Another myth exploded

[NikonUser]

In November 2014 I posted some images under the title "A Nikon myth exploded"

http://www.photomacrography.net/forum/v ... hp?t=25343

My statement also turned out to be somewhat of a myth in itself.

With the risk of again being incorrect I submit:

"A full-frame vs. objective myth exploded" without giving any evidence but basing it on hearsay.

I have read somewhere that microscope objectives can not be used on full-frame (36 x24 mm) sensors as the objective's image will not cover the sensor but give dark corners.

This image is from a Nikon CFN Plan finite 10x on 150mm extension; Nikon D610. MM ruler superimposed; straight out of the camera, no post processing except for size reduction.

OK, so is there some darkening at the corners but not enough, IMO, to negate using a full-frame camera with 160 finites.

[ChrisR]

Deleted - I misread the text,

[TheLostVertex]

...I have read somewhere that microscope objectives can not be used on full-frame (36 x24 mm) sensors as the objective's image will not cover the sensor but give dark corners...

Objectives have set fields of view they cover at their rated magnification. This varies from objective to objective. Some will work well with larger sensors, while some may not even cover the whole sensor of an apc-c sensor. There can also be a certain amount of flex between what the manufacture specifies and what you can get away with, which again varies between objectives. The only way to know really is to test

Most of the common issues preventing an objective from working on larger sensors would be vignetting, spherical aberration, and field curvature(not a huge deal for stacking AFAIK)

[Chris S.]

My read of NU's test: His extension of 150mm is standard for this optic, and measured subject field looks pretty close to the 3.6mm expected of a 10x objective on a full-frame sensor. So what NU showed here is what happens if you don't effectively crop the lens' output by limiting it to an APS-C sensor.

I have read somewhere that microscope objectives can not be used on full-frame (36 x24 mm) sensors as the objective's image will not cover the sensor but give dark corners.

I don't think dark corners are usually the issue. Rather, it's how well image quality holds up as you approach the corners. But as there are probably a lot of folks who think the corners go abruptly dark (I used to be among them), a demonstration like this certainly helps clarify things.

NU, any sense of how image quality compares near the corner of the full-frame, vs. the corner of an APS-C crop?

--Chris

[ChrisR]

Actually, which 10x is it - I'm assuming the NA 0.3? If so yes, I used one on a D700 a fair bit.
I have dim memories that the CFMplans (160 and 210 TL) are more even, also that the 4x and 10x apos are bad in the corners.
The 4x NA 0.13 I think is much more even than the 0.2 but less good centrally so I stopped using it.

And the 2/2.5x versions give a proper vignette - a circular image!

(The current(?) CFI 10x NA0.3 Fluor has a "FN" 25mm, which is pretty good, too.)

Stacking can lose the very corners because of misalignments and scale change - especially when I get the camera crooked.

[NikonUser]

Thanks everyone for the comments.
Yes, the Nikon CFN 10x NA 0.3.
Once again I was over-enthusiastic about this objectives ability to fill a 36x24 mm sensor with a tack-sharp image.

Even though a lot of detail is lost in the corners on the large sensor I still prefer the total image I get over the image from the 23.6x15.8 mm I get from my D90.

For the type of work I do, mostly bug images, lack of sharpness in the corners is rarely of any significance.

Top: fly wing showing the full-frame image with the 10x on a D610; superimposed is the outline of the image that I would get with the D90. Original image is 6016x4016 pixels reduced to 1024 for posting.

Bottom: right corner cropped to 1500x1500 px and then reduced to 1024px. Awful corner beyond the D90 frame; even the extreme corner of the D90 frame lacks sharpness.
But the full frame reduced to 1024px is still more useful to me than the image size from the D90.

[Rylee Isitt]

NU,

That first image... is it supposed to look that way? It looks corrupted to me, like two images superimposed in an odd way.

In addition to an objective that delivers a wide enough image, I would imagine that getting vignette-free results on a full frame would either require the right photo eyepiece (one with higher magnification and/or wide field of view), or a wide diameter tube if doing direct projection onto the sensor.

With direct projection I suspect you can also extend the tube past 150mm/200mm to "zoom" in on the image, and as long as your tube's inner diameter is wide enough, that will do the job. You should be able to do the same with infinity objectives by adding further extension after the tube lens. As far as I can tell this is no different than placing extension tubes on any lens... you get an enlarged image.

[Chris S.]

That first image... is it supposed to look that way? It looks corrupted to me, like two images superimposed in an odd way.

Rylee, you probably missed this in NU's text:

MM ruler superimposed.

So the strip in the center of NU's first image is a scale bar, demonstrating that the field of view is about 3.6mm, just as we'd expect from a 10x objective projecting on a 36mm-wide sensor.

In addition to an objective that delivers a wide enough image, I would imagine that getting vignette-free results on a full frame would either require the right photo eyepiece (one with higher magnification and/or wide field of view), or a wide diameter tube if doing direct projection onto the sensor.

With direct projection (which, if I understand correctly, NU is doing here), the extension tube needs to be chosen with a bit of care, but need not be unusually wide. Extension tubes that include electronic connections and aperture control are typically too narrow, but cheap manual extension tubes are fine. My PB-6 bellows will handle full frame without a problem.

With direct projection I suspect you can also extend the tube past 150mm/200mm to "zoom" in on the image. . .

NU specified that he used 150mm of extension, which is central to what he's demonstrating. He is not enlarging image beyond its normal 10x use. He's showing us that the image circle of this objective is big enough to cover a 24x36mm full frame--at least for some uses. Shooting it on a 16x24mm APS-C ("DX" in Nikon-speak) sensor simply crops it.

NU's images with the APS-C frame superimposed are really illustrative, and show why full-frame may be the most rational choice for one photographer, and APS-C most rational for another. Lenses differ in how far away from center their quality holds up, but this demonstration of NU's is not atypical. If one mostly shoots centered subjects, as NU does, corner sharpness may be relatively unimportant, compared with increased field of view. Personally, I do want sharp corners, and also want to place as many pixels as possible in the high-quality central portion of the image circle. Right now, the highest megapixel counts within the APS-C frame come from APS-C bodies, so they are mostly what I use with microscope objectives. But NU definitely makes his point.

Cheers,

--Chris

[Rylee Isitt]

Chris,

Makes sense... I typically assume that the biggest allure of full frame is the image quality - both in terms of resolution and image noise. But I can understand the allure of getting a wider field of view as well.

I have in the past shot a few images with less than 150mm extension, pushing 10x objectives down to 5x in some cases. So far, I've not noticed vignetting when doing this with a Nikon Finite Conjugate 10/0.25, and it's pretty good at the corners too. Although that's not the same situation as shooting at 150mm on a full-frame camera, it does suggest that it has good corners and a nice wide field of view.