 8x10 with a certain lens the more I stop down. This is noral isn't it? True for all lenses, right? The more you stop down, the larger the IC of a given lens? Or at least I can see the hole in the diaphragm through the corners of the GG better when stopped waayyy down...

2. So on a rare occasion, when I need to stop down to f45, f64, or even off the scale toward f128 for MAX coverage (or even for any other reason to stop way down), is diffraction limiting really something I need to be....cognizant of? Will it really be apparent in my contact print made from an 8x10 neg? Will it really make a visible difference in sharpness for a small (11x14) enlargement from a 4x5?

I hear about "diffraction limiting" all the time, I'm just trying to imagine how much impact it might actually, practically speaking, have on a typical large format photograph?

Anybody here stop down to f64 and beyond with impunity?

Leigh
16-May-2013, 05:07
Diffraction limiting is one of those threats that mothers use to get their kids to eat Brussels sprouts, as in:
"If you don't eat your sprouts you'll be diffraction limited for the rest of your life."

Yes, it is a real physical phenomenon that can be measured in the laboratory.
It should be left in the laboratory.

Process lenses, which are the highest quality optics available, typically go down to f/128 or smaller.

The image degradation attributable to DL is generally indistinguishable from defocus.
It's interesting that DL appears to increase defocus while smaller apertures increase DoF.
For a given sharp point removed some distance from the plane of focus, DL will defocus it but DoF will sharpen it.

I use whatever aperture is appropriate to the situation, as did famous photographers of the past.

Yes, image circle diameter increases as you stop down.

- Leigh

rdenney
16-May-2013, 05:45
Generally, insufficient depth of field is a more noticeable effect than diffraction. But subject motion blur is a bigger effect still.

No effect is always a defect, of course.

I stop down as necessary for depth of field. That may force so long a shutter time that I can't control motion blur. That's when I either try to turn motion blur to my advantage, choose a different composition, wait for a calmer day, use faster film, or spend more effort finding the optimum focus plane position using tilts and swings.

But I never worry about diffraction. If diffraction is severe enough to undermine a given enlargement, I just don't make as big a print. That has happened only a couple of times in my decades of large-format work, though generally I don't make big prints. It happens more frequently with small format.

Rick "unwilling to eliminate a fine effect by creating a coarse effect" Denney

Leigh
16-May-2013, 06:12
Permit me a short discourse on diffraction, if you please.
Those disinterested in theory can skip this post.

Diffraction occurs when a beam of energy hits an edge that's opaque to that energy.

Some years ago I had the opportunity to work with diffraction in a very different setting, that being a circular
beam passing over a straight edge, rather than going through a hole.

In that situation it's very easy to see that the degradation due to diffraction depends entirely on the ratio of
the energy hitting the edge to the energy passing freely through the system.

Putting this in a photographic context, it's the ratio of the area of the aperture to its circumference that's important.
That's why long FL lenses are "diffraction limited" at smaller apertures (larger f/ numbers) than short FL lenses.

This also explains why small formats are more susceptible to DL issues than large format, since they use shorter lenses.

The problem with the usual explanation of DL and the Airy disc is that it is modeled using a small hole. Thus almost all
of the energy hits the edge, resulting in a worst-case effect.

- Leigh

N Dhananjay
16-May-2013, 06:35
Diffraction is real enough - as you say, you can see the grain turn to mush with a grain focuser under an enlarger lens. How big an issue it is varies according to taste and intention.

The usual trade-off is DOF vs. diffraction. Diffraction results in a uniform softening of the image whereas inadequate DOF shows up as areas of very good sharpness contrasted against areas of absolutely inadequate sharpness. Different people can, and do, make different calls about the appeal of these two 'looks'.

If you are shooting in a smaller format, you probably need to worry about it since the resolution drop can easily become noticeable when you enlarge the negative. The usual number bandied about is that you require about 5-10 lpmm resolution (let's say 7.5 lpmm) at a minimum on the print. Assuming about a 10X enlargement from 35mm, that will require a resolution of about 75 lpmm on film, making the unrealistic assumption of no losses elsewhere in the system - let us make the more reasonable assumption of requiring about 100-150 lpmm - this would mean probably the best lenses, very careful technique etc and I seriously doubt that the upper end is feasible but let's play along. Using a formula such as approx. 1500/f-stop to calculate diffraction limits, you can see that by about f/11 or f/16, you no longer get sufficient resolution on film. Now you will not see this when viewing the image because the 100 lpmm is well above what the human eye's limits are, so you might judge that you have got the DOF you need and are getting away without diffraction effects, but once enlarged, you will see the softness in the image. The problem is that by now the softness can get attributed to lots of other things (camera shake etc.)

With a typical contact printing process, you see the impact on the ground glass. Also, large negatives (and especially contact printing) is forgiving of many sins. So, if it looks good on the ground glass, (let's say a resolution 15 lpmm), the print is gong to look fine to the eye. You can get that level of resolution at f/90 - f/128. Beyond this point, of course, you are essentially using a pinhole and that money you spent on the expensive glass is a waste (at least for this image).

So, what does this mean in summary? If you are contact printing, you typically do not worry about diffraction because it is highly unlikely that you will see the effects on the print, even at extreme f-stops. With smaller formats, I would say you do need to worry about this because a) you do not see the effect when viewing the image - it looks fine to the eye (because you are seeing an image of say 50 lpmm - well above the requirements for the human eye) but when enlarged, the resolution drop is easy to discern.

Also, the IC increasing with greater stopping down depends upon the lens deign. What you are referring to (in terms of seeing the aperture from the corners - does it look 'football' shaped or round?) is a function of mechanical vignetting. Unless you've used extreme movements, it should not require stopping down more than a couple of stops to get past this region. On some lens designs (such as the Dagor), the lens can throw out a big cone of light but the resolution off-axis might be very poor and unusable - in fact, this is often the reason to mechanically vignette a lens). Off-axis resolution gets better as you stop down, which means more of that cone of light becomes usable as you stop down - i.e., the image circle does not get bigger but a larger part of it becomes usable.

Cheers, DJ

Cletus
16-May-2013, 06:47
Okay, so in relation to lenses, depth of field, image circle and coverage -

Does a smaller aperture, or f-stop ALWAYS increase the size of the image circle projected by the lens?

Does a smaller aperture, or f-stop ALWAYS increase the depth of field of the image?

And last, regardless of shutter speed (assume a bright, sunny day with no wind at all) at what point will stopping the lens down begin to impact the overall sharpness of the image?

I'm imagining me of those "infinity focus" landscapes of a road in the open desert, diminishing toward the horizon. I happen to be looking at a print I made like this right now, and it's not very 'sharp' at all - it occurred to me that I probably used a fairly small aperture, maybe f45, and the overall image just isn't very sharp in the details. If that makes any sense....I'm sure the other variables were under control - stable support, no wnd, etc...I was thinking the sharpness issue might have been related to diffraction limitation,or whatever the correct term is.

N Dhananjay
16-May-2013, 06:51
Permit me a short discourse on diffraction, if you please.
Those disinterested in theory can skip this post.

Diffraction occurs when a beam of energy hits an edge that's opaque to that energy.

Some years ago I had the opportunity to work with diffraction in a very different setting, that being a circular
beam passing over a straight edge, rather than going through a hole.

In that situation it's very easy to see that the degradation due to diffraction depends entirely on the ratio of
the energy hitting the edge to the energy passing freely through the system.

Putting this in a photographic context, it's the ratio of the area of the aperture to its circumference that's important.
That's why long FL lenses are "diffraction limited" at smaller apertures (larger f/ numbers) than short FL lenses.

This also explains why small formats are more susceptible to DL issues than large format, since they use shorter lenses.

