Mitutoyo objectives on full frame -- teleconverters too

[Chris S.]

Following up on questions raised in other threads, I tested the Mitutoyo M Plan Apo series objectives on a Nikon D700 full-frame (36 x 23.9 mm) digital SLR. The trial was of interest because it was not known to what degree, if any, these objectives would vignette on a full-frame sensor. It is known that they perform well on crop frame sensors, such as the D200 I normally use, which has a 23.6 x 15.8 mm sensor).

The short answer is that every objective in this group vignettes on full frame, as demonstrated below.

I also tested some Mitutoyo objectives in conjunction with teleconverters placed between camera and tube lens. Again, the short answer is that vignetting occurs with both 1.5x and 2x teleconverters, for the lenses I tested. Images below.

For all images: Nikon D700 camera with Mitutoyo tube lens (focal length 200mm). Each is a single unstacked image. No image is cropped. For the test subject, I used 4 point type from a laser printer (much smaller than the 10-12 point type in common use, lest anyone get a wrong impression about the field of view).

Objectives were:
Mitutoyo M Plan Apo 5x/0.14
Mitutoyo M Plan Apo 10x/0.28
Mitutoyo M Plan Apo 20x/0.42
Mitutoyo M Plan Apo 50x/0.55
Mitutoyo M Plan Apo 100x/0.70


2x Mitutoyo



5x Mitutoyo



10x Mitutoyo



20x Mitutoyo


50x Mitutoyo


100x Mitutoyo



With the addition of teleconverters:

2x Mitutoyo plus 1.5x Kenko teleconverter



5x Mitutoyo plus 1.5 x Kenko teleconverter



2x Mitutoyo plus 2x Kenko teleconverter



5x Mitutoyo plus 2x Kenko teleconverter



20x Mitutoyo plus 2x Kenko teleconverter


--Chris

[ChrisR]

Thanks for doing the test.

So why does this (unconverted) setup vignette when Morfa's 172mm(?) tube lens, and some telephotos, don't?
Apart from the 2x image, the vignetting appears to be the same for the 5x, 20x, 50x. I think that says the vignette is occurring somewhere other than in the objective.
That one's testable, if you can hold the subject, objective and tube lens in place and remove the other metalwork, and see how big the projected image comes up?

It looks like the converter(s) are adding their own vignetting, as it's pretty constant.
I think that means that the exit pupil of the tube lens is too far away from them.
Just how annoying is that?

[RogelioMoreno]

Chris,

What is the inside diameter of the adapter (the one that we can see the thread) that is before the adapter I on the last picture that you posted on the following link:

http://www.photomacrography.net/forum/v ... 0&start=15

Maybe that adapter is causing the vignette.

Rogelio

[rjlittlefield]

I agree with ChrisR. This looks like physical vignetting in the tubes and TC near the sensor.

Another test: illuminate the subject, remove the camera, and look down the bore. Slide your head sideways until light from the objective gets cut off. What does it get cut off by?

--Rik

[ChrisLilley]

The fact that vignetting is not observed on a 1.5x crop sensor, but is observed on an FX sensor even with a 2x teleconverter, does seem to point to the tube lens/adaptors/TC as the source of vignetting rather than the objectives themselves.

I would worry slightly about putting the light through a C-mount, in a setup intended for FX use.

[Chris S.]

I think you gents are right.

I only had a few minutes today when I was not attached to my desk, but finally did sneak down to the “Brat Cave” (as my brother-in-law started calling the photo area of my basement). I shot some images of a plain white background with just the tube lens, then with the tube lens and each teleconverter—no objective lens. All three arrangements vignette. I can post pictures, later, if anyone wants to study the size and shape of the vignetting. (If so, just say the word.)

Why didn’t I think to start with this simple test? Doh!

That one's testable, if you can hold the subject, objective and tube lens in place and remove the other metalwork, and see how big the projected image comes up?

Chris R., too bad the “optical erector set” that came up on eBay a few weeks ago went for far more than I could afford. It might have had the doodads to set this up easily. Right now, with spring busting out all over, the thought of making a jig to hold this stuff in place without using the parts I have to hold it in place makes me go “Ack!” Might be worth doing, though.

