Experimenting with very small effective apertures

Messages
5,691
Name
Nick
Edit My Images
Yes
I am currently doing an exercise to explore the use of very small effective apertures to get greater depth of field than I can currently get with my close-up lens setups, with which I use f/45 full frame equivalent when photographing insects, spiders and other invertebrates. The eight images below used effective apertures varying between f/32 or so to f/132 full frame equivalent.

The first four images were captured with a Venus Optics KX800 twin flash mounted on a full frame Sony A7ii mirrorless camera, using a Laowa 25 mm macro lens set at f/22 to give effective apertures around f/72 to f/100 for the magnifications of 2.5X to 3.5X that I was using. Images 5,6 and 7 were captured with a Yongnuo twin flash front-mounted on a Sigma 105 mm macro lens with a 2X teleconverter on a Sony A7ii, with the aperture set to f/45 to give effective apertures up to f/132 at 2X magnification. (Image #7 used effective f/132. Images 5 and 6 probably used slightly larger apertures). Image 8 was captured with the Yongnuo flash front-mounted on an Canon MPE-65 on a Canon 70D APS-C dSLR with the lens set to f/16, to give an effective aperture no larger than 1/32, and probably much smaller. (As with images 5 and 6, I don't recall what magnification I used for image 8, so I can't work out the effective aperture for any of these images.)

The thinking behind the exercise and some of the issues involved in this approach are described in this post in my Journey Thread.

#1 These flies were quite small, I think around 3mm from head to the tip of the wings when folded back along the body. Some animals take their time when mating, but not these. It was quick, and they were moving around as well. I think I was rather lucky with this one. I didn't have time to reduce the magnification so I could get the subjects fully inside the frame.

1647 1 1642 13 2020_06_02 A7ii+Laowa25 DSC02305_PLab3 SP9LR 1300h-DNAIc-DNAI-PS-AISh
by gardenersassistant, on Flickr

#2 These long-legged flies were hyperactive, generally settling for only a few seconds. You have to be quick with these. The magnification and flash power (the KX800 is a manual flash) were set for the small flies and so it was tricky when one of these larger flies turned up briefly to get it completely inside the frame.

1647 2 1645 19 2020_06_01-05 A7ii+Laowa25 DSC02444_PLab3 SP9LR 1300h-DNAIc
by gardenersassistant, on Flickr

#3 The working distance of the Laowa 25mm macro is only 40mm or so. I was surprised that such a large subject as this hoverfly let me get that close, especially as they are quite jumpy.

1647 3 1642 16 2020_06_02 A7ii+Laowa25 DSC02332_PLab3 SP9LR 1300h-DNAIc-AISh2
by gardenersassistant, on Flickr

#4 This was another of the small flies. However, since it was grooming it stayed in one place for longer, making it possible to have a number of attempts.

1647 4 1645 23 2020_06_01-05 A7ii+Laowa25 DSC02451_PLab3 SP9LR 1300h-2DNAI-AIS
by gardenersassistant, on Flickr

#5

1647 5 1644 01 2020_06_07 A7ii+2X+105 DSC02819_PLab3 SP9LR 1300hDNAI-AIS
by gardenersassistant, on Flickr

#6 This Wasp beetle was in motion and quickly disappeared back into the foliage.

1647 6 1644 10 2020_06_07 A7ii+2X+105 DSC02835_PLab3 SP9LR 1300hDNAI-AIS
by gardenersassistant, on Flickr

#7 This Cucumber spider was moving around as it worked on its prey.

1647 7 1644 15 2020_06_07 A7ii+2X+105 DSC02874_PLab3 SP9LR 1300hDNAI-AIS
by gardenersassistant, on Flickr


#8

1647 8 1645 04 2020_06_01-05 70D+MPE-65 F16 IMG_8498 ISO100_PLab3 SP9LR 1300h-2-DNAI-AIS
by gardenersassistant, on Flickr
 
Last edited:
Messages
7,148
Name
Steven
Edit My Images
Yes
Interestingly, the effective aperture doesn't affect diffraction; only exposure. I believe that being able to see more of the larger details clearly due to increased DOF more than makes up for the loss of smaller details due to diffraction with macro... often the details are so small that a viewer doesn't even know they exist, so they don't know that the details are missing.
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Incredibly detailed and they look like they were a real technical challenge.
Thanks. The first three felt quite challenging, because of the need to react quickly, especially for the first one where the subjects were rather small, which always makes things more difficult (for me at least). The others felt fairly straightforward.

