Choosing infinity objectives

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: rjlittlefield, ChrisR, Chris S., Pau

cueben
Posts: 30
Joined: Fri Sep 02, 2011 2:12 pm

Choosing infinity objectives

Post by cueben »

Hi,

I am considering the purchase of new Infinity Corrected objectives for use with my existing camera lenses (100mm and 70-200mm) as tube lenses. I'm trying to wrap my head around the technical details to consider. I need help confirming or correcting what I think I’ve figured out so far. If this thread turns out to be insightful, I’ll write up the conclusions for use in the FAQ section. Here are my notes:


- With the proper adaptors and step down rings, infinity corrected objectives can be attached to DSLR lenses from 100mm to 200mm without significant loss of image quality. The rated magnification of the objective is achieved on a 200mm while shorter focal lengths will result in proportionately lower magnification. E.g., a 20x objective attached to a 100mm lens will yield a 10x image. Using lenses with more than 200mm or less than 100mm works but quality may degrade.

- When attaching the objective to the lens, it's best to keep it very close so that the light goes through the center of the lens. This helps image quality in the image corners and makes sure to avoid vignetting.

- The resolution of detail in the image depends on the resolution of the objective and the magnification. E.g., a 20x objective that can resolve 1um detail on the subject would create details on sensor of 20um when used at 20x and of 10um at 10x. A camera with high pixel density (say 5um pixel pitch) should be able to capture all that detail. Therefore, shooting at 10x and cropping the image should give results very similar to shooting at 20x directly. For comparison, an expensive 10x objective that can theoretically resolve detail up to 0.6um on the subject could present an image onto the camera sensor with 6um detail. When used at 5x it could yield 3um detail on sensor. This is smaller than what most cameras can record. With that objective, shooting at 10x would yield more image detail than cropping an image shot at 5x.

- The depth of field of the objective does not change when changing magnification via a different tube lens. An advantage of using a 10x objective at 10x instead of a 20x objective at 10x is that you would need fewer images per stack.

- Combining the last two points, one should ideally use the objective with the least magnification that yields the desired subject detail resolution without exceeding the camera’s ability to record it. In other words, first try a longer tube lens before resorting to more expensive objectives with higher magnification and/or higher resolution.

- If the depth of field is less than 2um, it may be difficult to get consistent results with StackShot because although it's theoretical smallest step size if around 0.5um, in practice it starts to jitter randomly at small step sizes and some steps can exceed 1.5um even when configured for 0.5um. Note: there are many objectives with DoF in the range of 1.5um to 1.8um and I am hoping that these will work well even if sometimes running the stack a few times is needed to get all the source shots you need to stack without focus banding.

- A small working distance can make it very hard to use incident light on a solid (3D) object like an insect. Unless using flat subjects like microscopy slides, try to have at least 10mm of working distance (more is better).

- For use without slides, try to avoid objectives that expect oil or water between the objective and the subject. “Air” objectives are best and the image quality difference should be noticeable.

- Objectives optimized for use without slide covers are ideal but there usually is only a small difference in image quality when using objectives designed for use with covers.

- Many older microscope objectives rely on corrective eyepieces and will have degraded image quality when mounted on a DSLR lens. Many newer lenses do not rely on this and have good performance in this setup. This is sometimes hard to determine on the maker or vendor web site.

- Chromatic aberration correction is hard to predict before buying. Within the same brand, apochromatic objectives do much better than achromatic ones but they tend to cost more.

- The stacking process eliminates some of the CA because it is most visible in OOF areas. It does not help for spherical aberrations.

- Nothing replaces tests. There is a leap of faith whenever buying an objective without having seen it used in the work of others.

Also, I’ve not found many places that sell new objectives to individuals (as opposed to labs or schools). Edmund Optics seems reputable and I am considering it. Any suggestions about other sources are welcome.

