Diffraction Limit on Macro Lenses

[DolphinDan]

Hi, I searched for an answer to this question on the forum but did not find it. So here is my question: what is the diffraction limit on modern macro lenses, for example the Rodenstock 180mm f5.6 APO-Macro-Sironar lens?

I have been shooting flowers at 1:1 on 4x5 and 5x7 with this lens set to f64 to maximize DOF, and the results seem OK on my computer. However, I have not printed any of my photos at 16"x20" or larger yet (I like printing my images in the 16"x20" to 32"x40" size range). I know regular lenses are diffraction limited at about f22, but I was wondering if macro lenses were designed differently such that the diffraction limit kicks in at f32 or f45.

I would like to maximize DOF without getting into visible diffraction (I usually photograph landscapes at f32 without noticing any diffraction; at f45 I start to see some diffraction). I usually print my scanned photos at 300DPI or 360DPI (EPSONs). Can anyone suggest usable apertures for modern macro lenses, formulas for figuring it out, and/or online information on the subject?

Daniel

[Bob Salomon]

f22, but if you are happy at f32 then use it.

[Dan Fromm]

Dan, the standard rule of thumb is that diffraction limited resolution is 1500/(effective aperture). For lenses with pupillary magnification = 1 (that's most of the lenses we use), effective aperture is aperture set * (1 + magnification).

So, at 1:1 with f/64 set effective aperture is f/128 and the highest resolution possible is 1500/128, i.e., about 12 lp/mm at very low contrast. Shooting at f/64 limits you to contact printing; enlargements will look fuzzy.

One of the paradoxes of macro work is that stopping down can reduce depth of field. To learn more about this, read:

Gibson, H. Lou. Close-Up Photography and Photomacrography. 1970. Publication N-16. Eastman Kodak Co. Rochester, NY. 98+95+6 pp. The two sections were published separately as Kodak Publications N-12A and N-12B respectively. Republished in 1977 with changes and without the 6 page analytic supplement, which was published separately as Kodak Publication N-15. 1977 edition is ISBN 0-87985-206-2.

This is the most terrifying book on photographic technique I've ever read.

Incidentally and all kidding aside, unless you like uniformly fuzzy prints there no DoF to speak of at 1:1, regardless of the aperture used. If you like uniformly fuzzy prints, don't use a lens, use a pinhole.

Cheers,

Dan

[Nathan Potter]

The diffraction limit has nothing to do with the lens but only depends on the relative f/no. or more precisely the physical size of the aperture opening. A convenient metric for determining the size of an Airy disc diameter (the size of a resolvable spot on the film) is obtained by considering light at 400 nm wavelength. Then the diameter of the minimum resolvable spot is equal to the f/no in microns. Thus the Airy disc diameter at f/22 is 22 um. The diameter at f/64 is 64 um. The diameter at f/1.2 is 1.2 um!

One can convert a 22um spot to lines per mm (l/mm.) by dividing 1000 (1000 um per mm.)by the Airy disc diameter. Then get line pairs per mm. (lp/mm.) by dividing again by 2. So your f/22 represents about 22.5 lp/mm. and your f/64 equals about 8 lp/mm. f/5.6 the common maximum diameter of large format lenses is equivalent to about 90 lp/mm. as a diffraction limit.

Of course above says nothing about the optics in the lens which can degrade the image further although in some cases can offer some suttle but minor optical improvements. And we can all quibble about using the diameter of an Airy disc (the first dark fringe in the diffraction pattern) as a standard but it is a readily observable and measurable phenomena.:)

Nate Potter, Austin TX.

[Jeff Conrad]

The diffraction limit for the plane of focus may be different from that for the DoF limits. Once the lens is essentially diffraction limited, further stopping down decreases sharpness in PoF. At the DoF limits, there is the additional tradeoff between defocus blur and diffraction; stopping down decreases the former but increases the latter. Eventually, diffraction predominates, and stopping down further gives less overall sharpness.

As long as the combined blurring from defocus and diffraction is imperceptible, the tradeoff of sharpness at the DoF limits vs. that in the PoF isn't relevant. Sharpness in the PoF is always greater than that at the DoF limits, and if the sharpness at the DoF limits is acceptable, the sharpness in the PoF will be acceptable as well. The tradeoff may become an issue when great DoF is needed, especially in close-up work.