We never see nature this close normally!
Indeed so. I think that is part of the magic of close-up/macro. We get to see things that are around us but hidden from us.
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Interestingly, the effective aperture doesn't affect diffraction; only exposure. I believe that being able to see more of the larger details clearly due to increased DOF more than makes up for the loss of smaller details due to diffraction with macro... often the details are so small that a viewer doesn't even know they exist, so they don't know that the details are missing.
That's exactly how I feel about it. Well put.

As to the effective aperture not affecting diffraction, we may have to agree to differ about that. My understanding is that it does, but I'm no expert on this though; I tend to look to this article (the DOF/aperture section is probably most relevant in this context) and its author, who is active at dpreview and has helped me come to some sort of (basic) understanding about these things. Enough as I need anyway, which isn't much to be honest.
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
As to the effective aperture not affecting diffraction, we may have to agree to differ about that. My understanding is that it does, but I'm no expert on this though; I tend to look to this article (the DOF/aperture section is probably most relevant in this context) and its author, who is active at dpreview and has helped me come to some sort of (basic) understanding about these things. Enough as I need anyway, which isn't much to be honest.
Thinking some more about this, perhaps I misinterpreted what you wrote. Let me try to explain what I'm thinking. (tbh I'm struggling a bit with this.)

In thinking about this sort of thing I find it helpful to distinguish clearly (or at least try to) as between aperture, as in the diameter of the hole the light is coming through, and f-number, the ratio between the focal length of the lens and the aperture.

Using these definitions, then if you leave the aperture unchanged but change the effective aperture, as happens if you add an extension tube for example, then the amount of diffraction is going to be the same because the amount of diffraction depends on the physical size of the hole the light is coming through, and in this scenario the aperture isn't changing. However, the effect of that same amount of diffraction will be greater because the sensor is further away from the hole the light is coming through and the light will have spread out more by the time it reaches the sensor, magnifying the diffraction (along with magnifying the image of the subject). Does that make sense?
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Wow your kit has evolved a lot since I last properly look on here Nick.

Interesting experimentation. (y)
Hi Bryn. I've not been here so much myself, rather on and off, with quite a bit of off. My photographic journey has been a bit on and off too, but over the years I have accumulated a lot of kit, quite a lot more in the past couple of years in particular. That stems from my being told a couple of years ago that I had 18 months to live, and that rather loosened the purse strings, on the grounds that if I didn't get on with it quickly I would have missed my chance. That was how I came to get a full frame camera for example; I had been wanting to try one for ages but couldn't convince myself it would be justified. Suddenly, it didn't seem to need much justification. Similarly since then with some other kit. As it happens, it looks like it may have been a false alarm (the medical people are conflicted about just what is, or in this particular case, apparently isn't, going on.) I'm fine; fit, and happy. Confined to barracks because of covid-19, as is my wife, but that is ok.

Anyway, I've got lots of different kit now (software, as well as hardware), which gives me the opportunity to indulge in plenteous experimentation, which I very much enjoy.

Up until this latest bout of uncertainty about what kit to use my kit had settled down to being a small sensor bridge camera with close-up lenses for invertebrates and a micro four thirds camera with a macro lens for flowers etc, shooting raw for invertebrates and for single-capture images of flowers etc, and using 6K video for flower stacks (I'm doing very roughly 50/50 as between singles and stacks for flowers).
 
Last edited:
Messages
7,148
Name
Steven
Edit My Images
Yes
Using these definitions, then if you leave the aperture unchanged but change the effective aperture, as happens if you add an extension tube for example, then the amount of diffraction is going to be the same because the amount of diffraction depends on the physical size of the hole the light is coming through, and in this scenario the aperture isn't changing. However, the effect of that same amount of diffraction will be greater because the sensor is further away from the hole the light is coming through and the light will have spread out more by the time it reaches the sensor, magnifying the diffraction (along with magnifying the image of the subject). Does that make sense?
Makes sense to me; the way I understand it is:

At a true f/132 the maximum resolution due to diffraction would be approximately .3MP on a FF sensor (average RGB; regardless of sensor resolution). At f/45 it is ~1MP. 1MP may sound like pathetically little, but 1MP is essentially all that can be contained in an image that is 1024x in size. At f/45 true you are getting the potential of 1MP in sharp focus, and presenting at 1024x makes the most use of that (downsizing the full sensor area).

