Budget Friendly BH2 Video?

[ethicalhacker]

I did infer the following formula from the table at the bottom of the Introduction section here.

Used Image Sensor Diagonal Dimension / Photo Eyepiece Magnification = Field Number of Image Presented to Image Sensor

For example, 21.64mm (micro four thirds image sensor diagonal) / 3.3x (NFK photo eyepiece magnification) = 6.55 (field number). Of course, the WHK 10×/20 eyepieces have a field number of 20 for comparison.

I do not understand how the U-PMTVC affects the field number, etc. The following quote is from here

Reduction factors in C-mount adapters are normally quoted relative to a 1" sensor, so a 0.3× adapter would be used with a 1/3" sensor. However, in the specification of the MTV-3 (and the U-PMTVC), Olympus use 0.3× to refer to the reduction of the image circle of an NFK 2.5× eyepiece from approximately 48 mm (to safely cover 35 mm film) down to 16 mm for a 1" video sensor.

48mm x 0.3x is 14.4 - not 16mm. I really don't know what the formula is supposed to be.

[Tom Jones]

The bottom line is that you need an image circle at the sensor that is large enough to enclose the entire area of the sensor that is used to capture the image. If it's smaller, you get vignetting. You really have very few reasonable choices as to how to get there. You are limited to the available projection lenses unless you want to try something off-brand that might not have the appropriate color corrections built-in. That's it. Beyond that, you don't want any more crop factor than is necessary, but "necessary" is a personal, not scientific or photographic choice beyond making sure you are inside any inaccuracies in the lens geometry at the periphery. Some of the finest photomicrography on the web was done by Charles Krebs using an APS sensor Canon and the obviously mismatched NFK 2.5x projection lens.

I have all of the NFK projection lenses, including the 1.67x, and tried all of them before choosing what I did. YMMV.

You can always get a full-frame camera and use the NFK 2.5x. Any sensor cropping will fit into the image circle projected by the 2.5x. Unfortunately, full-frame 4K isn't generally "budget-friendly". Even the GH5 will more than double your camera cost.

Tom

[Scarodactyl]

A good speed booster above a 2.5x nfk would probably work well. Teleconverters are sometimes used in the opposite situation where a direct projected image is too small for full frame.

[ethicalhacker]

I've spent some time going over Alan Wood's stuff again. The field number quoted for a 2.5x photo eyepiece with a full frame image sensor is 17.3. Further down the page, the field number quoted for the 2.5x photo eyepiece (not relative to any particular sensor size) is 21.6. Since the 'real' field number for the 2.5x photo eyepiece is 21.6, one needs to multiply this number by 2.5 to get the real image diagonal of 54mm. A full frame image sensor has a diagonal of 43.3mm, which comfortably fits inside this beam of light. With a relative field number of 17.3 [(1/2.5x) * 43.3mm], one knows they're missing some of what can be seen through the [non-photo] eyepiece, but the field-of-view is pretty closely matched. In return, there shouldn't be any vignetting. Now let us consider what would happen if we attached a U-PMTVC with its 0.3x reduction lens. It takes the circular image with a 54mm diagonal from the previous stage and multiplies it by 0.3, which should give us a circular image with a 16.2mm diagonal. This matches Alan Wood's number, and the 15.9mm video camera image diagonal he mentions fits inside it.

Now that we've seen how the math worked with the 2.5x photo eyepiece, let's see how it should apply to the 3.3x photo eyepiece. Please correct me if my facts or math is wrong. The real field number quoted for the 3.3x photo eyepiece (not relative to any particular sensor size) is 16.7. When we multiply 16.7 by 3.3, we get a real diagonal of 55mm. Multiplying that by 0.3 gives us an image diagonal of 16.5mm, which is way smaller than the micro four thirds image sensor; however, as I previously noted, the Panasonic Lumix GH4 has a UHD 4K diagonal of 15.63mm, which definitely fits. This wouldn't be true by default on the GH5 though.

If I'm understanding the math correctly, it doesn't seem like the U-PMTVC affects the field number.

The image circle matches the sensor, and the eyepiece FOV very well.

Since you have all of the photo eyepieces, would you be so kind as to grab a UHD 4K frame with each photo eyepiece. I imagine centering that U-PMTVC is kind of a pain, so I'll be patient. You could resize the resulting frames down to something reasonable for posting here. If all of the frames were shot using the same objective and looking at a calibration slide and you told us the objective magnification, we should be able to know how much wider the field-of-view would have been had we been looking through your [non-photo] eyepieces.

[Tom Jones]

No, actually. I won't modify my photography systems to take pictures for you. And I have no inclination to check or comment on your math. I've explained what I use and why. Asking me to do testing for you so you can figure out why your math is confusing or misleading you is unreasonable, and not something I'm willing to do. That's your rabbit hole, not mine.

You asked for a "budget-friendly" solution and I believe I gave you one. I know it works just fine for me, and in the end that's all that really matters. I listened to some nice, experienced folks who came before me, and simply hooked it up the way they suggested and tried it. The particular configuration I use was suggested to me by a retired microscopy and photomicrography professor friend who has also written a couple of relatively recent books on the subject. I think the end results speak for themselves. But, if you want to reinvent the wheel, be my guest.