In his article in the March/April issue of Photo Techniquesi (available on this web site as GIF images), Paul Hansma discusses the combined effects of defocus and diffraction. I also discuss it in my paper Depth of Field in Depth (http://www.largeformatphotography.info/articles/DoFinDepth.pdf) (on this web site, in PDF) under Combined Defocus and Diffraction. Although Hansma and I take quite different approaches, we come up with very similar rules for optimal f-number. Hansma gives


N = sqrt(375 Δv),

where N is the f-number and Δv is the focus spread. I give


N = 20 sqrt(Δv)

for 12 lp/mm in the negative.

It should be noted that we both give the case of infinity focus (i.e., zero magnification). For close-up work, the effective f-number must be used, i.e., N (1 + m), where N is the marked f-number and m is the magnification. At 1:1, this cuts the optimal f-number in half.

[rdenney]

One comment: "Diffraction-limited" is high praise for a telescope objective. It means that the lens faults are so minor that diffraction controls the image quality.

Thus, I think it's more useful to ask at what point diffraction causes more problem than stopping down solves. As Nathan has described, diffraction increases as the aperture decreases. So, a smaller aperture will always show more diffraction than a larger aperture.

But lens faults and depth of field are the competing influences. Most people want to know the sweet spot of a given lens, where further stopping down shows net degradation--diffraction gets worse faster than improvements from correcting lens faults and increasing depth of field.

In the macro range, depth of field is always a gigantic issue, and the larger the format, the easier it is to get into the macro range just taking pictures of normal subjects. My suspicion is that for most images, it would be rather hard for an image to be ruined by diffraction (limited in enlargeability, perhaps, but not ruined), while not having sufficient depth of field or correcting for aberrations (such as field curvature, etc.) can ruin the result at any print size. The challenge is finding the aperture that balances those competing influences, and that point is different for every lens and subject, it seems to me.

For copy work or other technical application, one might have a specific performance standard, such as a print of X size that requires a resolution of Y line pairs/mm, which can then be backed into a calculation. For three-dimensional subjects, however, I suspect this has little to do with the results.

A 16x20 is a 4x enlargement. If the visual standard is 5 line pairs/mm, then you need at least 20 line pairs/mm in the original image. Based on Nathan's numbers, f/22 is indeed your limit. But you'll have to ensure that the other influences on the image don't do worse things than the minor problem you might see at, say, f/32.

Rick "perfection is unattainable" Denney

[Jeff Conrad]

Rick, I agree that in most cases, diffraction isn't what ruins an image. But this changes in close-up work, so that there really can be a visible tradeoff between DoF and overall sharpness. There's a good example on p. 84 of Lester Lefkowitz's The Manual of Close-Up Photography.

[Jeff Conrad]

Dan, I was writing my reply when you posted yours. It should be noted that the approach Hasma takes (i.e., root-square combination of defocus and diffraction blur spots) is essentially the same as Gibson took in 1960, when his analytic section was first published. So the idea is hardly new--it just hasn't received a lot of attention.

It should be noted that Gibson apparently made an error in his basic formula for DoF (2nd ed., Vol. II, p. 95), using overall magnification (subject to final image) rather than magnification in the negative. When the error is corrected, it is seen that nearly all the difference between getting all the magnification in the negative (e.g., large format, contact print) and from enlarging the negative is due to loss of sharpness in the negative emulsion.

In 1955, H.H. Hopkins developed a more theoretically solid approach calculating the MTF for combined defocus and diffraction. I used Hopkins's approach, and that my results were very similar to Hansma's suggests that the root-square combination of defocus and diffraction at least captures a fair amount of what's happening.

But it doesn't always seem to work. In the same issue of Photo Techniques that I mentioned, Stephen Peterson attempts to find the minumum f-number that would give acceptable sharpness (his article is also available on this web site). In effect, his technique subtracts diffraction blur from total blur, and he eventually arrives at an f-number beyond which the desired sharpness cannot be achieved. I got far less correlation between his maximum f-number and that predicted by calculating MTFs with Hopkins's method than I did with Hansma's optimal f-number; in fact, I got much better correlation with the method many people on this forum have long used, i.e.,


N = Δv / 2c,

where c is the circle of confusion. It's possible to regard this value as a minimum f-number and Hansma's as a maximum. Unfortunately, in close-up work, the minimum can sometimes be greater than the maximum, suggesting that you can't get there from here. The value from this formula is greater than Peterson's; the actual value is probably somewhere in between.

