Macro and diffraction - A test that has surprised me

[Stagecoach]

In my never ending quest to learn more in the field of macro and keep myself amused as my TVs still bust after 8 months a test tonight to determine the results of doubling the DOF by increasing aperture well above what I would normally use which is f16/f22.

I have always been made aware by reading of the increased diffraction at small apertures, have seen review controlled tests showing crops of the effect but never really looked at this in the real world.

So, the criteria for my test was to simulate a typical critter pic as per the shield bug I posted a few nights ago and to do this I used a match 15mm long (standard bug length) to simulate approx 3/4 length coverage in the frame and some green tomatoes with hairy stems, the hairs simulating fine hairs on a bug.

The object was to try and get an idea of image quality loss if I was only producing images for web viewing at 1100/1200 pixel width.

Equipment was D300 with 105VR and two SBR-200s attached to the lens front both in TTL mode. Tripod, remote and mirror up. Manual settings and focus.

Camera (film plane) to subject distance was almost the closest possible at 300mm.

From DOF calculator - 300mm @ f16 DOF = 3.2mm. 300mm @f40 DOF = 8.1mm

These Jpegs are SOOC at 'standard settings', opened in Elements 7, reduced to 1100 pixel width and then using unsharp mask at 100% 0.3 pixels on each. No other adjustments were made although I suspect there is a slight drop in exposure in the f40 one due to having the diffusers in place.

The focus point was the bright hair at the centre which I appear to have missed

f16, ISO400, 1/60th


f40, ISO400, 1/60th


I ran tests through to min aperature of f57

Any comments especially with respect to where any degradation in IQ between the two can be seen will be appreciated as always.

[Andrew76]

Cool experiment! The only real difference I can see at first glance (beyond the obvious increase in depth of field ), is that the tomato in the background of the second one appears to be much 'noisier'. Or is it that it's just more in focus?

Other than that, I can't see any degradation in the subject matter.

[DanK]

Grahame,

Fascinating. Thanks for posting this. I had learned that the effects of diffraction are greatly overstated, and I have found that I can print macros shot at f/22 at 8 x 10 (~ 20 x 25 cm) with no noticeable degradation. However, I generally never shoot much above that. Your test has prompted me to do some more tests. I'd like to look at higher resolutions. A small image on a computer screen is low resolution and therefore forgiving. Prints are more demanding.

Dan

[RustBeltRaw]

It'll be hard to judge a loss of detail when the images have been downsized this much. That may be the answer you're looking for. Personally, I try to stay under f11 (my 60D's diffraction limit) if that's sufficient to my needs. If it's not, then I stop down the aperture and don't sweat it. f22 creative effects are too delicious to worry about frequently-invisible image quality losses. On the other hand, I rarely need anything above f8.

That being said, the details near the focal point in your f40 test shot do appear slightly less crisp. What setup are you using to get that high? Is it a macro lens on extension tubes?

[HaseebM]

Only thing I see is a small dust mote and perhaps slightly less crisp as Lex mentioned in your second image. Which begs the question, are 'dust motes or specks' a done thing when shooting at more than f/11?

[jcuknz]

As one who started his photographic learning process shooting at f/64 I have often wondered at this diffraction business becuase I have never struck it to date and took/take heart from Colin Southern's comment on the subject that other aspects would be more likely to impinge of IQ before diffraction. There was also the f/64 Group of noted photographers in the past ... is it becuase they, like me in my early stages, were working with LF rather than the small toys we have today?

That is a genuine question becuase it is something which concerns me that for all the houhaa I have never struck it to recognise it.

With regard to TV ... I switched it off just prior to the Rugby World Cup in NZ, is that near two or nearly three years ago? .... and have only used it to amuse somebody spending the night with me once and I didn't watch it ... and now 'they' have stopped transmitting pictures that it can receive .... so it is back to its main purpose ... a monitor for video I have shot .... and it must be near a decade since I shot any of that

[rpcrowe]

It is really fun to satisfy my curiosity about my equipment. I played with testing my 300mm f/4L IS lens with a 1.4x Canon MkI teleconverter. Although my testing was not scientifically accurate, I satisfied myself that the combination was quite useful. But, the smallest aperture I have used on my Tamron macro is f/16 - because I will most often hand hold the lens and want the extra shutter speed available. However, the next time I work with my Tamron Macro lens on a tripod, I will shoot at f/32 to satisfy my curiosity.