Rogelio, the focusable T-tube has a clear aperture of 33mm. For comparison, the regular T-tube has a clear aperture of 37mm.

Rik, it appears to me that the camera-side of the C adapter is the visible bottleneck.

Chris L., the above sounds just like what you suggested. I’ll point out that this C-adapter is the very part specified by Mitutoyo for holding this tube lens, though this of course doesn’t mean it’s optimal, or that the lens can't be pushed beyond spec with another mount. The C-adapter does block the periphery of the tube lens—going from memory, I’d guess 2 mm. (Taking a measurement would require complete disassembly of the setup, which I don’t right now have the time to do.) Would that lost 2 mm (or 4 mm, if one considers diameter, rather than radius) prevent the vignetting? Would the quality be OK way out at the edge of the tube lens? I doubt any of us could say.

So, ugh, back to Chris R.’s suggested test, or even an upgraded version of it? But don’t anyone hold their breath waiting for me to get to it, as spring is here. . . .

Cheers,

--Chris

[Horstl]

....Why didn’t I think to start with this simple test? Doh! ...

Chris,

only the 2x lens shows vignetting on 24x36mm format:
http://www.photomacrography.net/forum/v ... 7&start=15

Even on 6x4,5cm medium format with a 210mm lens as tube lens there is only corner vignetting (5x/20x).

However, beside vignetting, how are you satisfied with the quality?
In my opinion these lenses are really good (nearly perfect) up to around 20mm. Up to ca. 30mm they may be good (depends on sensor resolution and personal acceptance criteria), beyond 30 mm they can be used, but performance is far from what it should be (if the subject is in the center and the borders are not critical, this may be acceptable).
You have a very nice setup now, before worrying or changing anything I recommend you another test: Leave the optics as they are and try the lower magnifying lenses with a higher resolution APS-camera (16-18 MP). The D700 makes sense with the 50x and 100x lens but is not a good match for the lower magnifications.

Horst

[morfa]

The D700 makes sense with the 50x and 100x lens but is not a good match for the lower magnifications.

Horst

Judging from my own experiences as well as the results shown by fellow forum members with, for instance, the Mitutoyo 10x used at (or below) it's rated magnification on a 24x36 sensor (5D or 5DmkII) I have to disagree with this particular statement.

In fact, I consider the 10x Mitutoyo the best lens I've used on a full frame sensor at magnifications in the range 7-14x. And, in my current opinion, a full frame sensor makes even better use of the image delivered by this objective than a smaller sized sensor.

Perhaps I should point out that I'm basing this purely on results from setups incorporating other tube lenses than the original Mitutoyo one(s). I have no experience with Mitutoyo tube lenses but – so far – I don't recall seeing any really convincing results from it on a full frame sensor.

[ChrisLilley]

You have a very nice setup now, before worrying or changing anything I recommend you another test: Leave the optics as they are and try the lower magnifying lenses with a higher resolution APS-camera (16-18 MP).

I would tend to agree. Given the excellence of the setup from F-mount forward to the subject, the use of a D200 is looking like a weak link.

D200 is a really well made camera, but a full magnesium shell and weather-sealing isn't really showing a benefit as your studio setup is not really exposed to mud, rain and banging on rocks. The sensor, on the other hand, is of the 10Mpx CDD generation that eventually worked its way down to the D40x/D60 and was then discontinued. Same for the electronics (although Nikon calls all of their electronics EXPEED so they are hard to tell apart).

The D200 was replaced by D300 and D90 (very similar sensors, with the D300 being more optimised for faster frame rate and the D90, surprisingly, having slightly better dynamic range than the D300. D300s was a minor updated with essentially the same results as D300. This was the 12Mpx CMOS generation.

Then the D90 and perhaps the D300s were replaced by the D7000, with a partial mag shell, mirror up (missing from D90), 100% viewfinder, a live-view that does not uselessly bang the mirror twice right before the shot (!) and the 16MPx CMOS sensor and improved electronics. Despite the increase in pixel count, the noise and dynamic range is not worse then D90, and is better at the lowest ISO.

