Sharpness of microscopic photos at 10x?

A forum to ask questions, post setups, and generally discuss anything having to do with photomacrography and photomicroscopy.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

plantfan123
Posts: 21
Joined: Tue Dec 08, 2020 12:39 pm

Sharpness of microscopic photos at 10x?

Post by plantfan123 »

Hi. I am relatively new to doing focus stacking using microscopic objectives. I am shooting a moss with a Nikon CFI Plan 10X Achromat objective attached to a Nikon Z7 45 MP mirrorless camera and am noticing that the focus stacked images are a bit blurry when zoomed in at 100%. If I view the photo at half or a quarter of the resolution (but at the same physical size on my screen), I barely notice a difference. I am shooting using the electronic shutter.

The 10x objective is attached to a Nikon 105 mm f/2.8 macro lens which itself is attached to the camera via the Nikon F-Z adapter, so maybe this part could be improved upon? It's nice because it makes the 10x objective's light occupy the whole sensor.

I am looking at getting a second mirrorless camera for another focus stacking set-up. When shooting with microscope objectives are high resolution cameras like the Z7 even worth it? Is there a way that I could modify this 10x set-up to get sharper images? Would extension tubes make for sharper images?

I've attached the focus stacked image (made up of 50 constituent images) as well as one of the constituent images. Thanks!

Focus stacked image:

https://www.dropbox.com/s/eby4ktnomxabs ... images.jpg

Single constituent image:

https://www.dropbox.com/s/0tw9g9fqpzwy6 ... -image.jpg

palea
Posts: 40
Joined: Mon Dec 17, 2018 8:09 pm

Re: Sharpness of microscopic photos at 10x?

Post by palea »

Hmm, a Plagiomnium? This is definitely softer than what I get with similar species and either an Olympus MPlanFL 5x 0.15 around 5x or a UPlanFL 10x 0.3 around 10x before deconvolving diffraction from the higher effective apertures on a tighter pixel pitch than the Z9's. From the amount of chromatic aberration I'm guessing it might be worth changing the objective but, just as a check, which 105 f/2.8 are you using at which marked aperture? Even if there's no reason to suspect alignment or other issues in the objective or rear lens, it's conceivable the Micro-Nikkor could be making a noticeable contribution.

The rear lens is stopped down to f/10 or so from the front in this configuration, meaning it's probably not the main limiting factor, but I test that with a series of autofocus brackets progressively stopping the rear lens down to f/11 or so. While it's not what simplified coupled lens theory suggests, I often find there's some pixel peeping improvement to closing the rear lens down to f/5.6 or maybe f/8 compared to leaving it fully open. It's an easy check to run and helps to avoid underutilizing the optics you have before chasing others.

plantfan123
Posts: 21
Joined: Tue Dec 08, 2020 12:39 pm

Re: Sharpness of microscopic photos at 10x?

Post by plantfan123 »

Thanks for the response. Yeah I thought Plagiomnium too. I am using the Nikon 105mm f/2.8 FX AF MICRO-NIKKOR (1990-2007) with an aperture of f/4.8. There's also a ProMaster circular polarizing filter right before the 10x objective.

The quality of the objective would be the simplest explanation! The UPlanFL 10x 0.3 is close to $900 US while my Nikon CFI Plan 10x/0.25na was $250. So it makes sense that a more expensive objective would produce sharper images, I would imagine? Would love to see your images of similar mosses with either objective to compare if you have them handy.

Would you mind explaining what the pros and cons are of using a lens versus extension tubes for having microscope objectives on cameras? Just curious.

Thanks for the info on your set-up! I'll play with the aperture to see if I get any improvements that way. Will post them here.

Scarodactyl
Posts: 1079
Joined: Sat Apr 14, 2018 10:26 am

Re: Sharpness of microscopic photos at 10x?

