Exploring photomicroscopy camera setups that offer the "best bang for the buck"
Have a Zeiss GFL standard trinocular microscope w/neofluars.
Currently using a 14.0 MP OMAX digital camera, its 23 mm tube adapter, and ToupView software.
Considering simply substituting the 14MP Omax with an EOS-R series mirrorless Canon @30mp w/ appropriate TFL-Mount
but I don't want to waste time and money if I'm at the ceiling of usable resolution.
Reviews, literature, and specs are prohibitivly massive and hard to keep up with to say the least.
Can anyone shed some light (no pun intended) on their experiences with the newer digital cameras?
May explore returning to full frame 35 mm film format. Still have that Zeiss setup.
thanks.
Optimal DSLR cameras vs 35mm film resolution
Moderators: rjlittlefield, ChrisR, Chris S., Pau
Re: Optimal DSLR cameras vs 35mm film resolution
You may have better luck if you post this question in the 'Equipment Discussions' section,
as this area is actually for Images made through a microscope.
as this area is actually for Images made through a microscope.
Zeiss Standard WL & Wild M8
Olympus E-p2 (Micro Four Thirds Camera)
Olympus E-p2 (Micro Four Thirds Camera)
- rjlittlefield
- Site Admin
- Posts: 23608
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Optimal DSLR cameras vs 35mm film resolution
I have moved it, leaving behind a shadow that leads here.
--Rik
--Rik
-
- Posts: 1631
- Joined: Sat Apr 14, 2018 10:26 am
Re: Optimal DSLR cameras vs 35mm film resolution
One big factor not really addressed here are the eyepiece corrections. Zeiss 160mm objectives are not fully complete in and of themselves, they require a particular eyepiece which adds final corrections to the image. An ideal camera setup will use a Zeiss photo eyepiece and a relay lens, which will be easier to accomplish with a ~50mm lens on a full frame dslr than a small sensor microscope camera.
Re: Optimal DSLR cameras vs 35mm film resolution
My arguments about (against) getting a dedicated full-format R: (just a personal summary, there are extensive discussions and tests in other threads)
- You are already at the resolution ceiling, that is about 4MP. More megapixels on the sensor is oversampling, that is useful somehow but won't do miracles.
(Calculation: take the 10x 0.25 objective; has resolution wavelength/(2*NA)~1 micron ; field of view is 2 x 2 mm -> 2000x2000 points = 4MP. Some gain in doubling that for oversampling/Nyquist, take photos at 4000x4000 pixels = 16MP)
- The cameras with large sensors main advantage is that they collect more light, but just because they also mount larger lenses. On the microscope the objective is the same, light available is the same, so this advantage disappears. It actually gets worse, because the available light is spread out on the large sensor, local brightness decrease, so the large sensor must work at high ISO and long exposure, while on a small sensor the light is more concentrated. This issue may become serious in some low-light conditions like darkfield, with live-view almost impossible with the large sensor.
Of course large, good, expensive sensors retain some advantage in terms of dynamic range, color accuracy etc.
- Professional, high-grade microscope cameras from Zeiss etc use small sensors (max 1' format) with moderate MP, but focus on other features (FPS speed, sensitivity, color calibration etc)
- A DSLR/mirrorless works on battery (max 1 hr of live view) and isn't really optimized for live-view on computer. The (free) Canon computer live-view app is well made and usable, but a dedicated camera, USB or HDMI with ToupView, is more convenient I think.
So I'd go with Canon only if I have it already or if I plan to use it also with its lenses on human and landscape subjects.
My experience: I have two camera setups, one DSRL (old Canon 600D, APS-C, 18 MP) and the other with a crappy 5MP USB cam, both afocal over photo eyepiece. The 600D gives a littlebit more resolution and much better colors/contrasts. But for the practical reasons above, I use the crappy 5MP more often.
In the house there is an even better camera (wife's new Canon M6, 32 MP) but I never use it because of fear of damaging or getting dust on the sensor, and the improvement in microscope image quality over the old 600D was modest; while the improvement for humans or flowers is substantial.
Another (pro) point, given that decent USB cameras aren't cheap, DSLR and mirrorless may be a quite good sensor-for-the-dollar proposal. In particular I advocate ancient second hand thing like my 600D of course!
So said, consider well the post #4 by Scarodactyl above: eyepiece compensation and ease of implementing an optimal setup.
A quick way to check that you OMAX is ok regarding eyepiece compensation: take a photo with the phone camera at the eyepiece (with good care of having it steady and centered); compare it with the image of the OMAX, specially at the corners - if the phone is better, well, the Omax is missing the eyepiece compensation.
- You are already at the resolution ceiling, that is about 4MP. More megapixels on the sensor is oversampling, that is useful somehow but won't do miracles.