I suggest comparing D60 (D200 is to old for the DxOMark tests), D300 and D7000 at the DxOMark site (the forum software rejects the unshortened url: results ) and for archiving:

image credit: DxOMark

There are of course disadvantages to this approach - the D7000 uses SDXC and SDHC rather than CF, and uses the small rectangular port for a shutter release instead of the circular port that your D200 and D700 use. But given the overall quality and investment represented by the rest of the setup,this is in my opinion an approach worth considering.

[ChrisLilley]

Chris L., the above sounds just like what you suggested. I’ll point out that this C-adapter is the very part specified by Mitutoyo for holding this tube lens, though this of course doesn’t mean it’s optimal, or that the lens can't be pushed beyond spec with another mount.

Yes, so it was the correct part to try, and its sufficiently far forward that it might have been expected to work: but now you have reported vignetting that would be my suspect. In fairness, Mitutoyo expect the image goes to a 24mm or 30mm field diameter eyepiece, not a 43mm diameter FX sensor.

The C-adapter does block the periphery of the tube lens—going from memory, I’d guess 2 mm. (Taking a measurement would require complete disassembly of the setup, which I don’t right now have the time to do.) Would that lost 2 mm (or 4 mm, if one considers diameter, rather than radius) prevent the vignetting? Would the quality be OK way out at the edge of the tube lens? I doubt any of us could say.

I agree that we don't have the data to answer those questions, short of experiment. It could be that unsuitable parts of the tube lens are being deliberately shielded. On the other hand, they could have just selected a standard coupling whose size didn't look as if it would interfere for their assumed usage. I doubt a Mitutoyo microscope uses a C-mount anywhere internally.

[Horstl]

...
In fact, I consider the 10x Mitutoyo the best lens I've used on a full frame sensor at magnifications in the range 7-14x...

Not the lens is the problem, the sensor is it. The D700 has enough resolution for the 50x and 100x lens (and maybe the 20x) but not for the lower ones.
Putting these lenses (especially with a shortened tube lens) on a 12Mpix 24x36mm Sensor is -in my opinion- a waste of glass. Using them at 14x is a much better idea - with a 5DII. With a D700 I would recommend tube lenses in the range of 300 - 400mm.
Chris has the complete line of this fine lenses, why should he not use them in the way they are designed for? He can take the 5x instead of the 10x when a larger field is preferred - at the same time handling is much easier with the lower n.A., resolution is still sufficient for todays highest resolving APS-Sensors. For the price difference between a (second hand) APS camera and a 24x36 camera you can nearly get the whole set of lenses - 2x + 5x + 10x instead of only one which you have to abuse whenever it does not fit well.
I don't have the 10x lens, but the 2x and 5x and I'm not happy with them on the 5DII. This may sound like (and is for sure) a luxury problem, but there is a mismatch between optic and sensor. When the Sony Nex7 is coming (which is rumored to have 24 Mpix on APS) then I will try these lenses with a 180 mm tube lens on such a camera. This will take the full well corrected field that these optics can provide without sacrificing resolution. And I really doubt that there will be a way in the future to get significant more out of these objectives, whatever sensors are coming.

Horst

[seta666]

Not the lens is the problem, the sensor is it. The D700 has enough resolution for the 50x and 100x lens (and maybe the 20x) but not for the lower ones.

Sorry but I do not agree, when using the mitutoyo 5X and 10X at the designed magnification efective f number is f18 so the D700 sensor has more than enough resolving power. In fact I have used both the EOS 5D and the EOS 5DmkII; the 5D classic had better per pixel sharpness, I never use the full 21mpx of the 5D mkII but I go down to 17mpx with the 5X 10X and 20X and down to 10mpx with 40X
with the 100X lens 2mpx would be more than enough as diffraction severely affects resolution, with an effective aperture of f70
http://www.cambridgeincolour.com/tutori ... graphy.htm
The EOS 5d is difration limited from f13, EOS MKII from f10 and EOS 7D from f6.8
http://www.the-digital-picture.com/Revi ... eview.aspx
Also I have used APS-C cameras (1000D 40D) and Fullframe (5D and 5D mkII) I like output from fullframe cameras muchmore; I can not wait to see how will perform next generation fullframe cameras from Canon Nikon. Also the New APS-C D7000 and k7 have to work very well with that beautifull 14EV APS-C sensor