[Nathan Potter]

Jeffs' additional comments on the added effects of defocus has reminded me of the OPs' original desire to maximize the DoF for macro work. I think it has been mentioned here, maybe even by me, that there is a technique to obtain infinite depth of focus with a macro setup by using a scanning technique. This is variously called Scanning Light Photography as well as other terminology. It's been around for a long time.

The idea is to illuminate the plane of focus, which would include the defocus regions to an acceptable circle of confusion, using two or three slits of light that are of a thickness approximating the thickness of the plane of acceptable focus and intercept that plane exactly and be coplanar with that plane throughout the field of view. Then the subject is translated through the slit illumination at a uniform rate and a velocity
that is consistent with a proper exposure dose. (Not so easy to explain briefly).

If interested Google C. J. Kazilek, "Scanning Light Photography" for a description with images and references.

It needs to be a pretty sophisticated setup but I used it some years ago for documenting some MMIC modules with considerable success. If the OP is really serious about great depth of field this cannot be beat.

A unique feature is that the image obtained has no perspective distortion since every part of the subject is always the same distance from the camera.

Nate Potter, Austin TX.

[Jeff Conrad]

The technique Nate describes is also known as multi-plane scanning. Sidney Ray's Applied Photographic Optics (http://books.google.com/books?id=cuzYl4hx-B8C&pg=PA231&lpg=PA231&dq=%22multi-plane+scanning%22+Ray) gives a brief description, as well as several other references.

Another alternative is software combination of images at various focus distances. It doesn't need a sophisticated setup, but it's obviously more of a chore with LF than with small- or medium-format digital. And it benefits from a stationary subject, which can be tough with flowers.

I've not personally used either technique, so I can't offer much additional comment.

Perhaps after providing a lot more information than the OP may have wanted, an attempt at practical answer to his question might be appropriate. Let's assume that DoF as well as sharpness in the plane of focus is important, and consider an example at 1.0 magnification, and for which the focus spread is 10 mm. Using my formula for optimal f-number at the DoF limits,


N = 20 sqrt(10) / (1 + 1) = 32 .

There's no point in using a smaller aperture because the image will get less sharp, even at the DoF limits. Assuming the standard 4x5 CoC of 0.1 mm, the conventional geometrical formula gives


N = 10 / [2 x 0.1 x (1 + 1)] = 25 ,

so it should be possible to get the desired DoF. The geometrical formula ignores diffraction, and probably underestimates the minimum f-number, so a better value might be 32, the same as given for the optimal f-number.

What happens if the minimum f-number is greater than the optimal? Use the optimal, because using a greater f-number will make the image less sharp. When this happens, you can't get the desired DoF, and you also may be getting sharpness at the DoF limits at the expense of that in the plane of focus, and may need to decide which is more important.

[Emmanuel BIGLER]

Hello all !
Coming late into this discussion ....
what is the diffraction limit on modern macro lenses
... my understanding of the question is : how far can I stop-down before diffraction "appears" in practical use.
As answered by Bob, Rodenstock says : you can go for f/22.

But I am wondering whether there would be some difference on this best f-stop between the classical apo ronar (a 4/4 formula) and modern LF macro lenses like the apo macro sironar (a 6/4 formula) from the same company.
Since the apo macro sironar (and other competing LF macro lenses from other well-know companies) is made of 6 lens elements, one should expect to gain something with respect to the classical 4/4 apo ronar design. Sure, the apo macro sironar is listed for a coverage of 60° @1:1 magnification, whereas the apo ronar was listed for 48°.
So if the extra two lens elements mainly serve to increase the angular coverage, a best f-stop@f/22 (effective = f/45@1:1) is the same for the good ol' apo ronar and the modern apo macro sironar...