Put another way, diffraction losses are more about the loss of recorded resolution/details that are w/in the plane of sharp focus... the actual aperture determines this.
Whereas DOF (acceptable focus) has more to do with the relative sharpness of things (primarily outside the plane of sharp focus)... the effective aperture (effective magnification) determines this; as does the display size and viewing distance (apparent magnification).
 
Last edited:
Messages
8,784
Name
Bryn
Edit My Images
Yes
Hi Bryn. I've not been here so much myself, rather on and off, with quite a bit of off. My photographic journey has been a bit on and off too, but over the years I have accumulated a lot of kit, quite a lot more in the past couple of years in particular. That stems from my being told a couple of years ago that I had 18 months to live, and that rather loosened the purse strings, on the grounds that if I didn't get on with it quickly I would have missed my chance. That was how I came to get a full frame camera for example; I had been wanting to try one for ages but couldn't convince myself it would be justified. Suddenly, it didn't seem to need much justification. Similarly since then with some other kit. As it happens, it looks like it may have been a false alarm (the medical people are conflicted about just what is, or in this particular case, apparently isn't, going on.) I'm fine; fit, and happy. Confined to barracks because of covid-19, as is my wife, but that is ok.

Anyway, I've got lots of different kit now (software, as well as hardware), which gives me the opportunity to indulge in plenteous experimentation, which I very much enjoy.

Up until this latest bout of uncertainty about what kit to use my kit had settled down to being a small sensor bridge camera with close-up lenses for invertebrates and a micro four thirds camera with a macro lens for flowers etc, shooting raw for invertebrates and for single-capture images of flowers etc, and using 6K video for flower stacks (I'm doing very roughly 50/50 as between singles and stacks for flowers).
Wow well I hope it all turns out to be a false alarm Nick, must have been worrying but getting the kit must have given you some motivation, your experiments have always been so concise and well executed. Often proving so many misconceptions about macro along the way :D.

I'm rushing 2020 along and hoping to get to my 40th next year quickly as I've been allowed to indulge into a new camera (hopefully the r6 will be released by then). Want to get a mirrorless and full frame :D.
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Makes sense to me; the way I understand it is:

At a true f/132 the maximum resolution due to diffraction would be approximately .3MP on a FF sensor (average RGB; regardless of sensor resolution). At f/45 it is ~1MP. 1MP may sound like pathetically little, but 1MP is essentially all that can be contained in an image that is 1024x in size. At f/45 true you are getting the potential of 1MP in sharp focus, and presenting at 1024x makes the most use of that (downsizing the full sensor area).

Put another way, diffraction losses are more about the loss of recorded resolution/details that are w/in the plane of sharp focus... the actual aperture determines this.
Whereas DOF (acceptable focus) has more to do with the relative sharpness of things (primarily outside the plane of sharp focus)... the effective aperture (effective magnification) determines this; as does the display size and viewing distance (apparent magnification).
Would you please help me make sure I am correctly understanding what you are saying here. I think you are saying (or perhaps implying rather than saying) that the sharpness within the plane of sharp focus depends on the actual aperture, but is independent of the effective aperture. Am I right in thinking this?

An implication of this would be, I think, that if you captured an image and then changed the effective aperture without changing the actual aperture, then within the plane of sharp focus the two images would be equally sharp. Would that be correct?
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Wow well I hope it all turns out to be a false alarm Nick, must have been worrying
It was rather worrying, yes. I know this sounds weird, but for the foreseeable future there is no way of knowing whether it is a false alarm or not. There are currently no signs that I have mesothelioma, but there was a solid diagnosis of it - symptoms, scans and biopsy - and it never regresses, and most certainly not with no treatment, and I haven't had any treatment. Two professors, leading experts in this field, one a medical doctor and one a pathologist, have conflicting views about what is going on. I can't see any way they are going to get new evidence to prove it one way or the other, unless it rears up again of course. It's just a case of carry on regardless at this point and hope for the best.

but getting the kit must have given you some motivation, your experiments have always been so concise and well executed. Often proving so many misconceptions about macro along the way :D.
Thanks. I think I do enjoy being a bit contrarian. :) And yes, having a fair amount of kit I can mix and match has increased the range of experiments I can do, and I do so enjoy doing experiments.

I'm rushing 2020 along and hoping to get to my 40th next year quickly as I've been allowed to indulge into a new camera (hopefully the r6 will be released by then). Want to get a mirrorless and full frame :D.
I hope your wishes come true, and that you and yours stay healthy and happy.
 