Thanks,
-Ben

Some references (I didn't note them all but am adding as I find them again):

Aberrations:
http://www.nikon.com/products/instrumen ... /index.htm

Mounting microscope objective on DSLR:
http://www.photomacrography.net/forum/v ... hp?t=12147

Image quality comparisons for various objective options:
http://www.photomacrography.net/forum/v ... hp?t=15876
http://www.photomacrography.net/forum/v ... php?t=9393

Similar thread about choosing infinite objectives:
http://www.photomacrography.net/forum/v ... hp?t=15844

Blame
Posts: 342
Joined: Fri May 14, 2010 11:56 am

Post by Blame »

You don't say what your camera is.

If it is a cropped sensor like the canon 550D or nikon D7000 then your 100mm lens should do the trick. No promises. If it is a Full frame sensor like the canon 5D you will need something longer.

Your 70-200mm will only work, at best, near to the 200mm end.

It is worth remembering that the light path is different when a camera lens is used as a tube lens. The center of the image is the same but corners may be dark or blurred.

At low powers you can expect the objective and the camera to both significantly contribute to the blur. Cropping is not optimal.

Try to forget about the magnification rating of a lens. The actual magnification is dependent on the combination of tube lens and objective. What counts is the NA, the field of view (subject side as limited by loss of image quality) and the working distance.

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

Re: Choosing infinity objectives

Post by rjlittlefield »

Ben, this will be a long reply but I'd like to address each of your points as you asked.
With the proper adaptors and step down rings, infinity corrected objectives can be attached to DSLR lenses from 100mm to 200mm without significant loss of image quality. The rated magnification of the objective is achieved on a 200mm while shorter focal lengths will result in proportionately lower magnification. E.g., a 20x objective attached to a 100mm lens will yield a 10x image. Using lenses with more than 200mm or less than 100mm works but quality may degrade.
Yes, with some caveats. You're pretty safe at 200 mm with an APS size sensor (around 15mmx23mm). Going to full frame or using a shorter tube lens stresses the objective by forcing it cover a wider field. If you drop down as low as 100 mm the corners will almost always suffer. The Nikon CFI Plan Achromat 10x NA 0.25 WD 10.5mm (part number MRL00102) works great at 5X on an APS sensor, using a 100 mm tube lens. Most other objectives will have bad corners under those conditions.