Regarding the apo ronar, I have handy a copy of an old brochure of the eighties aimed at those discerning users doing high quality repro work ; the brochure shows a variety of MTF plots presented like in the tutorials in optical engineering as a function of the spatial frequency up to the cut-off.
One can say that in yellow monochromatic light (535 nm), at the centre of the field and @f/22 the apo ronar's MTF cannot be distinguished from a perfect diffraction-limited lens, the absolute cut-of frequency being 50 cy/mm.
At 20° from the optical axis, the result is still incredibly close to the absolute limit, the cut-off does not change that much but the contrast @20 cy/mm gradually drops until the very end around 25° (total field about 50°)

Hence if we consider that a modern 6/4 LF macro cannot be less good than an apo ronar, for sure @f/22 and at the centre of the field the lens is so good that residual aberrations are negligible. Hence I'm sure that you could stop down@f/16 only (f/32 @1/1 magnification) and be happy... for taking a picture of a flat object.
For something fitting inside a Scheimpflug wedge.... one might get increased DOF as usual by using an appropriate tilt (like in the movie "Microcosmos"
http://www.16-9.dk/2007-04/side11_inenglish.htm
featuring ants & other insects moving on a plane slanted with respect to the camera's optical axis)

[GPS]

Emmanuel, I don't get it. The diffraction depends on the aperture size, for a given wave length. Now, how you arrive at that aperture - if through a 4 elements lens or a 10 elements elephant has no effect on the aperture size. 10 amp current is 10 amp if you transformed it 2 or 10 times before you came to the 10 amp.
Who cares about the number of elements if one wants to know diffraction?

[Dan Fromm]

George, its a question of residual aberrations and the aperture, if any, at which diffraction swamps them. Remember that not all aberrations are affected by aperture. Lens design matters.

Its also a question of how large the diffraction-limited field is. Remember that some off-axis aberrations are affected by aperture, so that the diffraction-limited field grows on stopping down.

Some years ago I remarked somewhere that the 55/2.8 MicroNikkor is diffraction-limited at f/4. Brian Caldwell, who has the prescription and sells, among other things, LensView, a lens design package with a database of around 30,000 prescriptions, commented that it was, but only over an 8 mm field ...

Cheers,

Dan

[rdenney]

What happens if the minimum f-number is greater than the optimal? Use the optimal, because using a greater f-number will make the image less sharp. When this happens, you can't get the desired DoF, and you also may be getting sharpness at the DoF limits at the expense of that in the plane of focus, and may need to decide which is more important.

I submit that for art photos of three-dimensional subjects, the balance between acceptable depth of field and acceptable diffraction is as much an artistic decision as a photographic decision. Outside of highly sophisticated techniques such as the scanning method mentioned above, which to me seems to be reserved for science projects rather than for artistic work (and recognizing that just occasionally those two objectives do overlap), the question for macro work isn't how to get rid of the out-of-focus stuff, but how to use it to support the artistic intent of the photograph.

My point was that it is not a case where we have three zones in the solution space, where the first zone includes pictures with too much that is out of focus, the second zone provides both acceptable subject focus and acceptable diffraction, and the third zone provides unacceptable diffraction. We have to set standards for the final product rather arbitrarily to be able to draw those boundaries.

Rather, the range of subject that appears to be out of focus smoothly decreases as the lens is stopped down, and the diffraction smoothly decreases as the lens is opened up. Superimposing those gradations will result in a U-shaped curve with a low point where both diffraction and insufficient focus are mutually minimal. But for the artistic purpose of the photographer, that spot may not provide an acceptable performance either in depth of field or in diffraction. There may be no solution space.

This seems to me rather easier to assess on the ground glass than with calculations, unless in the domain of a science project where there is complete control over (and geometric understanding of) the subject.

A little experience will allow the photographer to correlate how the image looks on the ground glass under a loupe with the performance they will see on a print of a given size. It may take a more powerful loupe than we normally use for focus and composition.

Rick "an engineer who calculates lots of things but often prefers an empirical approach" Denney

[Jeff Conrad]

Rick, I agree that DoF, like composition and many other things, is an artistic decision in any photograph. A good example is the one I mentioned from Lefkowitz: which is the better of the two images? The one you prefer, of course.

"Look at the groundglass" hardly ever seems bad advice, but were it always that simple, there would have been no need for the original question.

The point of my comment that you quoted was that it's easy to determine a maximum f-number for given conditions. The maximum f-number isn't necessarily the right one, but there's never any reason to use a greater one unless you want to decrease DoF and decrease overall image sharpness. The OP had indicated he was interested in maximum DoF and was using f/64; a quick calculation suggested that if his conditions were similar (i.e., 10 mm focus spread), with 4x5, there wasn't any benefit in going beyond f/32. To be honest, I usually determine settings for close-up work visually, but I sometimes find a calculation handy as a sanity check.