Messages
7,148
Name
Steven
Edit My Images
Yes
Would you please help me make sure I am correctly understanding what you are saying here. I think you are saying (or perhaps implying rather than saying) that the sharpness within the plane of sharp focus depends on the actual aperture, but is independent of the effective aperture. Am I right in thinking this?
Yes.
Put another way, if you have a detail of x sharpness and enlarge it (reduce the effective aperture) it doesn't become less sharp technically, it just covers more pixels.
An implication of this would be, I think, that if you captured an image and then changed the effective aperture without changing the actual aperture, then within the plane of sharp focus the two images would be equally sharp. Would that be correct?
Yes, when the subject/detail is displayed/viewed at the same size.

Decreasing the effective aperture is very much like zooming in on a computer screen... I.e. the increased magnification makes things that are less sharp more apparent. And then if you get up and move away from the screen, the image becomes relatively smaller and those details become relatively sharper again. DOF (recorded/displayed/viewed) is all a result of relative magnification (making things more/less visible); and the effective aperture is also just a result of magnification (enlarging the image circle).
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Question and Answer #1

Would you please help me make sure I am correctly understanding what you are saying here. I think you are saying (or perhaps implying rather than saying) that the sharpness within the plane of sharp focus depends on the actual aperture, but is independent of the effective aperture. Am I right in thinking this?

Yes.
Put another way, if you have a detail of x sharpness and enlarge it (reduce the effective aperture) it doesn't become less sharp technically, it just covers more pixels.
Question and answer #2

An implication of this would be, I think, that if you captured an image and then changed the effective aperture without changing the actual aperture, then within the plane of sharp focus the two images would be equally sharp. Would that be correct?

Yes, when the subject/detail is displayed/viewed at the same size.

Decreasing the effective aperture is very much like zooming in on a computer screen... I.e. the increased magnification makes things that are less sharp more apparent. And then if you get up and move away from the screen, the image becomes relatively smaller and those details become relatively sharper again. DOF (recorded/displayed/viewed) is all a result of relative magnification (making things more/less visible); and the effective aperture is also just a result of magnification (enlarging the image circle).
Yes, when the subject/detail is displayed/viewed at the same size.

Decreasing the effective aperture is very much like zooming in on a computer screen... I.e. the increased magnification makes things that are less sharp more apparent. And then if you get up and move away from the screen, the image becomes relatively smaller and those details become relatively sharper again. DOF (recorded/displayed/viewed) is all a result of relative magnification (making things more/less visible); and the effective aperture is also just a result of magnification (enlarging the image circle).
Thanks. I have to say I'm still having difficulty getting my head around this. I discuss some practical tests below. If you are concerned that my approach to capturing the test images may have influenced the results in some way then I will happily describe how I did the captures (In summary, best of 5 for each example, tripod, manual focus, magnified view, time for rig to settle, remote release, flash.)

Here is a comparison at 100% between 1:1 at f/16 on the left below and 5:1 at f/16 on the right below. Given your answer of "Yes" in Question and Answer #1 above I would expect these to have the same level of detail/sharpness, but the 1;1 image looks much sharper to me. (Full 1440p screen size version at Flickr)


1649 01 Compare 1to1 at F16 with 5to1 at F16 both at 100pc
by gardenersassistant, on Flickr

To my eye 1:1 at f/16 on the left below looks rather similar in terms of detail/sharpness to f/5.6 at 5:1 on the right below. In this case both have the same effective aperture of f/32.


1649 04 Compare 1to1 at F16 with 5to1 at F5.6 both at 100pc
by gardenersassistant, on Flickr

On the other hand in your answer to Question and Answer #2 above you wrote "Yes, when the subject/detail is displayed/viewed at the same size." To do that in this case we would I think need to view the 1:1 image at 5X the magnification of the 5:1 image, so for example at 500% for the 1:1 image versus 100% for the 5:1 image. However the 1:1 image is so pixellated at 500% that the comparison becomes rather iffy I think. Here is a compromise, with the 1:1 image at 200% on the left and the 5:1 image at 40% on the right. The 5:1 image looks a bit sharper to me, but allowing for the fact we are upscaling the 1:1 image and downscaling the 5:1 image, they might be essentially similar in terms of sharpness/detail. But assuming this is the case, I'm not sure how useful that is in practical terms.


1649 03 Compare 1to1 at F16 at 200pc with 5to1 at F16 at 40pc
by gardenersassistant, on Flickr

The similar level of detail/sharpness at 1:1 f/16 and 5;1 f/5.6 seems to me to be consistent with my understanding that when trying to keep to a similar level of diffraction softening one should adjust f-numbers to equalise effective apertures, which is what I have been doing for some years now.
 