Also, as mentioned by Blame a zoom telephoto will work well only at or near its maximum length. At shorter settings it will vignette.
When attaching the objective to the lens, it's best to keep it very close so that the light goes through the center of the lens. This helps image quality in the image corners and makes sure to avoid vignetting.
Yes, especially about the vignetting. In some cases mounting very close may give a little loss of contrast due to reflections between the objective and the front element of the tube lens. This can be minimized by flocking or masking all parts of the objective except the actual glass.
The resolution of detail in the image depends on the resolution of the objective and the magnification. E.g., a 20x objective that can resolve 1um detail on the subject would create details on sensor of 20um when used at 20x and of 10um at 10x. A camera with high pixel density (say 5um pixel pitch) should be able to capture all that detail. Therefore, shooting at 10x and cropping the image should give results very similar to shooting at 20x directly. For comparison, an expensive 10x objective that can theoretically resolve detail up to 0.6um on the subject could present an image onto the camera sensor with 6um detail. When used at 5x it could yield 3um detail on sensor. This is smaller than what most cameras can record. With that objective, shooting at 10x would yield more image detail than cropping an image shot at 5x.
Yes.
The depth of field of the objective does not change when changing magnification via a different tube lens. An advantage of using a 10x objective at 10x instead of a 20x objective at 10x is that you would need fewer images per stack.
Sort of. The NA of the objective does not change when changing magnifications. The depth of field may or may not change, depending on your definition of DOF. If you're using the standard definition involving blur circles, then DOF definitely will change because the size of the blur circles scales with magnification. If you're using the stacker's definition of "step size needed to avoid focus banding", then that will change too, but typically a lot less. In your example, it's the smaller NA of 10X objectives that allows fewer images per stack.
Combining the last two points, one should ideally use the objective with the least magnification that yields the desired subject detail resolution without exceeding the camera’s ability to record it. In other words, first try a longer tube lens before resorting to more expensive objectives with higher magnification and/or higher resolution.
Sort of. In general, stack length is minimized by using the smallest aperture that still gives the resolution you care about. With microscope objectives, the aperture size generally goes along with magnification, so your advice makes sense. However, you can also adjust the effective aperture size by adding an iris or a fixed mask just behind the objective. Stopping down the objective in this manner increases the DOF and with some subjects can improve image quality by reducing the "fog" caused by OOF foreground elements.
If the depth of field is less than 2um, it may be difficult to get consistent results with StackShot because although it's theoretical smallest step size if around 0.5um, in practice it starts to jitter randomly at small step sizes and some steps can exceed 1.5um even when configured for 0.5um. Note: there are many objectives with DoF in the range of 1.5um to 1.8um and I am hoping that these will work well even if sometimes running the stack a few times is needed to get all the source shots you need to stack without focus banding.
Correct regarding that StackShot rail. If more precision in the micron range is required, it's easy to buy also a separate motor to use with the StackShot controller, and connect that to drive the focus block of a microscope. See HERE and HERE.
A small working distance can make it very hard to use incident light on a solid (3D) object like an insect. Unless using flat subjects like microscopy slides, try to have at least 10mm of working distance (more is better).
Yes.
- For use without slides, try to avoid objectives that expect oil or water between the objective and the subject. "Air" objectives are best and the image quality difference should be noticeable.
- Objectives optimized for use without slide covers are ideal but there usually is only a small difference in image quality when using objectives designed for use with covers.
Sort of. Objectives at NA 0.10 and below can look through several mm of liquid with no problems. Objectives at NA 0.30 and below are not visibly affected by presence/absence of a 0.17mm cover slip. At NA 0.40, the cover slip makes a very slight difference in image quality. Above that the differences quickly become more significant. At NA 0.50 and above, you probably won't like the results if you use the objective differently from its design.
Many older microscope objectives rely on corrective eyepieces and will have degraded image quality when mounted on a DSLR lens. Many newer lenses do not rely on this and have good performance in this setup. This is sometimes hard to determine on the maker or vendor web site.
Correct, except that I'm not aware of any maker or vendor web site that does specify this information.
Chromatic aberration correction is hard to predict before buying. Within the same brand, apochromatic objectives do much better than achromatic ones but they tend to cost more.
I don't know what "hard to predict before buying" means. We have not noticed any variation in CA between different samples of the same model lens, but it does vary widely between makers and models. So, if a lens works well for somebody else then it probably will for you too, but if nobody has tested it then I don't know any way to predict reliably.
The stacking process eliminates some of the CA because it is most visible in OOF areas. It does not help for spherical aberrations.
Correct, but I'm not sure what point you're trying to make in mentioning spherical aberration. With infinity objectives, spherical aberration is a concern only if a) you focus the tube lens closer than infinity or b) you look through the wrong amount of glass or liquid. Other than that, you just have to accept whatever amount is designed into the objective, or you can reduce its effect by stopping down the objective as mentioned above with DOF.
Nothing replaces tests. There is a leap of faith whenever buying an objective without having seen it used in the work of others.
Amen.
Also, I’ve not found many places that sell new objectives to individuals (as opposed to labs or schools). Edmund Optics seems reputable and I am considering it. Any suggestions about other sources are welcome.
Yes, this is definitely a problem, especially for major brands like Nikon. Optics Planet does offer Nikon CFI BE, but other than the CFI BE 4X NA 0.10 (HERE), I'm not aware of tests of the CFI BE series.

--Rik

cueben
Posts: 30
Joined: Fri Sep 02, 2011 2:12 pm

Post by cueben »

Thank you very much for all the information. It's very useful to clear the uncertainty and doubt. It might be several days before I have time to push further (work, travel, etc) but this is certainly an excellent boost.