The situation is a bit different in general photography, where at least for what I do, diffraction is almost never an issue, so I almost never think about it. I usually try to find the minimum f-number for the DoF I want and hope I can get a short enough exposure to avoid motion blur.

I think whether one calculates settings or gets them strictly from observing the groundglass is a matter of personal preference. I've always been much better at determining the near and far points of focus that at determining what is and what it not sharp with the lens stopped down. If nothing else, I can get the settings from the focus spread a lot faster than I can by observation. This happens to be my approach. It's certainly not the only one and isn't necessarily the best. Now I usually want everything sharp, so there arguably isn't much art involved in choosing focus and f-number, and rote procedure may be more appropriate than it would be in the case of selective focus.

[Dan Fromm]

Jeff, not to be a complete idiot or anything, but at 1:1 focusing wide open (f/5.6 set, f/11 effective) I find the near and far limits of "in focus" indistinguishable. How do you do it? Please don't tell me to look at the GG, that's what I do now.

Cheers,

Dan

[Lynn Jones]

Stop reading books and start testing! I've tested several hundred lenses on film, which as a photographer, would make sense.

In reality, diameter of aperture is the real test, provided the lens is good, and that Rodenstock should certainly be good. Start to get suspicious when apertures at about 1/8" or about 3mm. Below that, some lenses start to get a bit less sharp, however in LF with good lenses 2mm of aperture is where you can see the difference. I often shoot with 90mm super wides, both Schneider and Ilex 6 element lenses at f 45. With the contrast of chrome films the results are fully commercially viable. Contrast is the name of the game, chrome GBar is 2.7 times negative GBar. The Germans going back to the 1800's used optical contrast as their criteria for acceptability, EG: Zeiss Contax lenses and others.

There are virtually no Dioptrics (refractive) optics that are Diff. Limited.

Some Catoptrics such as a schmidt cameras approach and possibly achieve DL performace.

Catadiptrics can be very good but aren't DL.

Lynn

[Jeff Conrad]

Dan, I doubt that I can focus any better than you can. I wonder if we're talking about the same thing here. I focus on the most distant part of what I want sharp, then focus on the nearest part of what I want sharp, and get the focus spread from the difference in the two image distances.

I certainly can't focus to submicron precision. Under the conditions we're talking about, I'd guess my depth of focus at either the near or far point is at least a millimeter, and it probably varies with what I'm looking at. But when the two points are 10 mm apart, I usually can tell which one I'm focused on. I don't always get the same focus spread--it might vary by a millimeter or even more. If this happens, I just use the average of several tries.

And with some subjects and bad lighting, I may have a hard time getting consistent values. As I mentioned, I usually make close-up settings visually because I'm usually more interested in the appearance of the out-of-focus background than the exact extent of the DoF. This probably is fortunate, because there would likely be times when I'd have a hard time getting an accurate focus spread. The operative word for using the calculation as a sanity check was "sometimes", and I guess I gave the impression that this is more frequent than it actually is.

Most of the time, I don't work at 1:1, so the image is brighter and the depth of focus is smaller, and it's usually fairly easy to get the focus spread. I don't suggest that I always get it perfect, but simply that I usually get it a lot worse if I try to assess sharpness with the lens stopped down to taking aperture.

[Dan Fromm]

Jeff, I believe that you focus by moving a standard. Front, rear, it makes no difference. What you're doing is changing magnification. Standard practice when shooting closeup is to set magnification and to move the plane of best focus by moving the camera/lens assembly as a unit. That's why we use focusing rails. Put the plane of best focus where desired and let DoF be what it will.

I suggested that the OP buy and read Lou Gibson's little pamphlet because Gibson's illustrations (photographs, not drawings) show DoF decreasing on stopping down. Its been a while since I looked in my copy of Lefkowitz, but I don't recall that he showed the effect. It is real. That's what's scary about Gibson; he makes what can't be done very concrete.

Good practice at normal distances is often not good practice closeup.

Cheers,

Dan

[Lynn Jones]

I see they eliminated my response on this subject which is too bad because it was valid, disagreeing with all the above. My comment was based on thousands of LF lenses from 4 different optics manufacturers.