(Calculation: take the 10x 0.25 objective; has resolution wavelength/(2*NA)~1 micron ; field of view is 2 x 2 mm -> 2000x2000 points = 4MP. Some gain in doubling that for oversampling/Nyquist, take photos at 4000x4000 pixels = 16MP)
- The cameras with large sensors main advantage is that they collect more light, but just because they also mount larger lenses. On the microscope the objective is the same, light available is the same, so this advantage disappears. It actually gets worse, because the available light is spread out on the large sensor, local brightness decrease, so the large sensor must work at high ISO and long exposure, while on a small sensor the light is more concentrated. This issue may become serious in some low-light conditions like darkfield, with live-view almost impossible with the large sensor.
Of course large, good, expensive sensors retain some advantage in terms of dynamic range, color accuracy etc.
- Professional, high-grade microscope cameras from Zeiss etc use small sensors (max 1' format) with moderate MP, but focus on other features (FPS speed, sensitivity, color calibration etc)
- A DSLR/mirrorless works on battery (max 1 hr of live view) and isn't really optimized for live-view on computer. The (free) Canon computer live-view app is well made and usable, but a dedicated camera, USB or HDMI with ToupView, is more convenient I think.
So I'd go with Canon only if I have it already or if I plan to use it also with its lenses on human and landscape subjects.
My experience: I have two camera setups, one DSRL (old Canon 600D, APS-C, 18 MP) and the other with a crappy 5MP USB cam, both afocal over photo eyepiece. The 600D gives a littlebit more resolution and much better colors/contrasts. But for the practical reasons above, I use the crappy 5MP more often.
In the house there is an even better camera (wife's new Canon M6, 32 MP) but I never use it because of fear of damaging or getting dust on the sensor, and the improvement in microscope image quality over the old 600D was modest; while the improvement for humans or flowers is substantial.
Another (pro) point, given that decent USB cameras aren't cheap, DSLR and mirrorless may be a quite good sensor-for-the-dollar proposal. In particular I advocate ancient second hand thing like my 600D of course!
So said, consider well the post #4 by Scarodactyl above: eyepiece compensation and ease of implementing an optimal setup.
A quick way to check that you OMAX is ok regarding eyepiece compensation: take a photo with the phone camera at the eyepiece (with good care of having it steady and centered); compare it with the image of the OMAX, specially at the corners - if the phone is better, well, the Omax is missing the eyepiece compensation.
Last edited by patta on Sun Jan 29, 2023 2:26 am, edited 1 time in total.
- rjlittlefield
- Site Admin
- Posts: 23608
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Re: Optimal DSLR cameras vs 35mm film resolution
To clarify this point...
wavelength/(2*NA) gives the distance per cycle at the diffraction cutoff limit. The sampling density at Nyquist limit is 2 samples per cycle, hence 4000 pixels in 2 mm, 16 MP total. This is not oversampled; it is actually minimum sampling. But, it is minimum sampling at the cutoff frequency where MTF=0. For detail at twice that size, say 2 microns per cycle, diffraction-limited MTF will be around 40% and simultaneously that level of detail will be oversampled at 4 samples per cycle. The combination of Nyquist sampling at diffraction cutoff is the criterion used by Nikon's calculator for Matching Camera to Microscope Resolution.
--Rik
Re: Optimal DSLR cameras vs 35mm film resolution
Yes, fully agree with the above clarification. Sorry for the Oversampling blunder.
More than math&physics, my argument against high MP sensor comes from practical experience with old objectives & sub-optimal arranged setups:
After editing the image, cropping the corners because they're invariably a smear, I usually end up with an image worth about 1MP, max 4MP in few fortunate cases. Maybe is just getting content, but a good 1MP microscope image, when is good, is often enough for me.
And, truth said, I use the crappy 5MP camera over a smaller field, about 1.2x1.2mm in the example of 10x objective, so the smear in the corners is cropped already, while sampling is tighter.
The story changes with good modern objectives, but I have only the usual Nikon 20x 0.75 that systematically gives more (slide and sample allowing...)
More than math&physics, my argument against high MP sensor comes from practical experience with old objectives & sub-optimal arranged setups:
After editing the image, cropping the corners because they're invariably a smear, I usually end up with an image worth about 1MP, max 4MP in few fortunate cases. Maybe is just getting content, but a good 1MP microscope image, when is good, is often enough for me.
And, truth said, I use the crappy 5MP camera over a smaller field, about 1.2x1.2mm in the example of 10x objective, so the smear in the corners is cropped already, while sampling is tighter.
The story changes with good modern objectives, but I have only the usual Nikon 20x 0.75 that systematically gives more (slide and sample allowing...)