And to answer Blame's question, I am using a Canon 7D for macro work.

-Ben

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

Post by rjlittlefield »

One more clarification... Stacking reduces color fringes caused by longitudinal color aberration, which are most visible in OOF regions. But it does not reduce color fringes caused by lateral (transverse) color aberration, because those are obvious in focused regions.

--Rik

Blame
Posts: 342
Joined: Fri May 14, 2010 11:56 am

Post by Blame »

Nice camera.

100mm should do it. Less is possible. It is worth noting that 85% of the area of the sensor is within a circle with diameter equal to the sensor's width and only about 80% of the diagonal.

Remember to put everything in manual including white balance, focus, aperture and iso. You don't want the camera changing its mind halfway through a stack.

cueben
Posts: 30
Joined: Fri Sep 02, 2011 2:12 pm

Post by cueben »

Hi,

Finally got some time for photography again and I am still trying to decide which 10x and 20x infinity objectives to buy. Using the forum search, I didn't find any information about the Edmund Optics "High Resolution" series of microscope objectives. Does anyone have experience with them?

They seem to offer a tradeoff: higher resolution than the Mitutoyo but not as well corrected for aberrations. Theoretical resolution is listed on the specs and I hope is correlated with actual resolution enough to allow comparison of different models. For aberrations, the Mitutoyo is "Apo" and the EO isn't which does not really help me figure out how much difference to expect in terms of performance for applications typical of these forums such as insect photography.

Specs for the EO:
http://www.edmundoptics.com/products/di ... uctid=2674

Specs for the Mitutoyo Apo:
http://www.edmundoptics.com/products/di ... uctid=1942

And yes, I saw there is a Mitutoyo 10x High Resolution Apo but that is pushing past my price range. Right now I'm kindof thinking of getting the Mitutoyo 10x Apo and the the EO HR 20x so I can push for cleanest images or highest resolution depending on the situation. Any advice or links are appreciated since it feels like a gamble and these things are not cheap.

Thanks,
-Ben

Charles Krebs
Posts: 5865
Joined: Tue Aug 01, 2006 8:02 pm
Location: Issaquah, WA USA
Contact:

Post by Charles Krebs »

Ben,

I believe that the Edmund "HR" objectives are designed to produce an image "circle" (at design magnification) much smaller in size than than the Mitututoyo (and current Nikon and Olympus) objectives. If I am not mistaken, these are intended to be used with devices that utilize sensors up to about 2/3" (diagonal of 11mm). I would be very concerned that they could provide an image that would be useful on an APS-C sized camera.

A few additional thoughts to your questions above:
rated magnification of the objective is achieved on a 200mm
Remember, this is not "universal". Nikon, Mitutoyo, and Leica "infinity" were designed for use with a 200mm tube lens. Zeiss is 164.5mm Olympus is 180mm. Motic is 180mm. Leica and Zeiss do some chromatic correction in the tube lens, the others do not. (Older Olympus infinity utilized corrective eyepieces).
E.g., a 20x objective that can resolve 1um detail on the subject would create details on sensor of 20um when used at 20x and of 10um at 10x. A camera with high pixel density (say 5um pixel pitch) should be able to capture all that detail.
For several reasons it is usually best to have around 3-3.5 camera pixels across the smallest detail as seen on the sensor if you want to "capture all that detail". (But even if you only have two, it should still look pretty good!)
Many older microscope objectives rely on corrective eyepieces and will have degraded image quality when mounted on a DSLR lens. Many newer lenses do not rely on this and have good performance in this setup. This is sometimes hard to determine on the maker or vendor web site.
This is perhaps more the case with finite tube length objectives. The original Olympus LB infinities (1981-1993) did used the same corrective eyepieces as their "biological" LB finite brothers, but those made in the past 20 years do not. (I don't know the status of even older infinities like American Optical or Reichert in this regard). Other than that I don't know of any infinity systems that used corrective eyepieces. But one thing to remember about infinity style optical systems is that these corrections can also be accomplished in the tube lens (as is currently the case with Zeiss and Leica), and that info is not easy to determine for older infinity systems.
(As for eyepiece correction with the old timers, you can sometimes get a clue in an old catalog and if you check the recommended eyepieces offered at the time. If, in the eyepiece designation you find "C" or a "K" (such as WHK 10X) you know these are chromatically corrective).