The problem with the usual explanation of DL and the Airy disc is that it is modeled using a small hole. Thus almost all
of the energy hits the edge, resulting in a worst-case effect.

- Leigh

This is true. I find it helpful to think of a beam of light as a number of waveforms balanced with each other. So, each wave has wave besides it keeping it in check and preventing it from 'spreading' (speaking loosely). However, when an obstruction is placed in the path of the beam, the waveforms are unbalanced. The central wavelets are still balanced by other wavelets on the sides but the ones at the periphery (which got cut off by the obstruction) are not balanced and held in check anymore and are now free to spread.

It is not only a function of the ratio of area to circumference (which yields a function of r/2 - pi*r^2/2*pi*r). You also need to consider the fact that the resolution results from angular functions. So, a resolution limit at the diaphragm spreads as an angular function towards the film. The longer the distance travelled, the greater the angular spread. Which is why diffraction is usually defined in photography as only dependent on f-stop and not the format size or focal length.

Cheers, DJ

BrianShaw
16-May-2013, 06:52
I've believe diffraction to be a real phenomonon of optical physics, but in practice I've never seen it, especially at prints 11x14 and under. I suppose if one prints bigger than that, stands too close to the print to view, or scrutinizes with a microscope it might be detectable... but in terms of practicality I doubt it is something to fear.

p.s. I always ate my sprouts as a kid, and still do!

Cletus
16-May-2013, 06:53
Thanks DJ, nice explanation! And thanks Leigh and Rick too, just that some of the theoretical stuff is a tad over my head. If we want to talk tech about electrical theory and principles, I would be on more solid ground! Optical, not so much.. :)

N Dhananjay
16-May-2013, 06:58
Okay, so in relation to lenses, depth of field, image circle and coverage -

Does a smaller aperture, or f-stop ALWAYS increase the size of the image circle projected by the lens?


>>>>> I think it is more fair to say that smaller apertures increase the uniformity of resolution across the image area (i.e., on-axis and off-axis).



Does a smaller aperture, or f-stop ALWAYS increase the depth of field of the image?

>>>>> Yes.


And last, regardless of shutter speed (assume a bright, sunny day with no wind at all) at what point will stopping the lens down begin to impact the overall sharpness of the image? >>>>> This is a difficult question to answer without more details. If you were contact printing, it is highly unlikely that diffraction is an issue , even at f/90 or f/128. And it certainly won't be at f/45.

I'm imagining me of those "infinity focus" landscapes of a road in the open desert, diminishing toward the horizon. I happen to be looking at a print I made like this right now, and it's not very 'sharp' at all - it occurred to me that I probably used a fairly small aperture, maybe f45, and the overall image just isn't very sharp in the details. If that makes any sense....I'm sure the other variables were under control - stable support, no wnd, etc...I was thinking the sharpness issue might have been related to diffraction limitation,or whatever the correct term is.

>>>>> Again, assuming you are talking about a contact print, at f/45 I very much doubt diffraction is causing the softness. I would suspect other problems - shake, vibration, improper ground glass - film plane registration etc.

Cheers, DJ

E. von Hoegh
16-May-2013, 07:03
Some of you may want to do a star test. Put a light, say a 15w incandescent, behind a small hole, say .030" - this is your star. Place the star in a darkened room 15 or 20 feet from the camera, and observe the star on the GG. You'll need a decent loupe.
You can see diffraction rings with most LF lenses starting around f:22-32. It's also edifying to compare on and off axis light patches.

I just love sprouts!

Brian Ellis
16-May-2013, 07:19
"Does a smaller aperture, or f-stop ALWAYS increase the size of the image circle projected by the lens?"

No. It depends on the lens design. Some lenses continue to increase as you stop down, others don't. G Clarons are ones that do, which is why they can cover a lot more than Schneider's specs would indicate.

As has been said here over and over again by many people with a lot of experience, unless you're making what used to be considered a really large print (maybe larger than 20x24) diffraction is pretty much a non-issue with 4x5 and larger film.

Peter Gomena
16-May-2013, 07:29
I once worked a jewelry shoot where one photographer couldn't understand why his RB67 with a macro lens stopped down to f/45 made a mushy-looking picture. I suggested he open up a stop or so, and his picture became much crisper. Diffraction does happen and it is real.

I rarely photograph anything with an aperture wider than f/22 on my 4x5 unless the subject is at a fair distance and I don't need to stop down that far. When I did studio work, it was not uncommon to have to stop a 210mm lens down to f/32 or f/45 to pull depth of field on a tabletop subject. Diffraction effects were negligible if they were visible at all, especially when subjects were being scanned and screened for print. Even then, print sizes were small, say about 6x8" at the most for a half-page in a catalog. Moderate enlargement should show little effect of diffraction, but considering a second lens is involved in enlargement, you have increased diffraction effects again.

Film resolution is going to play into diffraction effects as well. If your film can't resolve detail below X, diffraction effects below X resolution probably aren't visible.

Pinhole images can be amazingly sharp in contact prints. Enlarging them or inspecting them under a loupe tells a different story.

Leigh
16-May-2013, 07:31
Put a light, say a 15w incandescent, behind a small hole, say .030" - this is your star.
This is exactly the problem I mentioned in post #4.

The small hole is what's causing the diffraction rings, not the lens.

- Leigh

Gem Singer
16-May-2013, 07:43
Why do lens manufacturers of LF lenses list their image circles at f22 (some at f16)?

Does closing down to f32, or f64 actually increase the size of the image circle or increase the depth of field?

E. von Hoegh
16-May-2013, 07:51
This is exactly the problem I mentioned in post #4.

The small hole is what's causing the diffraction rings, not the lens.

- Leigh

Any diffraction from the "star" is constant. When you observe the image on the GG the image patch changes as you change the aperture, it's very easy to see diffraction due the lens. Try it. :)

Nathan Potter
16-May-2013, 08:49
This is just a numbers game involving resolution limit on film vs the degree of enlargement. The figures one obtains from calculating a diffraction limit refer to the minimum size of a resolvable point at the plane of best focus and roughly within the DOF at that aperture. Beyond and in front of that plane rays diverge to form a blur pattern (out of focus representation of the point at best focus). The tradeoff is you either sacrifice the resolution at the plane of best focus or reduce the size of the blur pattern outside the plane of best focus by stopping down.

When you view a print the sharpness is highly subjective due both to the nature of the print and ones' tolerance for sharpness. Real numbers tell the story. For me in those cases where sharpness is important I really like to see maximum line width on paper of around 75 Ám (3 mils), about the width of a human hair. Others less critical may be happy with 6 mils or even more.

If you shot at f/45 the resolution limit at best focus is about 64 Ám so a contact print will certainly be critically sharp at the plane of best focus and will degrade progressivly outside that plane. But at a 2X enlargement you've doubled the 64 Ám to 128 Ám (5mils) still hardly noticeable image degradation. At 4X enlargement you're at 256 Ám (10 mils) and will definitely see a somewhat blurred image within the range of the DOF.

Nate Potter, Austin TX.

Dan Dozer
16-May-2013, 09:01
Here's a little different take on the issue - story as told to me by Kim Weston. Edward Weston was looking for extreme depth of field on one of his famous shell photographs and the lenses he had at the time didn't give him that. He worked with a local metal shop to make him a waterhouse stop for one of his lenses (don't know what lens) that was something like F256 or smaller. I saw the print in person and it looked fabulous. Now, taking into consideration that it was taken with 8 x 10 film and contact printed supports what others have said that it probably isn't an issue with large format film.

