Am I getting this right?

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

perdu34
Posts: 50
Joined: Fri Nov 25, 2016 7:03 am

Post by perdu34 »

Thanks for the detailed reply Rik.

I am going to have to give this a bit more thought and read up on telecentric optics again. In the mean time, because I've already bought the lens and objectives, I am going to carry on setting up my 10x system.

There's a lot more to this macro stuff than there seems.

cube-tube
Posts: 107
Joined: Tue Oct 10, 2017 7:50 am
Location: Durham, NC

Post by cube-tube »

There's a lot more to this macro stuff than there seems.
As I am sure many photographers on this forum would agree, it really never ends.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Heh! And just as you were posting "it really never ends", I was composing this... :)
cube-tube wrote:About setting up your stacking software to create a telecentric image: That should work with most objectives, including objectives that are fairly far from telecentric, correct? I am still not totally clear on that, because everything from foreground to background should be reproduced at the same magnification, but there could still be large foreground objects blocking portions of the background that would be visible (unobstructed) to a telecentric lens, so it doesn't seem like true telecentricity in that way.
Unfortunately the answers will be much longer than your questions. But let me try.

With a truly telecentric lens, the aperture is positioned so that the chief rays are all oriented parallel to the optical axis. At first thought this might seem to imply that the lens gets a clear shot at background in orthographic projection. But that's not correct. What actually happens is that the lens captures all the light in an entrance cone that is centered around each chief ray. Near a foreground edge, points in the background are seen by only about half of the entrance cone. For the other half of the entrance cone, the background point is hidden behind the foreground object. The image of the background point will consist partly of light from the focused background, with the remainder coming from the unfocused foreground. Thus the unfocused foreground inevitably contaminates the focused background, even with a truly telecentric lens. Remember this illustration from http://www.photomacrography.net/forum/v ... 042#135042 :
Image

With a lens that is not telecentric, the chief rays are not parallel to the optical axis. As a result, background points near foreground edges may be hidden from the chief ray. Nonetheless, the entrance cones are often wide enough that some significant portion of the aperture can see the focused background that is hidden from the chief ray, albeit with a different amount of contamination. In the diagram below, yellow marks the part of the entrance cone that can see straight behind the edge of the foreground object even though it is completely hidden to the chief ray.

Image

In the case shown here, the contamination will be more than a telecentric lens would give. But if the occluding foreground were on the other side of the entrance cone, it would be less.

So, for a lens that is sufficiently close to telecentric, the fact that some background is logically hidden from view does not really prevent acquiring a focused image of that "hidden" background.

To look at a specific case, let's consider that Nikon CFI BE 4X NA 0.1 objective that I mentioned earlier. The entrance pupil of that objective is about 42 mm away from subject. On an APS-C sensor with diagonal about 28 mm, the field diagonal at the subject will be about 7 mm. As a result, at the extreme corners of the field, the chief rays will be inclined about 4.76 degrees away from the optical axis (=atan(3.5/42)). That's quite a bit, but on the other hand the entrance cone of an NA 0.1 objective has a half-width that is about 5.74 degrees (=asin(0.1)). So even in the worst case, some fraction of the aperture will be able to see focused background in orthographic projection. It is, however, a small fraction, and because of that, this lens would not be very good for this use.

There is another consideration that also affects how far away from telecentric is acceptable. That consideration is the amount of scale change there is within the DOF slab at each focus position. With a truly telecentric lens, magnification is perfectly flat as a function of distance. But when a non-telecentric lens is forced to stack as if it were telecentric, the magnification in a focus stacked image becomes a sawtooth curve that on average is flat but that varies systematically within each focus step. As a result, features that are seen in focus at the near and far boundaries of adjacent steps are seen at two different scales in those two source images. It should be clear that if those scales are too much different, then there must be some sort of artifact in the stacked image.

The acceptable tolerance for this consideration depends to some extent on the stacking method that is being used. But for the sake of discussion let's say that the scale change within each focus step should be no more than 1 pixel across the width of the frame. Then if we're using a sensor that is 5000 pixels wide, we want a scale change of no more than 1 part in 5000 per focus step. Again considering that same objective, we conclude that the focus step should not be larger than 42mm/5000 = 0.0084 mm. This is quite a lot shallower than the nominal DOF of 0.055 mm for an NA 0.1 objective, so again we conclude that the Nikon CFI BE 4X would not be a good choice to use for faking telecentricity.