Now the lenses; I have used both the mitutoyo 5X and 10X (and have become my favorites)with couple of different tube lenses (morfanon and SOMCO 240mm)and image circle covers full frame very nicely (there is some vigneting on the somco but surely due to the lens frame), as morfa says you can safely use the mitu 10x down to 7X in most cases
This week I have to get the mitutoyo 20X (now is held in customs) and I hope is as good as this other two lenses
Also 2 teleconverter I have tested give some corner vigneting; I solved this by adding a short extension tube after the Teleconverter. As morfa has proved many times infinity focus is not that important
Regards
Javier

[yeatzee]

. Also the New APS-C D7000 and k7 have to work very well with that beautifull 14EV APS-C sensor

Regards
Javier

Pentax K-5 is what you mean Both share that awesome new sony APSC sensor.

I own the Pentax K-7 which I love, but it can't touch the K-5's sensor.

(granted the K-7 is half as expensive)

[Horstl]

...In fact I have used both the EOS 5D and the EOS 5DmkII; the 5D classic had better per pixel sharpness,...

That indicates that the 5D has not enough resolution. The sensor only resolves at a level where the optic is in high MTF-regions. For full resolution you also have to sample in the region where the MTF is falling. Of course this will be seen when looking at pixel level. But looking at pixel level into a picture that has twice the pixel count also means looking at a picture that is twice as large.

.... I like output from fullframe cameras muchmore; I can not wait to see how will perform next generation fullframe cameras from Canon Nikon.

Regarding the next generation fullframe cameras my wishes my sound maybe a little respectless: how about a usable jpeg-engine in the 5DIII? Or a dust protection system that protects also the sensor and not only camera sales? Or a fullHD live-picture/video output per HDMI? Or a intelligent multiframe sensor as Panasonic has shown? Or simply a tilt-able display? Or a reliable liveview/live-histogramm prediction system with every lens?
There is a lot of basic homework that could be done before entering next level (stratospheric?) image qualities.

Horst

[morfa]

Thanks for the clarification Horst – it seems I misunderstood what you meant.

The phrase "Mitutoyo 10x is a waste of glass on a D700" can be read in two ways:

I) If you have a D700 then it's a waste to get a Mitutoyo because there are other (cheaper or in other ways less wasteful) alternatives that should be able to give equally good or better results with this sensor.

I would disagree with this but I'm now fairly certain that this is not what you're saying.

II) If you have a Mitutoyo 10x and are looking to buy a camera body to go with it, then choosing a D700 would be a waste of glass.

This may be true. At least in some sence. But for instance, if the buyer only wants to consider F-mount alternatives, the D700 may still be the best (least wasteful) alternative. I'm not saying it is, I'm just saying I wouldn't bet against it until I saw some hard evidence (pixels please; tables and numbers have long ago lost my respect in image quality matters). At this point I have seen some very impressive images (from a pixel-peeper perspective) taken with classic 5D and Mitutoyo 10x (at 10x or below). When I start to see some equally impressive images from smaller sensor cameras your statement will seem more convincing.

Besides, I think we have this fundamental, ideological dichotomy to keep in mind. Which is to be preferred:

A) Cover the entire sensor with a perfect image circle at the expense of not capturing some parts of the perfect image circle (and sometimes be running the risk of oversampling).
B) Capture as much as possible of the perfect image circle at the expense of at the same time capturing some less than perfect parts of the image circle (and sometimes be running the risk of undersampling).

Some (most?) seem to prefer A and others (like me) prefer B. Obviously, A-people tend to advocate APS-C sized (or smaller) sensors while B-people prefer FF.

how about a usable jpeg-engine in the 5DIII?

Could you elaborate on this, please? Why do we have reason to believe it would be unusable?