Jim Jones
16-May-2013, 09:04
I learned from experience to avoid stopping 35mm camera lenses smaller than f/8 for critical sharpness. Once I tried stopping a process lens down to f/128 on 5x7 film for maximum DOF. The resultant diffraction blur made the negative worthless. As Nathan suggested, other factors are often more important than diffraction. Consider Weston's macro photography for example. Some were shot with very small apertures.

(edit) Oops, Dan beat me to it.

Kirk Gittings
16-May-2013, 09:23
Here's a little different take on the issue - story as told to me by Kim Weston. Edward Weston was looking for extreme depth of field on one of his famous shell photographs and the lenses he had at the time didn't give him that. He worked with a local metal shop to make him a waterhouse stop for one of his lenses (don't know what lens) that was something like F256 or smaller. I saw the print in person and it looked fabulous. Now, taking into consideration that it was taken with 8 x 10 film and contact printed supports what others have said that it probably isn't an issue with large format film.

So true-size of enlargement is key to whether you can get away with small apertures or not. I only have shot 4x5 for many years. Sometimes dof is paramount and I need to stop down to f45. I know when I do that it will limit the print size I can make from the negative to say 11x14.

TimmyMac
16-May-2013, 10:01
Anyone care to explain what's going on with really retrofocal lenses? i.e. an 8mm circular fisheye on an EOS system with 45mm or so back focus, or a 14mm rectilinear? Is diffraction a bigger problem or is the apparent hole size larger thanks to some optical wizardry?

Ditto for the biogon derivatives where the aperture looks bigger from sharp angles... do they get some relief from diffraction at the periphery?

SergeiR
16-May-2013, 10:07
So, I've always been told, or read somewhere, never to stop down past, say, errrr, about f32-45 'cause anything at, or smaller than about f45 is getting into "diffraction limits" of the lens, or some such hokum.

Now, I know there is indeed such a thing as diffraction.
..
Anybody here stop down to f64 and beyond with impunity?

There is diffraction. However it hits you on different values for different formats. Also hits you differently from lens to lens. Ideally you got to experiment with every single lens to find "sharpest" if you are into these things.
I have closed down to 128/254 (diff lenses though) on 8x10 and didnt see diffraction effects yet.

Drew Wiley
16-May-2013, 10:17
The main variable is how much enlargement you need relative to the subject matter (it's a trade-off with depth of field considerations). You can get away with
very small f-stops in a contact print or magazine reproduction which might become obnoxious with even modest enlargement. Since I never print black and white
larger than 20x24, I commonly use f/64 for 8x10 and f/32 for 4x5. But I print color film bigger sometimes, so have to be a little more cautious. Edward Weston's
prints looked very sharp and snappy because they were contact-printed. But if they were only moderately enlarged, most of them would have looked horrible.
But sometimes he did just the opposite and shot rather wide open to get some wave splash nice and crisp; and again, it worked in a contact, but wouldn't have
at bigger scale.

Leigh
16-May-2013, 10:32
Why do lens manufacturers of LF lenses list their image circles at f22 (some at f16)?
Does closing down to f32, or f64 actually increase the size of the image circle or increase the depth of field?
Some manufacturers (Nikon and Schneider definitely) include IC and angle of view at both full aperture and f/22 (or f/16).

The choice of aperture depends on the speed of the lens, i.e. f/22 for f/5.6 lenses and slower, and f/16 for f/4 lenses.
It's generally at or close to the aperture that yields optimum quality. I suppose the industry standardized on those values
so customers could compare lenses from different manufacturers.

Stopping down smaller than f/22 will always increase DoF, and may increase the IC, but not beyond the design limits.

- Leigh

Tin Can
16-May-2013, 10:51
I like this plan, simple.

soon I try!


Some of you may want to do a star test. Put a light, say a 15w incandescent, behind a small hole, say .030" - this is your star. Place the star in a darkened room 15 or 20 feet from the camera, and observe the star on the GG. You'll need a decent loupe.
You can see diffraction rings with most LF lenses starting around f:22-32. It's also edifying to compare on and off axis light patches.

I just love sprouts!

rdenney
16-May-2013, 11:23
Why do lens manufacturers of LF lenses list their image circles at f22 (some at f16)?

Does closing down to f32, or f64 actually increase the size of the image circle or increase the depth of field?

Depth of field is a function of aperture size and focal length.

Diffraction is a function of aperture size.

Coverage is a function of lens design.

If you remove the film back and look through a lens from the corner of the film opening, you'll see the outline of the aperture, and the opening in the aperture should be illuminated by the scenery light coming through the lens. As you open up the lens, you'll see that part of the aperture is now providing you a view of the inside of the lens barrel. At apertures that large, the lens barrel will partly occlude the aperture and cause falloff. If it blocks half the opening, it causes one stop of additional falloff, etc. But once the aperture is small enough so that it is no longer blocked by any part of the lens barrel (or lens shade or filter ring), then stopping down further will not increase coverage.

More modern lenses are designed to perform edge to edge of their coverage circle to a given standard, and are designed to block image light outside that circle. But they are standardized at a small aperture which is what most photographers use. Thus, a Super Angulon is evaluated at f/22, and will provide the performance they specify to the point where the aperture is blocked by the lens barrel. At wider apertures falloff will be more pronounced and performance at the edge might not be as good. Older lenses are not blocked that way, and their image circles are specified as the limits of their specified performance. But the lens might well illuminate quite a bit more. That extra coverage will perform better at smaller apertures, at least until diffraction becomes critical (which is rare with large format), but perhaps not up to the specified performance of the lens.

Double-retrofocus lenses like the Super Angulon are designed to project a round aperture even at an oblique angle, which is why they suffer less from falloff than is theoretically assumed using the cosine rule, as long as the aperture is not occluded by the barrel. You can see that effect looking through the corner of the film opening, too.

Rick "who worries about coverage, but not really about diffraction" Denney

Tin Can
16-May-2013, 11:39
Good explanation, now I need to start really stopping down and quit fooling around with half-way measures. Oops, apertures...


Depth of field is a function of aperture size and focal length.

Diffraction is a function of aperture size.

Coverage is a function of lens design.

If you remove the film back and look through a lens from the corner of the film opening, you'll see the outline of the aperture, and the opening in the aperture should be illuminated by the scenery light coming through the lens. As you open up the lens, you'll see that part of the aperture is now providing you a view of the inside of the lens barrel. At apertures that large, the lens barrel will partly occlude the aperture and cause falloff. If it blocks half the opening, it causes one stop of additional falloff, etc. But once the aperture is small enough so that it is no longer blocked by any part of the lens barrel (or lens shade or filter ring), then stopping down further will not increase coverage.

More modern lenses are designed to perform edge to edge of their coverage circle to a given standard, and are designed to block image light outside that circle. But they are standardized at a small aperture which is what most photographers use. Thus, a Super Angulon is evaluated at f/22, and will provide the performance they specify to the point where the aperture is blocked by the lens barrel. At wider apertures falloff will be more pronounced and performance at the edge might not be as good. Older lenses are not blocked that way, and their image circles are specified as the limits of their specified performance. But the lens might well illuminate quite a bit more. That extra coverage will perform better at smaller apertures, at least until diffraction becomes critical (which is rare with large format), but perhaps not up to the specified performance of the lens.

Double-retrofocus lenses like the Super Angulon are designed to project a round aperture even at an oblique angle, which is why they suffer less from falloff than is theoretically assumed using the cosine rule, as long as the aperture is not occluded by the barrel. You can see that effect looking through the corner of the film opening, too.

Rick "who worries about coverage, but not really about diffraction" Denney

E. von Hoegh
16-May-2013, 11:59
I like this plan, simple.

soon I try!

Use a fine GG and at least a 10x loupe.