Lynn

[GPS]

I see they eliminated my response on this subject which is too bad because it was valid, disagreeing with all the above. My comment was based on thousands of LF lenses from 4 different optics manufacturers.

Lynn

It's still there, valid as before. :)

[Jeff Conrad]

Once the object distance is set, perhaps by moving the entire camera on the rail and fine tuning with the front standard, focusing with the rear standard doesn't change magnification, because the object distance doesn't change. So it shouldn't affect determining the f-number.

The images on p. 53 and pp. 62-63 of Gibson are similar to those on p. 84 of Lefkowitz, showing that past a certain point, overall sharpness decreases. When the blurring of an object becomes visible, because of either defocus or diffraction, that object is by definition outside the DoF. When this happens for the entire image (i.e., because of diffraction), the DoF is zero (Gibson actually talks of "depth of detail," which is DoF that takes diffraction into account). And the transition to zero DoF can occur quite rapidly; on p. 62, Gibson shows zero DoF occurring at less than a stop beyond optimal f-number.

I completely agree that close-up work often calls for different technique, and Gibson shows that in some cases, you can't get there from here.

But we should note that the examples cited from Gibson and Lefkowitz use considerable magnification (those on pp. 62-63 of Gibson are at 4:1). It's a very different situation at the OP's 1:1, as Gibson indicates in Figure II-80 on p. 62, and things at 1:1 should be very manageable. Depending on the required DoF, the OP might get better results by opening up a stop or two. But it's also possible that for enlargement to 8x10, losses from diffraction aren't noticeable, and the difference won't be noticeable, either. The smaller f-number would of course allow a shorter exposure time and reduce the chance of motion blur, and might allow greater enlargement if that were desired at some time in the future. The only real way to tell, of course, would be to try several different f-numbers and see if there's any difference in sharpness.

[Dan Fromm]

Well quoted, Jeff, but not satisfactory.

I do most of my closeup work with a Nikon because its so much easier than with a Graphic. Faster, too. I use flash to eliminate motion blur. With a 105/2.8 MicroNikkor AIS, at 1:2 my flash rig lets me choose between f/16 with the lens on its extension tube and f/22 with the lens fully extended and no tube. A shot at f/22 won't print as large satisfactorily as the same shot taken at f/16. The problem isn't the lens' CRC, which affects off-axis aberrations, it is diffraction. If we want to enlarge more than around 5-6 x, diffraction is a killer at relatively low magnifications and relatively large apertures. Contact printing changes the game, but 35 mm contacts lack something.

When I was trying to decide which macro lenses to use on my Graphics, I tested to find each candidate lens optimal range of magnifications and aperture. Short answer, all of the first-class macro lenses I tried -- all 5 Luminars, 63 Macro Nikkor, 100/6.3 Neupolar, a heap of Mikrotars -- were best wide open. End of that discussion, they are all diffraction limited. So is the 4"/5.6 Enlarging Pro Raptar that I actually use.

This was the case for the 100 Luminar, 100 Neupolar, and 90 Mikrotar at all magnifications tried from 1:8 to 4:1, also for the 4" EPR. I mention these lenses because in that range all cover 4x5 and because the range includes magnifications we're interested in.

I think we've all neglected to mention enlargement's role in print sharpness. The bigger the enlargement, the better the negative has to be. This is why my flash rig designer not only calculates f-stop required for good exposure given magnification, the flash's power, and the rig's geometry. It also calculates DoF and maximum enlargement given magnification, aperture set, and desired print quality (another interpretation of the dread circle of confusion). Its all very discouraging.

Cheers,

Dan

The designer is embodied in a spreadsheet. If you'd like, I can e-mail a copy to you.

[Jeff Conrad]

Dan, I don't think I really disagree with anything you've said here. My experience is similar, though I haven't played with nearly as many lenses.

But I'm not quite sure how this relates to the OP's question. I really read it to be "What is the optimal f-number in the plane of focus?", and several people gave pretty succinct answers; I may have added gratuitous complexity by throwing in the DoF limits. Even with that added complexity, though, the OP's situation would seem manageable. If, for the sake of argument, we assume that 4x5 is 4x the size of 35 mm, the usable f-number for 4x5 at 1:1 should be twice that of 35 mm at 1:2, so f/32 should work. And the OP indicated that f/64 looked OK, so it still would seem worth comparisons at perhaps f/22, f/32, and f/45.