Last edited:
Messages
7,148
Name
Steven
Edit My Images
Yes
The 5:1 image looks a bit sharper to me, but allowing for the fact we are upscaling the 1:1 image and downscaling the 5:1 image, they might be essentially similar in terms of sharpness/detail. But assuming this is the case, I'm not sure how useful that is in practical terms.


I agree, the 5:1 image looks sharper. And the details are a bit larger aren't they?

The main comment I have about the images is that I do not see where the details in the right image originate from in the left image... it almost appears to me as if two different systems were used simultaneously to record different areas; that would introduce many variables.

I think the best way of comparing actual lens diffraction limitation to the effect of effective aperture limitation would be when the lens aperture in the first is the same as the effective aperture of the second...
Ideally, I think you would use exactly the same setup at f/X actual w/ no extension compared to f/X effective with extension (and record the same central FOV area). E.g. f/16 actual w/o extension compared to f/2.8 at 5x (f/17 effective)... that would give the 5x a big advantage; but that's assuming a lens that is diffraction limited at f/2.8 which is unlikely. If that were the case the extended lens would just limit sensor resolution (<24MP FF), but the display/print would not be diffraction limited based on the 30um FF CoC.
A better/more realistic comparison might be f/24 (f/22) actual compared to f/4 at 5x (f/24 effective); which puts the extended lens just above the display/print diffraction limit (32micron airy disk). Going much smaller with the actual aperture setting/diffraction would likely require making/using aperture masks (e.g. f/135); but the same trend should follow.

When you stop down a lens (actual aperture), diffraction limits the lens' resolution... and when the lens' resolution is less than the sensor's resolution, it also limits the system/recorded resolution. But stopping down also makes the focus more "flat field," and it eliminates optical errors; both because you are using less of the lens' periphery... it also increases the DOF; there's a lot of tradeoff there which can be quite beneficial (up to a point).
When you magnify what the lens projects (effective aperture), you only change how much of it is recorded. If the lens is capable of resolving more than the sensor at its' actual aperture, then the effective aperture may not introduce any limitation. E.g. f/4 actual and f/16 effective on a 16MP FF sensor. If the lens is already the limitation (it typically is), then the magnification/effective aperture makes the shortcoming(s) more apparent. The net result in that case is more apparent softening (diffraction/reduced DOF), but for a slightly different reason... at the extreme apertures you're working with the difference may be negligible, IDK. But to me the f/134 effective images do not look like 1/3MP max resolution displayed at 1MP size (1024x).


Here's a thought experiment for you:
If you take an image and crop away 75%, what changes about that remaining 25%?
If you view the 25% crop at the same physical size as the original full image, what differences exist? (e.g. 8" print, or 8" on the monitor)
What are the limitations of using/outputting the crop compared to using the full original?
And then; other than exposure (required SS), what is the difference between cropping in post vs enlarging the image circle so that your sensor crops it at the time of recording?
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
I agree, the 5:1 image looks sharper. And the details are a bit larger aren't they?

The main comment I have about the images is that I do not see where the details in the right image originate from in the left image... it almost appears to me as if two different systems were used simultaneously to record different areas; that would introduce many variables.

I think the best way of comparing actual lens diffraction limitation to the effect of effective aperture limitation would be when the lens aperture in the first is the same as the effective aperture of the second...
Ideally, I think you would use exactly the same setup at f/X actual w/ no extension compared to f/X effective with extension (and record the same central FOV area). E.g. f/16 actual w/o extension compared to f/2.8 at 5x (f/17 effective)... that would give the 5x a big advantage; but that's assuming a lens that is diffraction limited at f/2.8 which is unlikely. If that were the case the extended lens would just limit sensor resolution (<24MP FF), but the display/print would not be diffraction limited based on the 30um FF CoC.
A better/more realistic comparison might be f/24 (f/22) actual compared to f/4 at 5x (f/24 effective); which puts the extended lens just above the display/print diffraction limit (32micron airy disk). Going much smaller with the actual aperture setting/diffraction would likely require making/using aperture masks (e.g. f/135); but the same trend should follow.