Tin Can
16-May-2013, 12:16
I have a couple of Steve Hopf's GG's that I like.

Loupes, not so fond of. For fine focus, I remove my very strong glasses and stick my eye right in there. Works for me.



Use a fine GG and at least a 10x loupe.

E. von Hoegh
16-May-2013, 12:18
I have a couple of Steve Hopf's GG's that I like.

Loupes, not so fond of. For fine focus, I remove my very strong glasses and stick my eye right in there. Works for me.

Unless you use a strong loupe, all you will see is a bright dot on the GG.

rdenney
16-May-2013, 12:20
I have a couple of Steve Hopf's GG's that I like.

Loupes, not so fond of. For fine focus, I remove my very strong glasses and stick my eye right in there. Works for me.

The august Herr Doktor von Hoegh (who is apparently still with us at age 148!) meant that you need a strong loupe for the diffraction star test he was proposing. For regular focusing, not so much.

Rick "who uses a 6x loupe for focusing 4x5, but would probably prefer a 4x on 8x10" Denney

Tin Can
16-May-2013, 12:27
Loupe it is. I think there is a song there..

Thanks!



Unless you use a strong loupe, all you will see is a bright dot on the GG.

Cletus
16-May-2013, 12:27
All these thoughtful comments are answering many questions I've had for quite along time now - what prompted this in the first place was me fooling around this morning with my new Fuji 210 W (inside lettering) and checking it's coverage on 8x10, which appears to be impressive, at least without having actually exposed any film with it yet. As I've been taught, with the lens stopped down to ~f/22 and the camera all zeroed out, I look through the clipped corners of the GG at the lens aperture, then rising the front until the visible circle of the aperture becomes elliptical. At that point, stopping down more until it's a circle again, then rising further, checking for the ellipse, stopping down, etc. until I am stopped down as far as possible, with as much rise as possible and thus, determining the coverage of the lens at given f-stop. As I understand it, as long as the visible aperture is still a circle and not an ellipse, you're still fully within the IC of the lens. As soon as the circle becomes an ellipse, that's the point where vignetting will start and is the limit of how much movement you can use with that lens. Does that sound about right?

Now, since I normally always just use f/32 (it always seemed like a good compromise to me between adequate DoF and reasonably short exposure) and only occasionally stopped down to f45 and very rarely f64 for fear of, um, getting any diffractions in the picture. From what I understand after reading, it appears I could comfortably use smaller than f/32 (wind and other conditions permitting) any time I want, for "maximum DoF". Or, if I'm approaching the limits of coverage on a particular lens and need to eke a few more mms of rise or whatever? If I just go from f32 to f64, I might get get another 5mms or so of rise, depending on the lens and format.

Bottom line, I will always, or almost always, prefer a smaller aperture to a bigger one and the only thing that's kept me from going smaller than f/32 was the belief that diffraction would become unmanageable and spoil my pretty picture. Since that doesn't seem to be the case and if you're not going for the 'selective focus' look and you can tolerate a longer shutter speed, I can't see any reason why "the smaller the aperture the better" wouldn't hold true under most circumstances. You get the most coverage, or IC, that the lens is capable of and you get more DoF - or at least depending the FL, subject distance and all the other things that DoF are related to and you get some forgiveness of focusing error related to imprecise camera movements - i.e. it may not be in perfect focus wide open, but stop down to f/128 and it's tack sharp all the way across!

So, unless I'm making BIG enlargements - and I'm not, most of the time I'm contact printing or making small enlargements - there's not much downside to always using the smallest possible aperture you can manage, for everything, all the time.

Heroique
16-May-2013, 13:24
...Other factors are often more important than diffraction.

The wind likes to blow away diffraction theories.

Vibration prefers to shake them off.

Let’s just say that if you worry about diffraction, and think you’ve protected your image from that narrow aperture, you might very well have saved it from the ever-sneaky breeze.

jp
16-May-2013, 14:50
How's this for my understanding (correct me if I'm wrong).

1. There is always diffraction where there is an aperture.
2. When you stop down you decrease the amount of refracted/transmitted light such that the diffracted light's proportion is increased and that mix negatively affects image quality.

An example of this is too small of a pinhole on a pinhole camera hurts rather than helps image quality.

Dan Fromm
16-May-2013, 15:03
Um, in fact the relative aperture to use in calculating the diameter of the blur circle/diffraction limit in lp/mm at low contrast is the effective relative aperture, i.e., the relative aperture adjusted for magnification. In out-and-about work the two are essentially the same, but in closeup work the effective aperture is smaller than the nominal.

The relationship between effective and nominal aperture depends in part on the lens' pupillary magnification and orientation. This is why a reversed retrofocus lens used for closeup work is much dimmer, given the aperture set, than a lens of normal construction.

For a clear explanation, see Lester Lefkowitz' book The Manual of Closeup Photography.

Oren Grad
16-May-2013, 15:46
All these thoughtful comments are answering many questions I've had for quite along time now - what prompted this in the first place was me fooling around this morning with my new Fuji 210 W (inside lettering) and checking it's coverage on 8x10, which appears to be impressive, at least without having actually exposed any film with it yet. As I've been taught, with the lens stopped down to ~f/22 and the camera all zeroed out, I look through the clipped corners of the GG at the lens aperture, then rising the front until the visible circle of the aperture becomes elliptical. At that point, stopping down more until it's a circle again, then rising further, checking for the ellipse, stopping down, etc. until I am stopped down as far as possible, with as much rise as possible and thus, determining the coverage of the lens at given f-stop. As I understand it, as long as the visible aperture is still a circle and not an ellipse, you're still fully within the IC of the lens. As soon as the circle becomes an ellipse, that's the point where vignetting will start and is the limit of how much movement you can use with that lens. Does that sound about right?

No. It's not enough to see where the lens starts to mechanically vignette. The circle of illumination can be and often is larger than the circle within which a lens renders with adequate performance (by whatever standard you'd like to apply).


So, unless I'm making BIG enlargements - and I'm not, most of the time I'm contact printing or making small enlargements - there's not much downside to always using the smallest possible aperture you can manage, for everything, all the time.

Run your own test. Seriously. Don't guess based on theoretical arguments or on others' rules of thumb; that can lead you astray. Real-world lenses with real iris diaphragms and different cocktails of aberrations depart in a variety of ways from idealized formulas, and people differ in their printing and viewing habits and perceptions. Spend four or five sheets of film on an aperture series with the lens you want to use and judge for yourself what your practical limits are.

Cletus
16-May-2013, 17:57
Thanks Oren, all what you just said makes good, common sense. Actually, the last remark about the "using the lens stopped way down all the time" was me being just a tad facetious, but I do get the understanding through reading these comments that stopping down past my usual f32 might not be quite as bad as I thought it to be.

Actually, if I were to dig back through the dusty cobwebs of everything I've ever read, I'm sure I'd remember something about using the largest possible aperture, or at least as close to whatever the optimum f-stop for the given lens is, after having properly applied the necessary movements to focus the camera. Theoretically, unless you're shooting a subject or scene that has many complex planes (so to speak), you should be able to focus pretty much tack sharp with the lens wide open. For those times when you just can't get everything you want in focus, you can stop down to try to compensate for the rest using DoF.

I'm (definitely) not saying all this because I think you don't know it - or even if my ideas are entirely accurate - I was just making the point that my habit of shooting stopped down all the time is partly to make up for some of my own shortcomings and abilities (or inclination) to spend a lot of time focusing and applying movements. Also because there seem to me to be a few desirable advantages, as well as some trade-offs (i.e. diffraction limitations), to shooting stopped down instead of at larger apertures for the kind of stuff I tend to shoot.