Post by Scarodactyl »

You are pushing the objective way put of spec with the 105mm lens, but that shouldn't hurt center performance. This objective can take great photos when set up right, and this definitely looks like underperforming. Just to be sure, the polarizer is between the objective and lens and the lens is focused to infinity, right?

rjlittlefield
Site Admin
Posts: 22295
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Sharpness of microscopic photos at 10x?

Post by rjlittlefield »

Explaining further about the sensor resolution issue...

Nikon has a rule for matching the resolutions of sensor and optics. Boiled down, that rule says to use a sensor that puts two pixels per line pair at the diffraction limit of the objective. The justification for this rule is that you need at least that many pixels because of the Nyquist sampling limit, and you don't need any more because the rule also puts 3 pixels per line pair at a level of detail where the objective cannot have contrast higher than 22% MTF. A good case can be made that more pixels are actually required, because of several issues with digital sampling, but for the sake of discussion let's suppose that Nikon got it right.

With a 10X NA 0.25 objective used at 5X, Nikon's rule says that the sensor would need pixels of size 2.75 microns, or 114 megapixels in full frame. Clearly your Z7 at only 45 megapixels is not overkill by this rule.

That said, a sensor that is operating anywhere near Nikon's rule will not give images that look sharp when pixel-peeping at 100%. That's because Nikon's rule says that a level of detail which would produce alternating light/dark/light/dark pixels would suffer from MTF=0 (no contrast at all). Even at 4 pixels per line pair, nominally giving a light/light/dark/dark pattern, the objective will give no more than MTF=39%. It should be clear that such images will have a lot of softening that is unavoidable, caused simply by diffraction. There is nothing you can do that will get around the fundamental limits of diffraction. The best that you can do, again as Scarodactyl suggests, is to digitally sharpen the captured images, best using deconvolution but a well chosen "unsharp mask" will do almost as well.

Now, regarding other issues...

I see your comment about f/4.8, and it concerns me. The simplest way that I know to get "f/4.8" from a Nikon f/2.8 macro lens is to focus it down to 1:1. As Scarodactyl says, infinity objectives like the Nikon CFI are designed to be used with rear lenses focused at infinity. At only NA 0.25, you can deviate from infinity focus by quite a bit, but I would expect some degradation from going all the way down to 1:1.

I have looked at your images to see if I can spot other problems that may be degrading them.

Comparing your stacked and original images, I also see that in some places the stacked image does not look quite as sharp as the original source. That discrepancy suggests that something is going wrong in the stacking process. You don't describe exactly how you're doing the stacking, so again I'm guessing. However, I do notice that (1) the stacked and single source images are shifted by several pixels, and (2) the single image contains a hot pixel (coordinates 2547,3692) that is near focused detail, but in the stacked output image only a single copy of the hot pixel appears, at different coordinates and in a different position with respect to the focused detail. That combination of observations tells me that you're probably using a depth map method, and that the method is not properly set for the level of detail that is found in your images. If you're using Helicon Focus, as suggested by the link that you provide, then I suggest (1) try Method C, and/or (2) with Method B, experiment with different Radius settings.

--Rik

plantfan123
Posts: 21
Joined: Tue Dec 08, 2020 12:39 pm

Re: Sharpness of microscopic photos at 10x?

Post by plantfan123 »

Scarodactyl wrote:
Sat Nov 27, 2021 12:26 pm
You are pushing the objective way put of spec with the 105mm lens, but that shouldn't hurt center performance. This objective can take great photos when set up right, and tjis dwfinitely looks like underperforming. Just to be sure, thr polarizer is between the objective and lens and the lens is focused to infinity, right?
Correct, the polarizer is between the objective and lens. When the lens is focused to infinity the moss only occupies about a quarter of the sensor, so I set the focus to almost the closest it can get. Maybe that's contributing to the bad quality? Is there a better way of having the moss occupy the full sensor? Thanks.

Scarodactyl
Posts: 1079
Joined: Sat Apr 14, 2018 10:26 am

Re: Sharpness of microscopic photos at 10x?