When you stop down a lens (actual aperture), diffraction limits the lens' resolution... and when the lens' resolution is less than the sensor's resolution, it also limits the system/recorded resolution. But stopping down also makes the focus more "flat field," and it eliminates optical errors; both because you are using less of the lens' periphery... it also increases the DOF; there's a lot of tradeoff there which can be quite beneficial (up to a point).
When you magnify what the lens projects (effective aperture), you only change how much of it is recorded. If the lens is capable of resolving more than the sensor at its' actual aperture, then the effective aperture may not introduce any limitation. E.g. f/4 actual and f/16 effective on a 16MP FF sensor. If the lens is already the limitation (it typically is), then the magnification/effective aperture makes the shortcoming(s) more apparent. The net result in that case is more apparent softening (diffraction/reduced DOF), but for a slightly different reason... at the extreme apertures you're working with the difference may be negligible, IDK. But to me the f/134 effective images do not look like 1/3MP max resolution displayed at 1MP size (1024x).


Here's a thought experiment for you:
If you take an image and crop away 75%, what changes about that remaining 25%?
If you view the 25% crop at the same physical size as the original full image, what differences exist? (e.g. 8" print, or 8" on the monitor)
What are the limitations of using/outputting the crop compared to using the full original?
And then; other than exposure (required SS), what is the difference between cropping in post vs enlarging the image circle so that your sensor crops it at the time of recording?
We can come back to some of this (e.g. I used the exactly the same kit for all the shots, an MPE-65 on a 70D, and you didn't mention the similarity of level of detail/sharpness I noted between the shots with the same effective aperture, and various other things), but let's park all that for now, along your thought experiment, which I'm happy to give you my answers to. For now though I'd like to clarify the test you suggest doing (which I will do).

You say "Ideally, I think you would use exactly the same setup at f/X actual w/ no extension compared to f/X effective with extension (and record the same central FOV area)"

Q1: Suppose I use the MPE-65 at 1X using f/22 and at 5X using f/4. Would that meet your criteria?
DOH! The MPE-65 doesn't go to f/22. Would 1X at f/16 and 5:1 at f/2.8 meet your criteria?
Q2: You want me to record the same central area. How would I do this with the magnification being so different? (I can centre the shots in the same place, but the area recorded won't be the same.)
Q3: What result are you expecting?
Q4: What conclusion would you draw from the expected result?
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
I agree, the 5:1 image looks sharper. And the details are a bit larger aren't they?

The main comment I have about the images is that I do not see where the details in the right image originate from in the left image... it almost appears to me as if two different systems were used simultaneously to record different areas; that would introduce many variables.

I think the best way of comparing actual lens diffraction limitation to the effect of effective aperture limitation would be when the lens aperture in the first is the same as the effective aperture of the second...
Ideally, I think you would use exactly the same setup at f/X actual w/ no extension compared to f/X effective with extension (and record the same central FOV area). E.g. f/16 actual w/o extension compared to f/2.8 at 5x (f/17 effective)... that would give the 5x a big advantage; but that's assuming a lens that is diffraction limited at f/2.8 which is unlikely. If that were the case the extended lens would just limit sensor resolution (<24MP FF), but the display/print would not be diffraction limited based on the 30um FF CoC.
A better/more realistic comparison might be f/24 (f/22) actual compared to f/4 at 5x (f/24 effective); which puts the extended lens just above the display/print diffraction limit (32micron airy disk). Going much smaller with the actual aperture setting/diffraction would likely require making/using aperture masks (e.g. f/135); but the same trend should follow.

When you stop down a lens (actual aperture), diffraction limits the lens' resolution... and when the lens' resolution is less than the sensor's resolution, it also limits the system/recorded resolution. But stopping down also makes the focus more "flat field," and it eliminates optical errors; both because you are using less of the lens' periphery... it also increases the DOF; there's a lot of tradeoff there which can be quite beneficial (up to a point).
When you magnify what the lens projects (effective aperture), you only change how much of it is recorded. If the lens is capable of resolving more than the sensor at its' actual aperture, then the effective aperture may not introduce any limitation. E.g. f/4 actual and f/16 effective on a 16MP FF sensor. If the lens is already the limitation (it typically is), then the magnification/effective aperture makes the shortcoming(s) more apparent. The net result in that case is more apparent softening (diffraction/reduced DOF), but for a slightly different reason... at the extreme apertures you're working with the difference may be negligible, IDK. But to me the f/134 effective images do not look like 1/3MP max resolution displayed at 1MP size (1024x).