I'm just glad to know that I can, from time to time, shoot at f45 or 64 if I thought the situation called for it, without worrying that I'm just throwing away film to do so. :)

Dan Fromm
16-May-2013, 18:27
I'm just glad to know that I can, from time to time, shoot at f45 or 64 if I thought the situation called for it, without worrying that I'm just throwing away film to do so. :)

Cletus, sometimes you'll waste film that way, sometimes you won't. And that's where knowing how large you intend to print when you're shooting and knowing the rules of thumb that resolution on film (in lp/mm) can't be better than 1500/(f/ number) and that the minimum acceptable resolution in the final print is 8 lp/mm are helpful.

For example, back when ISO Kodachrome was still available, I sometimes shot flowers at effective f/45. My resolution on-film can't have been better than 30 lp/mm. A print with at best 8 lp/mm, starting from 30 lp/mm on film, can't be larger than 4 x the negative. This rules out acceptable prints larger than 4" x 6" from my little 1 x 1.5 piece of Kodachrome.

Rough and ready, but sure enough 8x10s from those slides weren't sharp enough. One of the many ways in which larger format can beat smaller format.

All this assumes that sharpness is what matters most. It often isn't, and then considerations of resolution just go away.

Doremus Scudder
17-May-2013, 03:07
Cletus,

The goal is to optimize your aperture for the image you are making. The depth-of-field vs. diffraction tradeoff has been discussed in depth on the LF Homepage here: http://www.largeformatphotography.info/fstop.html

It is well worth reading. I've been using my version of this for years; I find close and near focus points, note the focus spread at the camera bed/rail, center the standard in the middle of the spread, and set the aperture according to a chart I have taped on my cameras. And, if I have to stop down to f/45 or more for a given subject, I know where the limits of print size are.

Of course, other factors include subject movement, wind, etc. This usually makes the largest aperture that is still smaller than the lens's "sweet spot" (usually f/22 or thereabouts for 4x5 lenses) that will do the job for DoF the best choice.

Best,

Doremus

E. von Hoegh
17-May-2013, 07:32
The august Herr Doktor von Hoegh (who is apparently still with us at age 148!) meant that you need a strong loupe for the diffraction star test he was proposing. For regular focusing, not so much.

Rick "who uses a 6x loupe for focusing 4x5, but would probably prefer a 4x on 8x10" Denney

148 years and 12 days. At this age the days count... ;)

Brian Ellis
17-May-2013, 07:40
I learned from experience to avoid stopping 35mm camera lenses smaller than f/8 for critical sharpness. Once I tried stopping a process lens down to f/128 on 5x7 film for maximum DOF. The resultant diffraction blur made the negative worthless. As Nathan suggested, other factors are often more important than diffraction. Consider Weston's macro photography for example. Some were shot with very small apertures.

(edit) Oops, Dan beat me to it.

As I'm sure you know, 35mm and diffraction is a whole different deal than LF and diffraction. Consider that even an 8x10 print from a 35mm negative is approximately an 8x enlargement or the approximate equivalent of a 32x40 inch print from 4x5 film. Which is one reason why the smallest aperture on 35mm lenses tends to be around f/22.

Jim Jones
17-May-2013, 18:48
A basic rule of thumb when using a lens with normal coverage for the format; stop down so the entrance pupil is about 5 or 6mm in diameter for a good compromise between DOF and diffraction limiting. For other focal lengths, adjust that diameter by the ratio that a lens is longer or shorter than a normal lens. Also, the hyperfocal distance for critical work is maybe 2000 times the diameter of the entrance pupil, depending on many factors. This simplifies other calculations.

paulr
17-May-2013, 21:56
The effect of diffraction on image quality is complex. Some of the simplified explanations we get fed often obfuscate the issues. The point that I think gets lost is that diffraction does not act as a kind of "brick wall" resolution filter. It simply diminishes MTF. Diffraction needs to be expressed as a curve, not a number.

A related phenomenon is that a great deal of the MTF loss from diffraction can be regained through sharpening, especially if you choose an appropriate sharpening algorithm. Deconvolution algorithms can do a remarkable job, almost eliminating the effects of diffraction down to f16, and reducing it substantially down to f32 or so. These algorithms require an image with a very high s/n ratio, so they're generally easier to use with digital capture than with film.

The more important point here is that diffraction blur is much easier to correct than defocus blur. So all else equal, you're probably better off stopping down too far than not enough.

It's true that stopping down becomes more desireable when there's a lot of shift. A lens that might be diffraction limited on axis at f11 might not be diffraction limited until f45 out near edges of the image circle. There's a complex conversation between shift, depth of field, and diffraction. And if the wind is blowing, good luck.

To the question about stopping down increasing the size of the image circle: the technical answer is that it never does. This is only an apparent effect. Some lenses have enough falloff at wide apertures that the outer edges of the image circle are too dim to be noticeable. But the edges are there! Lenses that don't exhibit this effect are simply ones that have more even illumination.

Dan Fromm
18-May-2013, 06:35
To the question about stopping down increasing the size of the image circle: the technical answer is that it never does. This is only an apparent effect. Some lenses have enough falloff at wide apertures that the outer edges of the image circle are too dim to be noticeable. But the edges are there! Lenses that don't exhibit this effect are simply ones that have more even illumination.

Interesting if trie. On the one hand, stopping down doesn't expand the circle illuminated. On the other, for most lenses stopping down reduces aperture sensitive off-axis aberrations and expands the circle of adequate definition until diffraction dominates.

E. von Hoegh
18-May-2013, 08:30
148 years and 12 days. At this age the days count... ;)

OOPS! At my age one tends to get confused. In this case, it was by the different conventions in writing dates, and the fact that I had to look up my own birthday (10.5.1865). It's 148 years and 7 days, on the date I posted this. Go ahead, make fun of an old, old man. An old, old, man who brought you the Dagor.

Randy Moe, Leigh pointed out that there will be diffraction due to the hole which forms the star. This is not really a factor in the test, but if you'd like to eliminate any possible diffraction from the star, simply glue a piece of cigarette paper over the hole you made in a piece of sheet metal. A 1 mm (.040~) hole will work just as well).

Tin Can
18-May-2013, 08:39
Rolling papers, who has those? Not me for 30 years.

Regarding the hole. Can a round enough hole be made by a #60 drill bit, which is 0.040"? That is my smallest bit.

I have not yet completed my Star Wars laser weapon...


OOPS! At my age one tends to get confused. In this case, it was by the different conventions in writing dates, and the fact that I had to look up my own birthday (10.5.1865). It's 148 years and 7 days, on the date I posted this. Go ahead, make fun of an old, old man. An old, old, man who brought you the Dagor.

Randy Moe, Leigh pointed out that there will be diffraction due to the hole which forms the star. This is not really a factor in the test, but if you'd like to eliminate any possible diffraction from the star, simply glue a piece of cigarette paper over the hole you made in a piece of sheet metal. A 1 mm (.040~) hole will work just as well).

E. von Hoegh
18-May-2013, 08:48
Rolling papers, who has those? Not me for 30 years.

Regarding the hole. Can a round enough hole be made by a #60 drill bit, which is 0.040"? That is my smallest bit.

I have not yet completed my Star Wars laser weapon...

Maybe there's a friendly dope fiend in your locale? :)

Well, actually I made mine with a smaller bit and broached it perfectly round. But I think it will work just fine, I can't think of any problem other than the diffraction rings you'll see on the GG might not be perfectly round. Be advised that twist drills often make a hole slightly larger than their nominal size due to the cutting edges not being perfectly centered. A 1/16" hole will work fine, too - it isn't critical.

Edit - Chamfer each side of the hole with a slightly larger bit.

Edit edit- put the paper on the side facing the camera.