Chris S.
Site Admin
Posts: 4042
Joined: Sun Apr 05, 2009 9:55 pm
Location: Ohio, USA

Post by Chris S. »

Ben, your query about Edmund Optics HR lenses raised some questions for me, as I use a range of Mitutoyo objectives from 2x-100x (none of them HR versions). Since Edmund Optics is a major dealer for Mitutoyo objectives, you got me to wondering how their "house brand" might compare.

So I called Edmund Optics and spoke with a very helpful support engineer. The engineer told me that EO Plan APO objectives (which are not the HR ones) are designed as direct alternatives to the corresponding Mitutoyo M Plan Apo objectives, at a bit less cost. He said that for these lenses, EO vs. Mitty MTF charts are mostly indistinguishable. (He'll be sending me some charts so I can have a look.) Notably, one objective in this series, the EO 100x, has a higher NA than the Mitty 100x (0.8 vs. 0.7), and in this case, he said, the EO out-resolves the Mitty. This engineer recommends considering the Edmund Optics Plan Apo objectives as nearly identical alternatives to the Mitutoyo M Plan Apos.

But the Edmund Optics HR objectives, (they cover 5x, 10x, and 20x) are indeed not apochromatic. They are corrected for 587 nanometers only. So where chromatic aberration characteristics matter, this engineer does not recommend the EO HR objectives. (I made this call before reading Charlie's reply, and did not think to ask about the image circle.)

Based on the engineer's answers and my own experience, I'd be very skeptical of purchasing EO HR objectives for most macro subjects, including insects. Chromatic aberration can be anormous hassle and detriment to top quality.

Also, I'm not sure if the additional resolution of an HR objective would make a noticeable improvement on current sensors. My Mitty 20x out-resolves my D200 sensor, and from recent discussions at this forum, I suspect it will out-resolve any APS-C sensor likely to appear for a long time to come.

Add to this the fact that 20x Mitty objectives regularly show up for sale used at discount prices, and the Edmund Optics HR objectives show up rarely if at all, and would likely have to be purchased at list prices.

Cheers,

--Chris

cueben
Posts: 30
Joined: Fri Sep 02, 2011 2:12 pm

Post by cueben »

Hi Charles and Chris,

Thanks you very much for your replies. The devil is in the details and some of the precisions you bring about pixels per smallest detail, image circle size, tube lens corrections, significance of CA for this type of work and explanation of EO products are all very useful. This will allow me to spend less money and get objectives that will work better for me.

For the theoretical resolution calculations, I think the Canon 7D I use has very small pixel pitch at about 4.3 microns. With the stated resolution of the Mitutoyo 20x at 0.7 micron, that gives 3.26 pixels per smallest detail when used with a 200mm tube lens. That seems like a very good match in equipment. It will be interesting to compare 100mm, 200mm and 300mm tube lenses with the 10x and 20x objectives to see how they performs in reality as opposed to theory. I understand that at 300mm, vignetting can be an issue but I'm still curious about cropping the center of the frame.

In the end, I think that with the Mitutoyo 10x and 20x, it's hard to go wrong. Like Chris said, they are quite common so there should not be any big surprises. Pixel peeping aside, they should be good enough objectives to point out my own technical (and artistic) limitations so that I can work on them without having nagging doubts about the results I could have had with better objectives. My dream is to make panoramas of stacks to make large prints and hang them in my dinning room :-). That should be challenge enough to keep me busy for a while.

-Ben

Post Reply Previous topicNext topic