Post by Scarodactyl »

Yeah, that's probably your problem. The objective is only designed to work with the tube lens focused to infity. As is it's pushed way out of spec. You need a longer focal length lens to fill the frame.

rjlittlefield
Site Admin
Posts: 22295
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Sharpness of microscopic photos at 10x?

Post by rjlittlefield »

plantfan123 wrote:
Sat Nov 27, 2021 12:18 pm
Would you mind explaining what the pros and cons are of using a lens versus extension tubes for having microscope objectives on cameras? Just curious.
There are two types of objectives. "Finite" objectives are designed to be used on empty extension. "Infinity" objectives are designed to be used with an additional "tube lens" that is focused at infinity.

The stuff behind the objective is included in the optical design of the objective. Using either type in the opposite way will introduce aberrations that degrade the image.

The same applies if you focus the tube lens away from infinity, or if you change extension with a finite objective.

For some images, see viewtopic.php?p=139499#p139499 , which includes comparison of Nikon 10X CFI Plan Achromat (MRL00102) on Canon 100 mm f/2.8 macro, focused at infinity versus 1:1. Then viewtopic.php?p=59673#p59673 shows a comparison of finite versus infinite, both on tube lens, which is not appropriate for the finite. I do not have comparison images for a 10X infinity objective used as finite. I do have that comparison for 50X NA 0.55 infinity used as finite, at viewtopic.php?p=103915#103915 . Short summary is that using the NA 0.55 in the wrong way is a disaster, though surprisingly well correctable with heavy digital filtering.

Infinity objectives have the advantage that you can adjust magnification without adding aberrations, by changing the focal length of the rear lens, while leaving it focused at infinity. More magnification can be obtained with both finite and infinite objectives, in theory without adding aberrations, by adding a teleconverter on the camera in its usual position behind all other lenses and extensions.

--Rik

palea
Posts: 40
Joined: Mon Dec 17, 2018 8:09 pm

Re: Sharpness of microscopic photos at 10x?

Post by palea »

plantfan123 wrote:
Sat Nov 27, 2021 12:18 pm
Thanks for the response.
You're welcome. +1 on what Scarodactyl and Rik have said about the mismatch between trying to use an infinity objective with finite conjugates and changing to a 200 mm rear lens to match the objective's designed tube lens focal length.

I'd also suggest testing with the polarizer removed. A filter in an infinity space between the objective and rear lens (created by focusing the 105 at infinity) isn't expected to have much effect on aberrations but it's an easy thing to check. My experience with short conjugates, which a 105 mm macro at its 1x focus position approaches, is filters have substantial effects on spherical aberration. (With the 105 towards close focus you may even find removing the polarizer decreases image quality and it's possible inserting an intermediate amount of glass, such as a protection filter, works better than not having a filter or using the polarizer.)
plantfan123 wrote:
Sat Nov 27, 2021 12:18 pm
So it makes sense that a more expensive objective would produce sharper images, I would imagine?
In the used objective market optical quality and price are often loosely coupled. Mitutoyos are especially notorious for this (there are many posts here about that) but, as another example, I paid US$ 180 for the UPlanFL N I have. What you wrote reads like you might be unfamiliar with the differences between plan, fluor(ite), and apo(chromatic) objectives, in which case I would recommend reading up a bit.

I am not sure what is going on with Plagiomnium cell walls optically but, having imaged them with a range of coupled lens configurations, it seems like they're always prone to chromatic aberration. I suspect the walls or intercellular material between then are prismatic and that this is challenging for the optics, along with maybe demosaicing from the sensor. There's possibly also some interaction with the continuous LED lighting I use. It's not unique to Plagiomnium—I've encountered similar behavior with hyaline awns of some other genera—but the plan fluors do control it better than the plan achromat I have. Olympus markets their UIS 2 plan fluors as semi-apo, which might be more about the NIR correction, .
plantfan123 wrote:
Sat Nov 27, 2021 12:18 pm
Would you mind explaining what the pros and cons are of using a lens versus extension tubes for having microscope objectives on cameras? Just curious.
If you don't have an autofocus rear lens then you can't autofocus bracket and are limited to focus bracketing with linear motion rails. For the most part autofocus bracketing is cheaper, faster, easier, and has smaller perspective effects than linear motion bracketing. Olympus and Panasonic introduced autofocus bracketing in 2015 and 2016 but, as Rik created this forum well before that and most people here have DSLRs, linear rails still have most of the attention. In the past couple years Canon, Nikon, and Fuji have increased availability of autofocus bracketing but Sony still lacks comparable support, I think.