Tin Can
18-May-2013, 08:53
Oh, I forgot, you do watch repair, no wonder you have tiny broaches!

I do need smaller bits for many purposes, time to shop for tools, my favorite activity right after shooting women.

That sounds so wrong, except on this forum!




Maybe there's a friendly dope fiend in your locale? :)

Well, actually I made mine with a smaller bit and broached it perfectly round. But I think it will work just fine, I can't think of any problem other than the diffraction rings you'll see on the GG might not be perfectly round. Be advised that twist drills often make a hole slightly larger than their nominal size due to the cutting edges not being perfectly centered. A 1/16" hole will work fine, too - it isn't critical.

Bill Burk
18-May-2013, 09:27
This has been a great discussion and the only thing I didn't see was Ralph Lambrecht's chart from Way Beyond Monochrome.

Ralph shared the chart on a special interest group site some of our members subscribe to.

To avoid violating LFF terms, I will just tell you a search term where you will find a thread started 03-18-11 by member T42: Puzzled about Diffraction

See post #5

Doremus Scudder
19-May-2013, 04:14
This has been a great discussion and the only thing I didn't see was Ralph Lambrecht's chart from Way Beyond Monochrome.

Ralph shared the chart on a special interest group site some of our members subscribe to.

To avoid violating LFF terms, I will just tell you a search term where you will find a thread started 03-18-11 by member T42: Puzzled about Diffraction

See post #5

Bill,

I see nothing in the usage guidelines for the LF forum that forbids linking to APUG. Here is the relevant excerpt from the guidelines.

"Links to published articles - Direct links (i.e. not through index pages that contain promotional text or blatant advertising) to on-topic published articles are OK to post, as long as those pages are freely available to all (i.e. no paid subscription required), and do not require further contact with the poster to gain access.
When posting links to material on other web sites, please provide a link to the material, not the home page of the site. Repetitive postings of links to other discussion boards are not appropriate."

So, since I found the discussion about diffraction you referred to informative, here's the link for those that may not have puzzled out exactly where you were referring to: http://www.apug.org/forums/forum48/88939-puzzled-about-diffraction.html

Also, here's a link to the Cambridge Color web page about diffraction; also very informative and with a couple nifty interactive diffraction calculators: http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm

Have fun,

Doremus

Tin Can
19-May-2013, 07:14
The tutorial at Canbridgeincolour is excellent. I suppose there is app for this. I liked using their interactive calculator.




Bill,

I see nothing in the usage guidelines for the LF forum that forbids linking to APUG. Here is the relevant excerpt from the guidelines.

"Links to published articles - Direct links (i.e. not through index pages that contain promotional text or blatant advertising) to on-topic published articles are OK to post, as long as those pages are freely available to all (i.e. no paid subscription required), and do not require further contact with the poster to gain access.
When posting links to material on other web sites, please provide a link to the material, not the home page of the site. Repetitive postings of links to other discussion boards are not appropriate."

So, since I found the discussion about diffraction you referred to informative, here's the link for those that may not have puzzled out exactly where you were referring to: http://www.apug.org/forums/forum48/88939-puzzled-about-diffraction.html

Also, here's a link to the Cambridge Color web page about diffraction; also very informative and with a couple nifty interactive diffraction calculators: http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm

Have fun,

Doremus

Bill Burk
19-May-2013, 08:11
Bill,

I see nothing in the usage guidelines for the LF forum that forbids linking to APUG...

Doremus

Doremus,

I really did want to link into that thread because I felt it would add value to this discussion. Thanks for checking and linking...

Bill

keithb
19-May-2013, 09:53
So, I've always been told, or read somewhere, never to stop down past, say, errrr, about f32-45 'cause anything at, or smaller than about f45 is getting into "diffraction limits" of the lens, or some such hokum.
[...]

I hear about "diffraction limiting" all the time, I'm just trying to imagine how much impact it might actually, practically speaking, have on a typical large format photograph?



I think of it this way: as I stop down to a tiny, tiny aperture I'm moving towards having a pin hole. All that lovely, expensive glass just isn't getting much chance to influence the light passing through. Recall the thin lens diagrams you drew in school physics classes, rays through the optical centre of the lens (where the aperture is, in non-telephoto designs) are not refracted. That might not be exactly true for a more complex lens design, but you get the idea.

How small is small? f/64 on a 90mm lens would be about 1.4mm But, how small is a pin hole? The Harman TiTAN comes with a hole that's quoted as f/250 when 110mm away from the film, so that's 0.44mm, about one third the size of my hypothetical f/64 aperture.

And what are the characteristics of a pin hole? Enormous image circle, "infinite" depth of field, but nowhere sharp. Is that "bad"? Depends on the effect you want to achieve.

Best Regards,
Keith

Peter Yeti
26-May-2013, 13:49
This has been a very inspiring discussion with lots of good information. The theory is clear but it's often even more convincing when verified by an experiment. So, I took advantage of the really lousy weather here this weekend, wasted one extra sheet of film, and spent some time in the darkroom.

Here is the still life that I abused for this test, taken with a Sinaron S 5.6/180mm at f22 on a Sinar p with 4x5'' back and lit with studio strobes. I'd think this is a fairly safe set up regarding wind or other sources of vibration.


https://sites.google.com/site/gourmetyeti/_/rsrc/1369595097402/photos/still-2013/eggs_sm.jpg


I exposed a second sheet at f64 for comparison, keeping everything else the same. The negatives revealed differences in definition when viewed with a 10x magnifying glass. I picked out the tip of the towel in the lower right corner for further scrutinizing. I made sure that this part was in the focus plane wide open at f5.6 in order to avoid any bias. Then I made 15-fold enlargements of this section, using my 50mm enlarging lens, and tried to keep the tonal values as close as possible between the two negatives. Here are the scanned results:

@f22:
https://sites.google.com/site/gourmetyeti/_/rsrc/1369595116855/photos/still-2013/f22_sm.jpg


@f64:
https://sites.google.com/site/gourmetyeti/_/rsrc/1369595134574/photos/still-2013/f64_sm.jpg


I think that the results are pretty convincing. Most obvious is the loss of contrast in the f64 version. The loss of resolution is much more subtle and only visible in the most minute detail, which is at the borderline of the (very high) resolving power of this lens. I think this is pretty much in agreement with the statement that MTF rather than resolution suffers most from too high diffraction.

However, keep in mind that the detail I enlarged is equivalent to a 13x18 cm fraction of a 150x190 cm print from a 4x5'' negative. The softer micro contrast will affect smaller prints as well but resolution may not really be a matter of much concern. Anyway, it's a good idea to keep diffraction as low as possible to obtain crisp detail and contrast.

I hope this test will be helpful for others as well.

Jim Andrada
26-May-2013, 21:19
I guess what I'd like to see would be a comparison based on enlarging to 16 x 20 rather than floor to ceiling size.

Interesting test and thanks for doing it.

This begs a question though - why does my 480mm Apo Ronar have a stop marked 256? I can't believe they would put "useless" stops on the lens

Struan Gray
27-May-2013, 00:01
Graphics arts lenses have very small stops for making halftone separations using crossline screens (or other types of screen). Roughly, the individual dots of the halftone are small images of the aperture of the enlarging lens. The exact shape of the dot is also influenced by the line screen, the height of the screen above the halftone film, and the way the halftone film is processed. In analogue halftone screening optimizing the shape and spacing of the halftone dots for different screens and different tonal transfer curves is as much art as science, but one of the tools available is to adjust the aperture of the lens. Fine screens need small dots, so lenses for this application would often stop down really far, much further than anyone would ever use for normal picture taking.