I have a 25mm extension tube and I want to play with my Tamron using the extension tube + 1.4x TC.

I didn't know that your Nikon micro lens had an f/40 aperture capability. I thought that f/32 was the minimum aperture, as with my 90mm Tamron f/2.8 Macro. Were you using an extender?

BTW: When I want to find specifications for any piece of photo equipment; I usually go to B&H Photo and Video website. Choose the equipment and click on "specifications". It was easier to find the specs for this lens on the B&H site than looking on Nikon's website or the various reviews on the internet. The reviews I looked on always stated the widest aperture but I couldn't easily find the smallest aperture.

Anyway, I have a hunch that the longer the focal length, the less diffraction you experience. The members of the "f/64 Club" usually shot with large format cameras and very long focal length lenses. Even wide angle lenses for 8x10 inch or larger format cameras are quite long in focal length.

[Stagecoach]

Hi all,

Thanks for the comments and observations and I'll respond to them in order;

Andrew, the noise appearance you mention I believe to be the actual colouring of the tomato which I have looked at with a magnifying glass today. My tomato growing is on a par with my photography, I'm still learning I can not think of a reason the noise would have been noticeable as I never have a concern at ISO 400 unless cropping or increasing exposure in PP which I have not done in this case.

Dan, agree as you say that images of this size are more forgiving and that's why I limited my criteria to this size for now. The purpose and size of the final image is going to determine the limits of individual acceptability eventually.

Lex, you mention that you personally try and stay under f11, are you referring to extremely close up macro work ? For comparison purposes between the two although I had manually focused on the bright hair in the centre I selected the sharpest hair I could find in the f16 and compared to the same hair in the f40 which in theory would have been in 'better' focus.

As for crispness between the two at the focal point, or any point of sharpness I also wonder how much our perception is affected at this resolution and greater by the sharpness or blur of the surface behind. That centre bright hair being a good example of this.

The setup for this test was the Nikon 105VR 2.8 macro, no filter attached and no tubes. Image taken at almost min focusing distance and measured with tape measure as near as possible at 300mm subject to sensor.

Haseeb, yes that black dot is something on my sensor and I have a few more less noticeable ones. They will become more of a concern as aperture is reduced and the significance is very much going to depend upon the subject surface. Unfortunately I can not cure them with a dry sensor clean and living on an island where I have no access to cleaning materials or servicing its a problem I just have to live with.

Jcuknz, as you can see from the gist of my test my intention was not to dispel technical theory which I in no way disagree with but to look at a specifically defined real world scenario within the realm of macro with the aim of seeing the affects.

As for the TV my decision not to replace it was based partly on your countries insistence in exporting 'Shortland Street' to us

So, what have I learnt ? I am now far more confident that I would be able to push aperture to f40 if needed to either increase overall DOF, OR, to give me a greater chance of getting that critical area such as a bugs eye in focus. Yes there is a degradation in IQ but within the limits I set I believe it is acceptable. In addition there is also the opportunity to compensate in PP of which I did not do for the test.

One other interesting observation is that in the set I took at ISO 200 there was a clear decrease in exposure as aperture was reduced above a certain setting which I put down solely to the limited power of the flashes. In all cases this did not require more than 1.5EV to correct. In the ISO400 set it looks to me as if the f40 image is brighter than the f16 ? The flashes were fully recharged before the f16 shot, a mystery.

[Stagecoach]

Richard, you make a very interesting observation regarding the minimum aperture.

The f No's indicated on the camera and within the exif are the same right up to f57. No extender was used.

I also had a look at the specs online and where minimum aperture is given it is always f32.

So what it actually is I have absolutely no idea but there has to be a logical reason somewhere and hopefully someone can explain. But then again it's irrelevant I suppose and just for fun here's the same image taken during the tests that is at the minimum aperture 'indicated' and the only difference in its post processing is that I have had to put a 1.15EV correction on the RAW. Also noticed that the RAW appeared brighter than the jpeg SOOC but really picking at hairs here.