The other major advantage to coupled lenses is their effective aperture is EA = MN rather than EA = (M + 1)N for magnification M obtained by extension, N being the f/stop of the limiting aperture in the lens. There's a transition zone from around 1x to 3x but at 4+x the limiting aperture of a coupled lens is very likely the microscope objective. In this case a coupled lens with an infinity objective is a stop faster than a DIN objective with the same numerical aperture, meaning you need a stop less lighting and get a stop less diffraction. It's one reason why compound microscopes started changing to infinity optical trains in the 1930s.

There are further details but that's probably enough for now.
Last edited by palea on Sun Nov 28, 2021 10:26 am, edited 2 times in total.

rjlittlefield
Site Admin
Posts: 22295
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Sharpness of microscopic photos at 10x?

Post by rjlittlefield »

palea wrote:
Sun Nov 28, 2021 9:07 am
In this case a coupled lens with an infinity objective is a stop faster than a DIN objective with the same numerical aperture, meaning you need a stop less lighting and get a stop less diffraction. It's why compound microscopes changed to infinity optical trains in the 1930s.
With respect, this is simply not true.

Manufacturers of microscope objectives label them with the NA (numerical aperture) that they will actually deliver, when used as designed. There is no difference in lighting or diffraction between 10X NA 0.25 finite and 10X NA 0.25 infinite. If you delve into the details of those objectives you would find that the finite has a wider intrinsic aperture, but that difference is exactly made up by the addition of the rear lens to the infinite. When used as designed, they accept and deliver exactly the same cones of light.

The reason that compound microscopes changed to infinity optics was for simplicity in designing add-on optics such as through-the-lens illumination and filtering. This is because planar elements can be freely inserted into the infinity space without adding aberrations. In contrast, adding elements into a finite system always introduces aberrations, unless the elements are made carefully non-planar to avoid them.

--Rik

palea
Posts: 40
Joined: Mon Dec 17, 2018 8:09 pm

Re: Sharpness of microscopic photos at 10x?

Post by palea »

rjlittlefield wrote:
Sun Nov 28, 2021 10:14 am
With respect, this is simply not true.
Then it would be appropriate to retract or revise the posts you've made indicating a one stop difference between extension and coupled lenses over the past years. FAQ entries also merit updating. Since MN and (M + 1)N has been in place for some time and is in good agreement with my effective aperture measurements it seemed plausible that corpus of information was stable for the purposes of this thread.

From your reply I'm afraid I can't tell what the replacement is. If you could link the current model that would be appreciated as I'm not finding it. What I'm seeing on search is you were still using MN and (M + 1)N a year ago and that this didn't come up in last month's similar FAQ discussion.

JKT
Posts: 315
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Re: Sharpness of microscopic photos at 10x?

Post by JKT »

palea wrote:
Sun Nov 28, 2021 10:51 am
Then it would be appropriate to retract or revise the posts you've made indicating a one stop difference between extension and coupled lenses over the past years. FAQ entries also merit updating. Since MN and (M + 1)N has been in place for some time and is in good agreement with my effective aperture measurements it seemed plausible that corpus of information was stable for the purposes of this thread.
Nope. The difference is real when you apply it to coupled lenses. Those are designed to work alone and marked according to that. On the other hand, infinite microscope objectives are designed to work coupled with another lens and the markings on those follow that principle.