Armin Seeholzer
27-May-2013, 02:38
Thanks Peter

It shows what I did know for a long time, diffraction is overrated! In my beginning days in LF I tried not to go past f32 and many pictures where just for the garbage, because not enough DOF. I still try not just to stop down if its not needed, but if its needed I really give a s.... of it!

Cheers Armin

Peter Yeti
27-May-2013, 03:39
I guess what I'd like to see would be a comparison based on enlarging to 16 x 20 rather than floor to ceiling size.


Jim, you are certainly right that 16x20 is the more relevant size for most of us. But at 4-fold magnification the effect would be so tiny that you couldn't display it on the web.

Armin: "It shows what I did know for a long time, diffraction is overrated!"

I think everyone can decide for himself what matters more. I usually do it like you and rate DoF higher when needed.

Jody_S
27-May-2013, 05:11
This has been a very inspiring discussion with lots of good information. The theory is clear but it's often even more convincing when verified by an experiment.

(...)

I hope this test will be helpful for others as well.
Thanks, this is very helpful. I'm going to stop worrying about it for now, because there's no way my scanner could pick up that difference. Someday, if I get a drum scanner, or I'm shooting something that will be sent out, I'll start thinking about it again.

BrianShaw
27-May-2013, 06:11
Thanks, Peter... very instructive! So what's for supper... eggs + basil sounds like fritatta to me.

Peter Yeti
27-May-2013, 15:39
OT: Hmm, frittata sounds very good. But not with those studio eggs, which are at least a couple of years old. The fresh basil is in danger, though.:)

Jim Andrada
28-May-2013, 02:29
Hi Struan

Thanks much - I get it. I've been so conditioned to think of these lenses the way we use them that I didn't think back to their original uses! When I used to work with Lith film I was using the neat stuff that Kodak used to sell with the elliptical dots already in the film.

Struan Gray
28-May-2013, 04:13
Always a pleasure to pass on some of the useless knowledge I've acquired over the years :-)


PS: incidentally, the halftone people are among the few who seem to have taken diffraction seriously, including second order effects (Fresnel diffraction) and coherence. This paper is quite useful:

W. Streifer, R.N.Goren, and L.M. Marks
Analysis and experimental study of ruled halftone screens
Applied Optics, 13, 1299 (1974)
http://www.opticsinfobase.org/ao/abstract.cfm?URI=ao-13-6-1299

Ed Richards
28-May-2013, 06:36
I have also done some real world testing. If you have important fine (high frequency) detail, you can get visible effects at 4x enlargement. Assuming you have a reason to stop down a lot, the benefit of the DOF outweighs the loss of fine detail. (While some obsess over fine detail, it is seldom critical to anyone other than other photographers.)

Bernice Loui
28-May-2013, 10:04
Depends on what the image maker demands in the finished image as to how important fine detail and other factors are.

It was not too long ago when I took a group of 10x14 B&W prints made from 5x7 negatives to a party/get together of non-photographers. As they looked at each print, it was clear and event that they were drawn into the print by the fine details, tonality and more... Viewing distance would have been 6" to maybe 24". There was no question among the viewers that the fine details had appeal to this group.

Fine details or extreme DOF (everything "sharp") or any other single obsession alone will never make an expressive print, it takes far more than that.

There are so many aspects of what makes a great expressive print and so many ways to achieve it.

The effects of Diffraction is very real, like other aspects of the image making process, it is up to the image maker to be aware of this effect and decide how it's effects are to their finished print.


Bernice







I have also done some real world testing. If you have important fine (high frequency) detail, you can get visible effects at 4x enlargement. Assuming you have a reason to stop down a lot, the benefit of the DOF outweighs the loss of fine detail. (While some obsess over fine detail, it is seldom critical to anyone other than other photographers.)

Bernice Loui
28-May-2013, 10:06
Maybe Green Eggs & Ham...?


Bernice


Thanks, Peter... very instructive! So what's for supper... eggs + basil sounds like fritatta to me.

Emmanuel BIGLER
28-May-2013, 10:53
Hello all!

Coming late to this discussion and after having carefully read the exchanges, I'd like to address a few questions that have been raised, and as far as I've seen, not yet answered.


From TimmyMac, 16-May-2013


Re: Stopping down and "Diffraction"

Anyone care to explain what's going on with really retrofocal lenses? i.e. an 8mm circular fisheye on an EOS system with 45mm or so back focus, or a 14mm rectilinear? Is diffraction a bigger problem or is the apparent hole size larger thanks to some optical wizardry?


Most models we use to explain what is going on in a large format lens are based on the approach of a thin, single lens element with the aperture stop located at its center. But most, if not all lenses that we use daily are thick compound lenses. It is a kind of a miracle that such a simple and naive model can explain so many photographic things: object-image formulae, depth of field, F-numbers and photometry, and perspective rendition.
So how can we combine what we know from diffraction in a single lens element with the true reality of a thick compound lens?

The simplest approach, which sounds reasonable when the lens is stopped down by several clicks with respect to its maximum aperture, is, first, to neglect diffraction effects that occur at the edges of all lens mounts inside the barrel, and take into account only diffraction by the iris itself.
Second, we'll be crazy enough to believe that all geometrical aberrations are negligible. Hence we are left with something like a perfect thick compound lens, pure geometrical optics, except diffraction effects by the iris. This problem is in itself very complex, but we can further simplify it by considering a single monochromatic source point, which generates a single spherical and monochromatic wave on input, to which corresponds, in a diffraction-free and aberration-free optical system, a perfect converging spherical wave on output. And we'll limit the lateral dimensions of this converging beam by something called: the exit pupil, i.e. the geometrical image of the iris as seen from the back of the lens, from the image point.

Hence after outrageously simplifying the problem, we are left with an academic exercise, "compute the distribution of light around a perfect converging spherical wave limited by a certain circular aperture located at a certain distance from the converging point."

If the lens is a retrofocus lens, the diameter of the exit pupil will be bigger than the diameter of the entrance pupil. Hence we could dream that this aperture being larger, diffraction effects will be smaller. But here we are not dealing with diffraction effects for a parallel beam of light, but with a converging beam of light, and we look exactly at the place where the beam converges. Assume, to begin with, for simplicity, that the input source is located very far away, so that the converging beam focuses in the focal plane. Let "a" be the diameter of the entrance pupil, "a' " the diameter the exit pupil, "M_p" the pupillar magnification ratio = a'/a, and "f" the focal length of the lens. The f-number N (=11, 16, 22, 32, ..) is defined as usual by N=f/a with reference to the entrance pupil diameter a and the focal length f.

Actually for a retrofocus lens, the magnification ratio M_p is greater than one, and it can be shown that the distance "d" between the exit pupil and the focal plane is larger than f, it is equal to
d = f . M_p
hence bigger than f in a retrofocus lens and smaller than f in a telephoto.

Now, the basic diffraction theory says that that the size of the diffraction spot, the famous Airy disk (if the aperture is circular) is given approximately by
1.2 lambda d/a'.
So the hope we had to get less diffraction with a retrofocus lens is in vain, since the exit pupil is bigger but when the source is located far away and the image located in the focal plane, the distance d is bigger, exactly in the same proportion :
d/a' = f.M_p / (a . M_p) = f/a = N
whichever the pupillar magnification ratio M_p can be.
Hence a retrofocus lens, or a telephoto lens, of nominal f-number N, for an object located at infinity, whichever magnification ratio M_p can be, exhibit exactly the same Airy disk of diameter 1.2 N lambda, N = f/a defined with reference to the entrance pupil.

What happens in the macro range? Nothing special, except that we have to replace the distance d = f.M_p by a modified distance f.(M + M_p) when M is the image/object magnification ratio. The extension beyond the focal plane is f.M for all lenses whichever value of M_p, hence the total distance is f.(M + M_p).