In contrast, I have just now tested a Mitutoyo M Plan Apo 10X NA 0.28 objective as having a scale change of about 1 part in 15,000 with a step size of 10 microns. Since the nominal DOF of this lens is even a little shallower than that, clearly the Mitutoyo 10X can be treated as telecentric with no problems at all. Similarly the Mitutoyo M Plan Apo 5X NA 0.14 shows about 1 part in 9000 at 40 microns step, so again, no problem treating that as telecentric.

--Rik

JH
Posts: 1307
Joined: Sat Mar 09, 2013 9:46 am
Location: Vallentuna, Stockholm, Sweden
Contact:

Post by JH »

Rik - thank you for the mitutoyo testing. One question - does the telecentric behavior of the mitutoyos significantly depend on the tube lens (raynox/itl200 etc) and or the distance between the tube lens and the mitutoyo?

Best regards
Jörgen Hellberg

Edit: Or a short instruction on how to do the testing so I can do it myself.
Jörgen Hellberg, my webbsite www.hellberg.photo

cube-tube
Posts: 107
Joined: Tue Oct 10, 2017 7:50 am
Location: Durham, NC

Post by cube-tube »

Thanks for the detailed explanation, Rik. I have another question, though: Do you know of any deconvolution software available to hobbyists? A while ago I started looking into open source deconvolution software, but honestly, the math was over my head. Convolution and "contamination" of the entrance cone is something that I have struggled with. I have also thought about trying to set up a light sheet system, but that would be really tough with opaque specimens like insects.

Image

People here post good images of complicated subjects all the time, so I feel like I'm missing something.

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

JH wrote:Rik - thank you for the mitutoyo testing. One question - does the telecentric behavior of the mitutoyos significantly depend on the tube lens (raynox/itl200 etc) and or the distance between the tube lens and the mitutoyo?
Telecentricity depends on the apparent position of the limiting aperture, which depends on the physical position of the limiting aperture plus the optics in front of it. Nothing behind the limiting aperture will affect it. Unless you add an aperture just behind the objective, then the limiting aperture is provided by the objective itself, so nothing in the tube lens setup will have any effect on telecentricity.
Edit: Or a short instruction on how to do the testing so I can do it myself.
Set up a flat target with detail all across the frame. I use a piece of paper covered with small toner dots, created by running a piece of light gray ink-spitted paper through a copy machine, glued to a glass slide.

Shoot that target with focus shifted by any convenient amount above and below perfect.

Load those two pictures into Zerene Stacker, set Options > Parameters > Alignment to Default, and Stack > Align All Frames. Then Options > View Console Log, and near the bottom of the log find the line that says something like this:

Code: Select all

Aligning, finalFit = RegistrationParameters: xoffset = -1.5519481019041215E-4, yoffset = 6.43139093306766E-5, scale = 1.0001148993820153, rotate = -1.7883162134191277E-5
It is the scale value that you want. In this case the change in scale is from 1.0 (implicit) for the first frame to 1.0001148993820153 for the second frame, or computationally 1 part in 8703. This is the value that I reported as "about 1 part in 9000" for the Mitutoyo 5X.

It is important that the target have focused detail over the entire frame. Otherwise the measurement becomes more sensitive to apparent changes in scale due to the way things change in appearance with focus, versus actual changes in scale due to lens geometry. This is also the reason why I recommend turning off scale correction when working with high mag objectives, to prevent the scale "correction" from introducing more problems than it solves.

--Rik

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

cube-tube wrote:Do you know of any deconvolution software available to hobbyists?
See the discussion at http://www.photomacrography.net/forum/v ... hp?t=33724 , and search the forum for other discussions of deconvolution.
Convolution and "contamination" of the entrance cone is something that I have struggled with.
If you ever solve that problem, I would be interested to know how. The big problem is that the amount of contamination varies from point to point and depends on a plethora of different aspects: depth distance between foreground and background, relative colors and brightness of foreground and background, and the exact 3D PSF of the lens, which can vary across the field. In principle what you're trying to do is figure out what geometry and surface coloring of the subject is most consistent with the stack of photos that you actually have in hand, given what you know about the image acquisition process. It is a very challenging problem.
I have also thought about trying to set up a light sheet system, but that would be really tough with opaque specimens like insects.