I'd certainly be interested in looking at the spreadsheet, if for no other reason than to see how it compares with Gibson. And it could reveal something self evident that I've completely missed ...

[Jeff Conrad]

Umm ... on rereading the OP's question, I think he did ask for the optimal f-number, so I hope I at least made a reasonable stab at providing an answer.

[Dan Fromm]

Hmm. Yes, the thread has drifted a bit.

Jeff, as I read the original post, the OP wanted to know what to do. I'm not sure we can tell him that, even with all the information he gave, but he did mention shooting 4x5 and 5x7 at 1:1 and wanting to enlarge to 16x20 to 32x40.

Starting from 4x5 that's roughly 4x or 8x. If we accept that he needs at least 8 lp/mm in the print, then for 16x20 he needs at least 32 lp/mm in the negative and for 32x40 he needs 64 lp/mm. For 16x20, he can't shoot at an effective aperture smaller than f/46; at 1:1 this means he can't set an aperture smaller than f/22. For 32x40 he can't shoot at an effective aperture smaller than f/23; at 1:1, this limits him to around f/11 set.

I'll let the OP do the arithmetic for 5x7.

Ain't much DoF at 1:1 at those apertures, so DoF considerations are moot. That's the joy of working closeup. We often can't accomplish what we'd like to.

Cheers,

Dan

[DolphinDan]

Thanks to everyone for responding. I understand the relationship between DoF and diffraction for macro lenses better now. I will try f16, f22, and f32, and compare the results to my older photos.

Dan: I will crunch the numbers for 5x7. Now that you have laid it out clearly I can handle it :-)

I still have one question which maybe you all answered by implication: do macro lenses have a different diffraction limit than "regular" (i.e., optimized for inifinity focus) lenses? If f22 is the optimal aperture, then the answer appears to be no.

Once again, thanks to everybody for answering my questions.

Namaste
Daniel

[Dan Fromm]

Dan'l, a lens is a lens is a lens. There's nothing magic about macro lenses.

Macro lenses differ from lenses made for shooting at distance only in how they're optimized. A lens can be optimized for only one pair of conjugates (film to rear node distance, front node to subject distance). Macro lenses are optimized for film far from the lens, subject close to the lens. Lenses made to be used at distance are optimized for film close to the lens, subject far from it.

Yours for solidarity among Daniels,

Dan

[willwilson]

All these numbers and math are fine, but really its about getting the shot. I can't think of a time when I was shooting 1:1 or higher mag with 4x5 and the subject didn't demand a certain aperture. There's not a lot of options with lf and macro. You just gotta burn some film at a few different apertures or stop down an extra stop and deal with a little extra diffraction.

What are you guys shooting at f22 and below with 4x5 at 1:1? It would have to be really flat.

[Dan Fromm]

Will, I shoot mainly flowers on 2x3 at magnifications up to 2:1. I try to stay at 1:1 or below, don't always succeed. I'd like to have sharpness in depth, have had to accept that I can't.

These days I shoot mainly at f/16 set, effective as small as f/45 or so. This limits enlargement considerably, but seems the best compromise. With my flash rig f/22 set is possible, but having tried it I'd rather not ... Others may choose otherwise.

I've learned from some of my fuzzier 6x6 shots at distance -- my Perkeo II's Color Skopar is in collimation, just isn't that sharp a lens -- that sharpness isn't all. Good exposure and composition can sometimes, not always, overcome softness. Its made me wonder why we care so much about which lens/aperture/... is best.

Cheers,

Dan

[GPS]

...
I still have one question which maybe you all answered by implication: do macro lenses have a different diffraction limit than "regular" (i.e., optimized for inifinity focus) lenses? If f22 is the optimal aperture, then the answer appears to be no.

Once again, thanks to everybody for answering my questions.

Namaste
Daniel


Dan'l, a lens is a lens is a lens. There's nothing magic about macro lenses.

...Dan

I'm glad, Dan, you realized that diffraction in a lens depends on its aperture only (for a given wavelength) and not on the fact of being constructed of 4 elements or more...
Diffraction is diffraction is diffraction...:)

[Jeff Conrad]

I don't think Emmanuel meant to suggest that lens design had any effect on diffraction, but simply that most modern 6/4 designs are sufficiently corrected by f/22 that they're essentially "diffraction limited," so that further stopping down won't improve sharpness in the plane of focus. Sharpness at the DoF limits is of course another matter.