Re: Optimal DSLR cameras vs 35mm film resolution
Here are a couple of links you might want to look at.
The first gives you a general overview: https://krebsmicro.com/photomic1/photomic1.html
See more must read articles at https://krebsmicro.com/
The second is more specific to Zeiss 160 finite microscopes: https://microscopyofnature.com/micropho ... -eyepieces
I am using the Kpl10-8-10 projection at the moment, with an Olympus E-p2
You might also want to look into vibration, something one does not get with USB cameras.
Video quality is also something you might want to take into account. 4k video is quite something!
The first gives you a general overview: https://krebsmicro.com/photomic1/photomic1.html
See more must read articles at https://krebsmicro.com/
The second is more specific to Zeiss 160 finite microscopes: https://microscopyofnature.com/micropho ... -eyepieces
I am using the Kpl10-8-10 projection at the moment, with an Olympus E-p2
You might also want to look into vibration, something one does not get with USB cameras.
Video quality is also something you might want to take into account. 4k video is quite something!
Zeiss Standard WL & Wild M8
Olympus E-p2 (Micro Four Thirds Camera)
Olympus E-p2 (Micro Four Thirds Camera)
-
- Posts: 1631
- Joined: Sat Apr 14, 2018 10:26 am
Re: Optimal DSLR cameras vs 35mm film resolution
4mp is certainly not enough to get the most out of a zeiss standard.
If you are getting badly smeared corners it more likely indicates a problem with the photo setup, which is an easy problem to run into with zeiss. 18mm isn't a huge FoV, limiting relative to a modern objective, but the zeiss objectives will generally cover it pretty well when paired with the correct eyepiece. They can actually be stretched a bit further with that rare demagnifying optovar option.
If you are getting badly smeared corners it more likely indicates a problem with the photo setup, which is an easy problem to run into with zeiss. 18mm isn't a huge FoV, limiting relative to a modern objective, but the zeiss objectives will generally cover it pretty well when paired with the correct eyepiece. They can actually be stretched a bit further with that rare demagnifying optovar option.
Re: Optimal DSLR cameras vs 35mm film resolution
You are forgetting about the Bayer sensor. To get complete color sampling at each "point", you need at least 4*16Mp = 64Mp(Calculation: take the 10x 0.25 objective; has resolution wavelength/(2*NA)~1 micron ; field of view is 2 x 2 mm -> 2000x2000 points = 4MP. Some gain in doubling that for oversampling/Nyquist, take photos at 4000x4000 pixels = 16MP)
Re: Optimal DSLR cameras vs 35mm film resolution
Could you explain -and illustrate- your setup? (kind of adaptation, lenses, field coverage... sample images...)Lambda 6 wrote: ↑Thu Jan 26, 2023 10:20 pmHave a Zeiss GFL standard trinocular microscope w/neofluars.
Currently using a 14.0 MP OMAX digital camera, its 23 mm tube adapter, and ToupView software.
Considering simply substituting the 14MP Omax with an EOS-R series mirrorless Canon @30mp w/ appropriate TFL-Mount
...
Why TFL-Mount?
Which setup you do you have? There are many Zeiss and aftermarket variants.May explore returning to full frame 35 mm film format. Still have that Zeiss setup.
In principle if you have a good setup for 35mm SLRs a FF DSLR or mirrorless will work well.
In my experience a DSLR (or mirrorless) are able to produce higher image quality than many dedicated c mount microscope cameras (likely excepting modern high end ones) although those small sensor USB cameras are much more practical for routine work or classroom demonstration.
I agree with most Scarodactyl, Rik, 75RR and Lou pointers
Pau
Re: Optimal DSLR cameras vs 35mm film resolution
I would add a couple of points. The relay lens setup may influence a camera choice. For example, when I started photomicrography in 1984 I had a BH2 with 2.5NFK relay and another lens in the Oly floating leaf shutter assembly for 4x5 sheet film. Then I progressed to a 6x7 SLR and then to a 35mm SLR, using 3.3NFK and 2.5NFK respectively. The SLR's had vibration and difficult focus issues. The 4x5 could focus on a ground glass, but $ per shot. My first DSLR was a full frame canon 6D. Its 20MP seemed adequate for 2X and higher SPlanApo objectives. I wanted to make high quality 4K movies, so I started using a Lumix GH5 which makes fantastic video and has a pure electronic shutter for stills. Mating this to the BH2 scope required a 1.67NFK and a .7X speed booster in order get the whole field onto the sensor. The booster delivered an extra stop of light and all is sharp corner to corner. It's vibration free, and has the ability to do stacking and stitching and HDR, which leads to final images that are far superior to anything I ever shot on film.
HTH, John
hart3d.com
HTH, John
hart3d.com