Exif and camera indicated ISO200, f57. 1/60th ..............



Aperature indication reason ..........................

A search on the web did not resolve this anomaly so I had a play to see exactly what was happening.

At maximum magnification 1:1 ratio giving minimum focus distance 0.314 mtr aperture is adjustable up to f57, this can clearly be seen by the reducing light level and difference in DOF. As focus is adjusted giving a lower than 1:1 magnification ratio and increased minimum camera subject distance indicated aperture reduces to a minimum of f32.

[Letrow]

This is interesting Grahame and shows me two things. One, diffraction is a bit overstated, which is good to know and two, the DOF is definitely better, which would benefit some of my insect shots.
My only problem will be available light. I mostly shoot handheld at F/11 and 1/200 or 1/250 and vary the ISO according to need. If I adjust the aperture ISO will surely suffer and that might affect the end result more than the diffraction I guess.

[Stagecoach]

Peter,

I fully agree with your comments. The chances of being able to push the aperture to these settings is pretty unlikely with available light when maintaining a reasonable hand held speed and I also suspect the noise from the increased ISO is going to be greater than diffraction affect.

There's always the possibility of flash to assist the natural light when hand holding but my attempts to date have never used smaller apertures than around f16. On no, I feel another test coming on ........................

[DanK]

Grahame,

The f No's indicated on the camera and within the exif are the same right up to f57. No extender was used.

I also had a look at the specs online and where minimum aperture is given it is always f32.

So what it actually is I have absolutely no idea but there has to be a logical reason somewhere

Indeed there is an answer, and this helps explain your post, if I understand how Nikons work with a macro lens. (I shoot Canon.) When you get close to minimum working distance, the effective f-stop increases. Canon cameras show the physical f-stop, but I believe Nikons show the effective f-stop. The explanation is on on the macro lens tutorial on this site:

The reason that the f-stop changes is because this actually depends on the lens's focal length. An f-stop is defined as the ratio of the focal length to aperture diameter. A 100 mm lens with an aperture diameter of 25 mm will have an f-stop value of f/4, for example. In the case of a macro lens, the f-stop increases because the effective focal length increases — not because of any change in the aperture itself (which remains at the same diameter regardless of magnification).

Again, I am not certain that Nikons display the effective f-stop, but that would explain your seeing values to f/57 with a lens that appears to be limited to <= f/32. That is also a difference of roughly 2 stops.

If so, your f/40 (effective) was probably a physical f-stop about 2 stops less, or f/20. That would resolve the mystery, at least for me. I had discovered diffraction in macros at (physical) f/20 is not very severe, but my impression, from a few shots without careful testing, it that it was very visible by f/32.

EDIT: scratch the last paragraph. It suddenly struck me that diffraction is probably a function of effective, not physical, f/stop, since diffraction is a function of f/stop, not physical aperture, regardless of focal length. So, I went back to the tutorials and found this:

The most important consequence is that the lens's effective f-stop increases*. This has all the usual characteristics, including an increase in the depth of field, a longer exposure time and a greater susceptibility to diffraction. In fact, the only reason "effective" is even used is because many cameras still show the uncompensated f-stop setting (as it would appear at low magnification). In all other respects though, the f-stop really has changed.

Dan

[RustBeltRaw]

Lex, you mention that you personally try and stay under f11, are you referring to extremely close up macro work ?

No. I'm a little too impatient for macro work. But diffraction affects all cameras at some point, especially when they have small photosites. My 60D's not terrible, and as I said, I don't usually get above f8, but I still worked out the diffraction limit in order to be mindful of the camera's limitations.

[PhotomanJohn]

Again, I am not certain that Nikons display the effective f-stop,

Dan - You are correct, they do.

John

[Stagecoach]

Dan,

Thank you for your explanations.

Having read them something suddenly 'clicked' and I recalled the often written about concerns when it first came out that this macro lens (as some others) also exhibits this at the other end of the scale in that it only achieves and indicates it's f2.8 capability at max focusing distance. Many owners wondered if there was a problem with their copy.