Tom

[hans]

...it doesn't seem like the U-PMTVC affects the field number...

https://www.microscopyu.com/microscopy- ... ld-of-view:

The diameter of the field in an optical microscope is expressed by the field-of-view number, or simply the field number, which is the diameter of the view field in millimeters measured at the intermediate image plane.

May help to think about projecting the sensor "backwards" using the inverses of the magnifications of the relay optics to find the equivalent size in the intermediate image plane. For Tom's setup, using your number for GH4 UHD 4K sensor diagonal, 15.63 mm * (1/0.3) * (1/3.3) = 15.79 mm FN.

[Hobbyst46]

...If I'm understanding the math correctly, it doesn't seem like the U-PMTVC affects the field number.

Quote from the same Alan Wood site:
"The MTV-3 produces a circular image approximately 16 mm in diameter, and to take full advantage of this a video camera with a 1″ sensor would be needed. The Nikon 1 J1 and V1 cameras have a CX format sensor with a 15.9 mm diagonal, and so they also appear to be suitable.

Reduction factors in C-mount adapters are normally quoted relative to a 1″ sensor, so a 0.3× adapter would be used with a 1/3″ sensor. However, in the specification of the MTV-3 (and the U-PMTVC), Olympus use 0.3× to refer to the reduction of the image circle of an NFK 2.5× eyepiece from approximately 48 mm (to safely cover 35 mm film) down to 16 mm for a 1″ video sensor.

The black U-PMTVC adapter (catalogue number U-V210, 037151) that is still available for the BX and other current Olympus microscopes can also be used on the BH-2. Like the MTV-3, it includes a 0.3× reducing relay lens. The only significant difference from the MTV-3 is that it provides some adjustment for parfocality between the video camera and the viewing eyepieces."

If the FN imposed by the MTV-3 is 16mm, and the PMTVC does not differ in this respect, then the FN of the PMTVC is 16mm, so it does determine the maximal field number.

[ethicalhacker]

I received enough parts to do a preliminary test with a NFK 3.3X LD photo eyepiece, a U-PMTVC adapter, a Fotodiox C-m4/3 adapter, and a borrowed Olympus OM-D E-M5 camera. My goals for this test were the following.

1. Measure the image diameter projected onto the image sensor.
2. See how closely the field-of-view matched that of the eyepieces.

There is significant vignetting when shooting in RAW (no image mode induced crop factor), which was to be expected when combing a C-Mount with a micro 4/3rds camera. The light beam hitting the image sensor was approximately 16.1mm in diameter (just clears the fringing). This means there was vignetting in the X axis and cropping in the Y axis. The 16.7 field number of the NFK was indeed the limiting factor for the field-of-view in this setup. When one looks through the NFK directly (without a U-PMTVC or camera attached) they will see the same field-of-view as they see when the U-PMTVC and camera is attached. This is what I meant when I said,

it doesn't seem like the U-PMTVC affects the field number.

I'm probably just not using precise enough terminology to communicate clearly. While this is less than the eyepieces, its certainly closer than I was expecting. The eyepieces and camera were parfocal.

[hans]

This is what I meant when I said,

it doesn't seem like the U-PMTVC affects the field number.

I'm probably just not using precise enough terminology to communicate clearly. While this is less than the eyepieces, its certainly closer than I was expecting.

I quoted the part about U-PMTVC not affecting field number to suggest it as the reason your math was not agreeing with Tom's experience. The original post before the edit implied a predicted field number around 6 mm for Tom's configuration. Paraphrasing from memory, something like "6.5 mm vs. 20 mm doesn't seem very well matched to me."

[ethicalhacker]

For the benefit of others that may find this post at some point in the future. I've produced the following renderings to demonstrate the approximate field-of-view one can expect.

This is what one would expect to see with a superwide SWHK 26.5 eyepiece (shown for context.)

2 SWHK 26.5mm.png (136.52 KiB) Viewed 5062 times

This is what one would expect to see with a standard WK or WHK 20 eyepiece.

1 WHK 20mm.png (136.77 KiB) Viewed 5124 times

This is what one would expect to see if they put their eye up to a NFK 2.5x photo eyepiece. (Please note that the field-of-view is slightly wider than the standard eyepiece. This is a paper calculation. I don't actually have a NFK 2.5x photo eyepiece, so I can't confirm whether a housing limitation will actually chop this down to a field number of 20 instead of 21.6.)

3 NFK 2.5x FN=21.6.png (137.21 KiB) Viewed 5124 times

This is what one would expect to see if they put their eye up to a NFK 3.3x photo eyepiece.

4 NFK 3.3x FN=16.7.png (135.87 KiB) Viewed 5124 times

This is what one would expect to see if they installed a NFK 3.3x photo eyepiece, U-PMTVC 0.3x adapter, and micro 4/3rds camera. The following assumes the entire sensor area is used (no image/video mode crop).

5 NFK 3.3x FN=16.7 U-PMTVC 0.3x m4-3.png (117.91 KiB) Viewed 5124 times

[ethicalhacker]

The original post before the edit...

Yes, I edited it to remove my error so as not to confuse other n00bs like myself that may read this later.

[tchall]

hi ethicalhacker
What did you feel was a cheap way to adapt a bh2
To take the closest photos to the eye piece
Many thanks TJ