I think we sometimes make all this seem more complicated than it really is.

[GPS]

... Sharpness at the DoF limits is of course another matter.

I think we sometimes make all this seem more complicated than it really is.

In that I agree with you, Jeff.

[Emmanuel BIGLER]

I don't think Emmanuel meant to suggest that lens design had any effect on diffraction,

Yes, Jeff, and sorry if I was not clear.
What I meant is that the best f-stop for which diffraction effects, if they could be separated from aberrations, would contribute "to an equal part" with respect to residual aberrations, DOES depend on the lens design.
We should also mention that between two lenses of the same focal length, stopped-down to the same f-stop, at the edge of the field, the wider the field is, the larger are diffraction spots, even in an ideally perfect diffraction-limited lens. But for the same position of the field, of course ultimate diffraction effects, if they could exist alone without aberrations, do not depend on the lens design (well, at least if we do not compare retrofocus lenses, where strange sthigs happen, with quasi-symmetrical lenses, this is another story)

However, my conjecture is that between a 6/4 macro lens and a classical 4/4 of the same focal length in the genre of the Apo Ronar, used in the same conditions, the gain for the 6/4 is probably in a wider field, not necessarily a wider best f-stop.

I say that because I have handy the full & official MTF curves of the apo-ronar at f/22, up to the diffraction cut-off, and at the centre of the field, the MTF curve of the real lens @f/22 is almost indistinguishable from the theoretical limit. At least as published in Rodenstock's specs, catalogue of 1986, where MTF curves are probably computed and not measured with a MTF bench.
Hence, nobody can design a lens which is substantially better at the centre of the field in the same conditions (same focal length, same magnification, same f-number), since the absolute limit is already reached.

For "digital" view camera lenses of late-style design, my understanding is that they are better corrected than previous "film" designs of the same focal length, at the expense of a narrower angular field, and the best f-stop is wider. I should say : HAS to be...

[dh003i]

It is also worth nothing that many large-format lenses are sharper even at f/11 and f/16 than at f/22. That isn't a bad thing, if the lens is already nearly diffraction-limited at f/22. It means that he lens is also very good at f/11 and f/16, and doesn't have enormous problems there. My Nikkor 90/4.5 probably has similar resolution to the 90/8, which is 73, 65, and 56 lp/mm at f11, f16, and f22, respectively.

[Bosaiya]

This is all headache-inducing and ultimately hopefully meaningless, it is very subjective. What is "good"? How high is up? Will raw numbers actually give you an answer, and if so, is that a good thing? Try a lens and see if it works for you.

If you find yourself in the Seattle-area stop by and try out my Schneider 180mm Macro and a bunch of others and decide for yourself based on the photographic results.

[dh003i]

Jeffs' additional comments on the added effects of defocus has reminded me of the OPs' original desire to maximize the DoF for macro work. I think it has been mentioned here, maybe even by me, that there is a technique to obtain infinite depth of focus with a macro setup by using a scanning technique. This is variously called Scanning Light Photography as well as other terminology. It's been around for a long time.

The idea is to illuminate the plane of focus, which would include the defocus regions to an acceptable circle of confusion, using two or three slits of light that are of a thickness approximating the thickness of the plane of acceptable focus and intercept that plane exactly and be coplanar with that plane throughout the field of view. Then the subject is translated through the slit illumination at a uniform rate and a velocity
that is consistent with a proper exposure dose. (Not so easy to explain briefly).

If interested Google C. J. Kazilek, "Scanning Light Photography" for a description with images and references.

It needs to be a pretty sophisticated setup but I used it some years ago for documenting some MMIC modules with considerable success. If the OP is really serious about great depth of field this cannot be beat.

A unique feature is that the image obtained has no perspective distortion since every part of the subject is always the same distance from the camera.

Nate Potter, Austin TX.


I wonder if this setup could also be used for macro-photography if you were willing to drop a *lot* of cash and work at night. You have the slitted side-lights and the camera setup, with a motor. But instead of the sample moving -- maybe it is a tree -- you couple the slitted side-lights to the camera and move the side-lights along with the camera. Not too practical.

Actually, I'm pretty sure that here focus-bracketing, scanning, and merging with the Tufuse or Enfuse program would be much more cost-effective.