In reality I suppose with practical photography the 'numbers' are simply a scale between the limits of adjustment and serve more as a reference to a setting within that scale that has been used.

[jcuknz]

Grahame ... I am not dispuiting the facts of the situation just suggesting it is not such a serious matter for most photography.
I gather your clever Nikon camera works out the effective aperture for you when I thought perhaps you had been using waterhouse stops for the experiment
I don't think anybody has answered my first question so I'll ask another. Assuming one has used a waterhouse stop placed either behind or in front of the lens what is the effect of doing this...I am aware that if it is too far away in front one gets a vignetting result. It is academic to me now but I did once organise them for an enlarger lens in the past, and a camera lens even further back in time without knowing the pros and cons just it was possible.

[rpcrowe]

Very interesting regarding the f/57 from an f/32 lens. I guess that would mean that as the lens is being focused closer, the focal length is increasing with the aperture size remaining the same. I wonder what the effective aperture would be if one was shooting at a 1:1 image ration at the maximum f/2.8 aperture.

Does the aperture of your Nikon lens read f/32 or f/57 in the viewfinder when you are shooting at a 1:1 ratio?

Tomorrow, I will try this with my 90mm Tamron Macro on a Canon 7D...

[Stagecoach]

Richard,

Yes, the focal length does change as focus (magnification ratio) changes. I have no idea what the focal length changes to from it's 105 mm but I suppose this could be calculated.

From what I have learnt Nikon bodies 'indicate' the effective aperture above f32 whereas Canon do not although both cameras and lenses are doing exactly the same thing. The only difference I can think that this would make practically is that the true effective aperture indicated could be used to determine the exact DOF from a chart but in the real world who's bothered about this and who works macro at these f stops ?

My Nikon, D300, does adjust and read up to f57 in the viewfinder when set at 1:1 ratio and can be adjusted down to f4.8

[DanK]

Tomorrow, I will try this with my 90mm Tamron Macro on a Canon 7D

AFAIK, all Canons show the actual, not effective, f/stop.

In practice, I think it makes very little difference unless you are using external metering. I think the Nikon approach is more informative, since the effective f/stop better describes how the lens is actually performing. However, as long as you know what f/stop range you find acceptable given the way your camera displays the information, I don't think it matters much.

According to the relevant page in the tutorials section, the difference at 1:1 should be roughly a factor of 2 regardless of the lens. It does not explain how quickly the two metric converge as you move out from 1:1.

[Arlen]

A good discussion, one that I missed when it first came around.

As has been noted above, whether or not the magnitude of resolution loss due to diffraction at higher f-stops is something to be concerned about depends on your output mode. If your output is never more demanding than a 1200px-wide web jpeg, you may not see much of a difference at f/32 or f/40 from shots at lower f-numbers. But in more demanding situations, like large prints, it may be noticeable. Not necessarily a deal breaker, especially after post-processing, but worth keeping in mind and deciding what factors are most important.

Unlike the lucky Nikonians, we Canon macro shooters have to remember that at close to 1:1, the effective f-stop is substantially higher than the camera reports. For my Canon 100mm f/2.8 USM Macro lens that is about 2 stops; if the reported value is f/22, the effective f-stop is really about f/45. Annoying that we have to remember that the camera is lying to us.

The situation gets worse when we add extension tubes to the macro lens, and here it seems that even Nikon owners don't get a break. From what I've read, Nikons are like Canons in that they don't report the increased effective f-stop that results from extensions tubes. For my 100mm macro lens, at 1:1 focus and set to a nominal f/16, the effective f-stop is f/32. Add a 12mm tube and that goes to f/37. With a 24mm tube the effective f-stop is f/40. With a 36mm tube it becomes f/47. Add all three tubes together, for a total of 68mm extension, and we are at an effective f-stop of f/60. The resolution loss at these high effective apertures is clearly apparent in the raw file and in even moderate-size prints (8x10).

So the take-home message for me is that if pushed too far, which can easily happen with extension tubes without realizing it, high f-stop diffraction effects can become large enough to matter.