The formula is slightly more complex than the simple factor (1+M) which is valid for symmetrical lenses (M_p=1) but this is not a big deal to compute: the Airy disk becomes equal to 1.2 N_{eff} lambda, where N_{eff} is the effective f-number = N(1+M/M_p), the same formula is used for the photometric bellows factor in a asymmetrical lenses (Hello Dan, yes I have a copy of the excellent book by Lefkowitz at home !), simply because the amount of light per square area in the macro range depends on the ratio d/a', same ratio in use to compute the size of the Airy disk in the basic diffraction theory.

Hence, after this lengthy explanation, we can sleep quietly, retrofocus lenses or telephoto lenses, at least for landscape use, behave exactly like a single lens element limited by an iris located at the centre of the lens.


Ditto for the biogon derivatives where the aperture looks bigger from sharp angles... do they get some relief from diffraction at the periphery?

Yes, in principle but the explanation for slanted rays in the corner of the image is very difficult to give without a drawing. In a few words: imagine that we observe the Airy disk on a piece of paper placed perpendicular to the beam and not parallel to the film. The lateral size of the Airy disk in this situation, measured on the paper and not on the film, is about 1.2 d/b. where "d" is the total distance between the center of the exit pupil and "b" the apparent width of the exit pupil which will be seen from the corners of the field as an ellipse even if the pupil is not cut by a lens mount (the famous cat's eye shape, we assume we do not see the cat's eye). Hence the Airy disk is no longer an Airy disk, but an Airy ellipse ;)

In a lens like the biogons, grandagons and super-angulons, pupillar distorsion partially compensates for the smaller size of the projected ellipse, like if the exit pupil would rotate and be 'less elliptical' in projection.
The analysis of the image quality becomes complex since we have to figure out how the image of a fine grid is projected in the image plane itself and not simply on our small piece of paper perpendicular to the rays.
Whether the grid bars are in the sagittal or meridional orientation, the projection differs, there two different factors with various cosine powers (cos^2 and cos^3 but do not ask me which is which) for the ultimate limit of sharpness for sagittal or meridional grid orientations. Hence even a perfect aberration-free wide-angle lens will exhibit sagittal and meridional FTM curves separating from each other in the corners.
And YES, pupillar distorsion, in principle, somewhat compensates for diffracttion effects, but in terms of image quality in the edges, actually residual aberrations are very important in the corners, I'm afraid that pupillar distorsion , even helpful, cannnot do much for this.

------------------

And another remark just to say that I totally agree with Jim Jones (17-May-2013, 18:48) when he says :
A basic rule of thumb when using a lens with normal coverage for the format; stop down so the entrance pupil is about 5 or 6mm in diameter for a good compromise between DOF and diffraction limiting.

If you look at this figure #12 (http://www.galerie-photo.com/images/scheimpflug-pf-mf_h012.png) in this article (in French, sorry) (http://www.galerie-photo.com/decentrement-bascules-scheimpflug-petit-moyen-format.html) which is a compliation of the best f-number for standard lenses (no wide angle no telephoto) either recommended by lens manufacturers or found by experience, you'll find, for classical lenses, that the best f-number obeys a simple trend: best N = f(in mm)/(8 mm) i.e. from 35 mm photography to the 8x10" format, standard lenses should have their entrance pupil open to about the same diameter, 8 mm. So Jim's rule of thumb @5 or 6m, within one or 2 mm exactly confirms this trend.

--------

And I promise that I'll engrave in solid gold (like the mythical golden disk engraved at the JPL and sent to ET in the Voyager or Pioneer Deep Space Program) those two most useful sentences, and hang them above my chimney:

Diffraction limiting is one of those threats that mothers use to get their kids to eat Brussels sprouts, as in: "If you don't eat your sprouts you'll be diffraction limited for the rest of your life." (Leigh)

The wind likes to blow away diffraction theories. (Jim Jones)

Dan Fromm
28-May-2013, 11:17
Diffraction limiting is one of those threats that mothers use to get their kids to eat Brussels sprouts, as in: "If you don't eat your sprouts you'll be diffraction limited for the rest of your life." (Leigh)

The wind likes to blow away diffraction theories. (Jim Jones)

All well and good, Emmanuel, but in coseup work diffraction can be very limiting because effective apertures are small. For illustrations (actual pictures!) of this, including illustrations that show depth of field getting smaller on stopping down, see Kodak Publication N-12B Photomacrography or N-16 Closeup Photography and Photomacrography.

Those who don't have either Kodak publication and are seriously interested in working closeup should buy one of them. If you haven't read either or tried the experiments -- I have the books and have done the experiments, the illustrations in the books aren't faked -- don't object to the statements above.

I used to shoot flowers and such around 1:1 on KM with a 105/2.8 MicroNikkor on a Nikon SLR. Flash illumination, of course. All mention of this tiny format is strictly forbidden in an LF forum. I often set the lens' aperture to f/22. That's f/45 effective at 1:1, and my slides' enlargeability suffered severely. Its a problem with no good easily portable solution. I'm aware of confocal techniques, haven't seen any way to take them to the field. Nowadays I shoot flowers with ISO 100 E6 in a 2x3 camera, another format too small to mention here, and suffer much the same problem. What the larger format buys me, simply, is a better chance to get an acceptably large image of the main subject without cropping out all of its setting.

In closeup work, diffraction is a killer.

Emmanuel BIGLER
28-May-2013, 11:28
In closeup work, diffraction is a killer.

Hello Dan and thanks for reading my impossible post ;)
Macro shots with a LF camera in windy conditions .. sure that your sharpness will get killed twice.

Dan Fromm
28-May-2013, 11:40
In closeup work, diffraction is a killer.

Hello Dan and thanks for reading my impossible post ;)
Macro shots with a LF camera in windy conditions .. sure that your sharpness will get killed twice.

Wind is a killer with any camera but an SLR even with flash illumination. Flash will stop motion but when there's any wind it can't keep the plane of best focus where intended when the shot was set up. The camera doesn't have to be affected by the wind, even a gentle breeze will blow the subject around enough to spoil a shot. One has to know when not to try ... I have many failed milkweed shots, alas.

Struan Gray
28-May-2013, 12:38
Emmanuel, thanks for the nice post.

Dan: I've seen some interesting work done by handheld focus stacking. I suspect this will gradually end up in firmware. Automated scan-and-stitch combined with focus stacking is fairly widely available on research grade microscopes.

It's getting faster and faster, so eventually you will be able to defeat a light breeze. With gale force winds you'll still have to make lemonade.


http://struangray.com/miscpics/thistle.jpg

Cirsium palustre (at f32)

Drew Wiley
28-May-2013, 12:47
LF shots of intricate details in windy conditions is what I do all the time. No choice around here. The wind rarely stops. But one does get good at sensing every little
nuance of the breeze. But March breezes which can literally lift and toss my 8x10 along with the big Ries tripod - that's another story! I allowed myself one shot
out on Pt Reyes this past Sat, and yes it was constantly windy. But a half second can seem like forever.... waiting, waiting for just the right instant for things stop moving.

Emmanuel BIGLER
29-May-2013, 02:15
Thanks to Struan for sharing with us this subtle image of "Cirsium palustre (at f32)".
This image demonstrates that, when using a fast shutter speed, not only waterfalls loose their specific mood ;)

And regarding the use of LF cameras for macro work, I had the privilege to have a close look at a selection of original prints by Karl Blossfled, an exhibition at the MontbÚliard castle & museum close to Besanšon. Blossfeld used to bring the plants back in his studio. This helps using a LF camera for macro work!