Here's a thought experiment for you:
If you take an image and crop away 75%, what changes about that remaining 25%?
If you view the 25% crop at the same physical size as the original full image, what differences exist? (e.g. 8" print, or 8" on the monitor)
What are the limitations of using/outputting the crop compared to using the full original?
And then; other than exposure (required SS), what is the difference between cropping in post vs enlarging the image circle so that your sensor crops it at the time of recording?
Here is my current understanding of the situation.

You believe that the sharpness within the plane of sharp focus depends on the actual aperture, but is independent of the effective aperture.
You believe that if you captured an image and then changed the effective aperture without changing the actual aperture, then within the plane of sharp focus the two images would be equally sharp when the subject/detail is displayed/viewed at the same size.
I believe that the sharpness within the plane of focus depends on the effective aperture.
I believe that if you captured an image and then changed the effective aperture without changing the actual aperture, then within the plane of sharp focus the two images would not be equally sharp when both images are viewed at the same magnification (e.g. both viewed at 100%).

The following table summarises a test you have proposed, and two tests that I propose.


1650 01 Table of test settings
by gardenersassistant, on Flickr

I have captured the images needed for your test as per the settings I suggested at Q1 in my previous post. As suggested in that post, you may want to clarify the objectives of your proposed test and confirm (or otherwise) my proposed approach to your test and suggested settings before we discuss the results of your suggested test. In the meantime I will report on the results from my proposed tests.

The images for your and my proposed tests were captured with the same camera, a Panasonic G80, and the same lens, a Canon MPE-65, using flash. The camera was on a tripod and boom arm which allows for movement perpendicular to the sensor plane (i.e. directly towards and away from the subject).


1649 05 Capture setup
by gardenersassistant, on Flickr

I used manual focus with a magnified picture in picture focusing area. After adjusting the focus, by moving the camera towards/away from the subject, I waited for the rig to settle before assessing whether the focus was good. I made numerous tiny adjustments back and forth using the handle on the boom arm, until I was satisfied that I knew what the best focus looked like and that the rig had settled in that position. I then used a wired remote shutter release to capture a raw+JPEG pair. I did this five times for each of the test settings. The following illustration shows the coverage of the four test settings.


1650 08 The four test shots
by gardenersassistant, on Flickr

I have used the out of the camera JPEGs so there would be no problem of inconsistent processing of the raw files. I chose the best focused/detailed of each set of five images. The selected full size out of the camera JPEGs are in this album at Flickr.

Here is the comparison proposed in reference lines 3 and 4 of the above table. It shows, at 100%, on the left, f/16 at 1:1 magnification, and on the right f/16 at 5:1 magnification. As per the above table, if diffraction blurring depends solely on aperture and is independent of effective aperture then these two images, which have the same aperture, should be similarly sharp/detailed. They are not.


1650 02 Same aperture, different effective aperture
by gardenersassistant, on Flickr

Here is the comparison proposed in reference lines 5 and 6 of the above table. It shows, at 100%, on the left f/16 at 1:1 magnification, and on the right f/5.6 at 5:1 magnification. As per the above table, if diffraction blurring depends on effective aperture then these two images, which have the same effective aperture, should be similarly sharp/detailed. They are.


1650 03 Same effective aperture, different aperture
by gardenersassistant, on Flickr

I believe these test results are consistent with the hypothesis that the amount of diffraction blur depends on effective aperture, and are inconsistent with the hypothesis that the amount of diffraction blur depends solely on aperture and is independent of effective aperture.
 
Last edited:
Messages
7,596
Name
Graham
Edit My Images
No
These are fantastic! May I ask what the three small globes are in #3? First thought is eyes, but it already has two giant ones.
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
Messages
4,332
Name
Laurence
Edit My Images
Yes
@GardenersHelper , thank you for starting a very instructive thread. I'm just starting to get into "serious" macro, there is so much to learn and so much experimentation.
In particular I want to thank you for that video you made in your alter ego of GardenersAssistant where you discussed the reasons for haloing of stacked images and how to overcome them. Saved me a lot of angst!
 
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
@GardenersHelper , thank you for starting a very instructive thread. I'm just starting to get into "serious" macro, there is so much to learn and so much experimentation.
In particular I want to thank you for that video you made in your alter ego of GardenersAssistant where you discussed the reasons for haloing of stacked images and how to overcome them. Saved me a lot of angst!
That is really good to hear. Thank you for the feedback. I'm glad the video was helpful.

I suppose all aspects of photography have their own complications, but I suspect close-up/macro may be one of the more complicated areas. There is certainly plenty to explore and enjoy.