I recall only one case where you'd want to use finite microscope objective as infinite. That is LOMO 3.7x. When you use that as infinite, you get the advantage as the markings are for finite use.

rjlittlefield
Site Admin
Posts: 22295
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Re: Sharpness of microscopic photos at 10x?

Post by rjlittlefield »

palea wrote:
Sun Nov 28, 2021 10:51 am
rjlittlefield wrote:
Sun Nov 28, 2021 10:14 am
With respect, this is simply not true.
Then it would be appropriate to retract or revise the posts you've made indicating a one stop difference between extension and coupled lenses over the past years. FAQ entries also merit updating. Since MN and (M + 1)N has been in place for some time and is in good agreement with my effective aperture measurements it seemed plausible that corpus of information was stable for the purposes of this thread.

From your reply I'm afraid I can't tell what the replacement is. If you could link the current model that would be appreciated as I'm not finding it. What I'm seeing on search is you were still using MN and (M + 1)N a year ago and that this didn't come up in last month's similar FAQ discussion.
Both MN and (M+1)N are correct when applied appropriately. I've added a "FAQ: How can I calculate effective aperture?" to try clarifying their use and interpretation.

--Rik

Photomicro
Posts: 109
Joined: Sat Aug 12, 2017 11:01 am
Location: UK

Re: Sharpness of microscopic photos at 10x?

Post by Photomicro »

A little off-topic in relation to the question, but this moss is quite delightful, and worth looking closer, at higher magnifications, see;

https://www.flickr.com/photos/66189529@N08/50969194091/

Mike
regards, Mike.

Time flies like an arrow, fruit flies like bananas.

https://www.flickr.com/photos/66189529@N08/

plantfan123
Posts: 21
Joined: Tue Dec 08, 2020 12:39 pm

Re: Sharpness of microscopic photos at 10x?

Post by plantfan123 »

Thank you *so much* everyone for the great advice. I've read it over a couple times and am gradually wrapping my head around all the concepts. Looks like the main thing was having the 105 mm lens focused at 1:1 instead of at infinity. I got a Nikon Nikkor Q 200mm lens (from the '60s!) today and it's perfect - the sharpness is great and the objective's light fills the frame focused at infinity.

Thanks Rik for the advice on Helicon Focus settings. I was using Method B. I'll experiment with other settings.

Still working on the lighting, colour and focus stacking settings but it looks so much better now with the 200 mm. This is a stack done with 100 photos in Helicon Focus (Method B, Radius 13, Smoothing 6).

https://www.dropbox.com/s/znddg6ioo1235 ... s.jpg?dl=0
plantfan123 wrote:
Sat Nov 27, 2021 12:18 pm
I am not sure what is going on with Plagiomnium cell walls optically but, having imaged them with a range of coupled lens configurations, it seems like they're always prone to chromatic aberration. I suspect the walls or intercellular material between then are prismatic and that this is challenging for the optics, along with maybe demosaicing from the sensor. There's possibly also some interaction with the continuous LED lighting I use. It's not unique to Plagiomnium—I've encountered similar behavior with hyaline awns of some other genera—but the plan fluors do control it better than the plan achromat I have. Olympus markets their UIS 2 plan fluors as semi-apo, which might be more about the NIR correction, .
Fascinating stuff! Lots to think about here.
palea wrote:
Sun Nov 28, 2021 9:07 am
If you don't have an autofocus rear lens then you can't autofocus bracket and are limited to focus bracketing with linear motion rails. For the most part autofocus bracketing is cheaper, faster, easier, and has smaller perspective effects than linear motion bracketing.
Also interesting! I'm using the Cognisys Stackshot rail. With the lens focus method can you do movements of a few microns that you can do with the Stackshot that are so useful in microscopy? How precise is it?
Photomicro wrote:
Mon Nov 29, 2021 1:36 am
A little off-topic in relation to the question, but this moss is quite delightful, and worth looking closer, at higher magnifications, see;

https://www.flickr.com/photos/66189529@N08/50969194091/

Mike
Thanks for sharing that! It's so beautiful.

Post Reply Previous topicNext topic