You might consider instead to build a system around the principles of confocal microscopy. The idea there is to concentrate illumination on the point that you're trying to image. It does not completely eliminate the contamination problem, but even a good DIY attempt might cut it way down.

--Rik

cube-tube
Posts: 107
Joined: Tue Oct 10, 2017 7:50 am
Location: Durham, NC

Post by cube-tube »

build a system around the principles of confocal microscopy
Ha! I'm sitting 20 feet from a confocal microscope right now.

Not mine though :D

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

cube-tube wrote:Ha! I'm sitting 20 feet from a confocal microscope right now.
Lucky man! Will they let you work with it during off hours?

--Rik

JH
Posts: 1307
Joined: Sat Mar 09, 2013 9:46 am
Location: Vallentuna, Stockholm, Sweden
Contact:

Post by JH »

Rik
Thank you very much!
Best regards
Jörgen Hellberg
Jörgen Hellberg, my webbsite www.hellberg.photo

perdu34
Posts: 50
Joined: Fri Nov 25, 2016 7:03 am

Post by perdu34 »

rjlittlefield wrote:build a system around the principles of confocal microscopy
I may be getting the wrong end of the stick here but I would think this is a really bad idea for macrophotography.

If you mean just blocking out a lot of light so that only a small part of the subject is illuminated then the "quality" of the light wouldn't be great as there is no diffusion. Get the angle even a little wrong and specular highlights would kill any image.

If you mean illuminating the subject through the objective then specular highlights are all but certain. And there is no way to separate the input light from the reflected light as it's not possible to use a dichroic mirror with "normal" photography.

This, I think, is why macro is so hard. There is a LOT of work been done over the decades to make microscopes work and produce stunning images with fluorescence and transmitted light for science. There's been well over 100 years of photography perfection too get the rules that apply to art. Then here we are, trying to apply art principals to science equipment.

cube-tube
Posts: 107
Joined: Tue Oct 10, 2017 7:50 am
Location: Durham, NC

Post by cube-tube »

Will they let you work with it during off hours?
Probably!

Unfortunately, it has been broken since I started working in this lab, but we just got some replacement parts so it should be fixed within the next few days.

Is that kind of thing relevant to this forum? I sort of think of this as a place to post about home studios and DIY setups, but if I get to spend some time with the confocal maybe I will post some images.

Alan Wood
Posts: 382
Joined: Wed Dec 29, 2010 3:09 pm
Location: Near London, U.K.
Contact:

Post by Alan Wood »

cube-tube wrote:
I have also thought about trying to set up a light sheet system, but that would be really tough with opaque specimens like insects.
I remember a light sheet system, probably 1970s or 1980s, long before digital, stacking, lasers and confocal, that was used to photograph insects.

There were amazing photos in magazines, sharp from head to tail but with odd perspective. I was using Olympus OM bellows macro lenses at the time, often at f/16 for depth of field, and I was envious of the sharpness and depth of field.

I wish I could remember how it was done. I have not been able to find any references.

Alan Wood

cube-tube
Posts: 107
Joined: Tue Oct 10, 2017 7:50 am
Location: Durham, NC

Post by cube-tube »

Oh man I would love to see that... was it in full color?
It definitely makes me more motivated to try it.

JH
Posts: 1307
Joined: Sat Mar 09, 2013 9:46 am
Location: Vallentuna, Stockholm, Sweden
Contact:

Post by JH »

perdu34
I would like to return to your initial questions. You said that this was 5x and that you usually used f4 and stacked 30-60 pictures and that the source images was sharper than the stacked pictures. Could you please post a side by side comparisson at 100% that shows the difference you are talking about?

Best regards
Jörgen Hellberg
Jörgen Hellberg, my webbsite www.hellberg.photo

Post Reply Previous topicNext topic