I would be GardenersAssistant here too, but that username was too long for this site.
 
Messages
7,148
Name
Steven
Edit My Images
Yes
Here is the comparison proposed in reference lines 5 and 6 of the above table. It shows, at 100%, on the left f/16 at 1:1 magnification, and on the right f/5.6 at 5:1 magnification. As per the above table, if diffraction blurring depends on effective aperture then these two images, which have the same effective aperture, should be similarly sharp/detailed. They are.


1650 03 Same effective aperture, different aperture
by gardenersassistant, on Flickr

I believe these test results are consistent with the hypothesis that the amount of diffraction blur depends on effective aperture, and are inconsistent with the hypothesis that the amount of diffraction blur depends solely on aperture and is independent of effective aperture.
The tests show exactly what they should...
It's hard to separate the effective aperture from the actual aperture; but what I am saying is that what is really making the difference here is f/16 vs f/5.6... and it is potentially a big difference/advantage.
Diffraction causes the loss of the ability to record fine details of the subject; and there is clearly a difference in the sharpness of the same fine details between the two images... if there wasn't, then you should be able to enlarge the 1:1 image and see the details just as clearly; but you can't... those small details are softer at f/16 due to lens diffraction. It's exactly what the first test shows with the lens at f/16 for both images... the additional magnification only makes it more apparent how soft the small subject details are.

At f/5.6 a perfect lens can project airy disks that are 7.5um in size (average/middle wavelengths) which equates to 60MP, whereas the same lens at f/16 can only project airy disks that are 21.5um and 7MP (FF). If you just calculate it mathematically; at f/16 you have the full image circle at 7MP, vs 1/5 of the 60MP f/5.6 image circle or 12MP.
This difference goes away quickly as you stop down, because the maximum possible lens resolution decreases by around 50% for every f-stop. And those resolutions are theoretical... but to my eye, the 5:1 f/5.6 image shows notably more resolution/detail/sharpness. FWIW, if these were f/32 actual images then the maximum recordable resolution of the original subject detail would be no more than 2MP (which would look quite bad at 100% zoom on a monitor).

It appears that what you are considering as being diffraction/diffraction loss is the sharpness of details that recorded the same size on the sensor (different sized subject details). That is really more of a depth of field/display consideration; but it is a common calculation/comparison. And in that case both the DOF and diffraction calculations use the same FF 30um COC/airy disk limit/standard... which equates to just under 1MP.
 
Last edited:
OP
GardenersHelper
Messages
5,691
Name
Nick
Edit My Images
Yes
The tests show exactly what they should...
It's hard to separate the effective aperture from the actual aperture; but what I am saying is that what is really making the difference here is f/16 vs f/5.6... and it is potentially a big difference/advantage.
Diffraction causes the loss of the ability to record fine details of the subject; and there is clearly a difference in the sharpness of the same fine details between the two images... if there wasn't, then you should be able to enlarge the 1:1 image and see the details just as clearly; but you can't... those small details are softer at f/16 due to lens diffraction. It's exactly what the first test shows with the lens at f/16 for both images... the additional magnification only makes it more apparent how soft the small subject details are.

At f/5.6 a perfect lens can project airy disks that are 7.5um in size (average/middle wavelengths) which equates to 60MP, whereas the same lens at f/16 can only project airy disks that are 21.5um and 7MP (FF). If you just calculate it mathematically; at f/16 you have the full image circle at 7MP, vs 1/5 of the 60MP f/5.6 image circle or 12MP.
This difference goes away quickly as you stop down, because the maximum possible lens resolution decreases by around 50% for every f-stop. And those resolutions are theoretical... but to my eye, the 5:1 f/5.6 image shows notably more resolution/detail/sharpness. FWIW, if these were f/32 actual images then the maximum recordable resolution of the original subject detail would be no more than 2MP (which would look quite bad at 100% zoom on a monitor).

It appears that what you are considering as being diffraction/diffraction loss is the sharpness of details that recorded the same size on the sensor (different sized subject details). That is really more of a depth of field/display consideration; but it is a common calculation/comparison. And in that case both the DOF and diffraction calculations use the same FF 30um COC/airy disk limit/standard... which equates to just under 1MP.
I think we are talking at cross-purposes and that there is little benefit in continuing this discussion. Thanks for taking the time to explain the aspects of this that interest you. For my part I will continue to use effective (and equivalent effective) aperture calculations to help me produce similar-looking results with the four different formats and two different types of close-up arrangements that I use.
 
Top