MicrobeHunter.com Microscopy Forum

Hey folks,

I wanted to make a post about something that surprised me the other day. I was helping one of my Discord buddies figure out a camera to use with his new DIC BH2. One of the classic problems with the BH2 is that it's difficult and/or costly to get a small format DSLR attached to it without compromising image quality. 2.5x NFK photo eyepieces and photo tubes are plentiful and mesh perfectly with a 35mm sensor, but full frame cameras have long been the domain of professionals and rich enthusiasts. Well I did a little digging around and was surprised to see that, in fact, many full frame cameras are available used for less than $1k and some for less than $500. We got my buddy set up with a Canon capable of 1080/60 video which should be way better than any of the alternatives. Maybe this is just news to me, but I think it's important to note for folks that are looking for camera advice: full frames are within reach!

Here's a link to Adorama showing the cheapest cameras with at least a "Good" rating: https://www.adorama.com/l/Used/Photogra ... D&sf=price

It's worth noting that full-frame consumer-grade digital cameras are likely overkill for most microscopy. I use a 16Mp APSC format camera and I'm always surprised by how crappy the photos are "out-of-the-box". There are very good physics-related reasons for that and they are well explained in one of Oliver's videos on the topic; I think it's called "How many pixels do you need?"

I suspect that the machine vision gang uses a camera technology that is very different than that used by general photographers. And despite maxing out at 8-10Mp, takes much sharper images. I suppose if you're scanning tissue samples for cancer cells or bacteria, you're not using regular photographer grade, digital camera technology, no matter how sophisticated it is.

To add to the discussion, it may seem intuitively wrong, but high-magnification images require far fewer pixels to capture the image than low magnifications. The lower-power macro photographers, say 1x to 20x, need the latest full-frame cameras to capture everything their lenses can see. As Josmann says, some of those cameras have become relatively affordable.

A hobbyist capturing cell images at 200-1000x can capture excellent images with fewer pixels. A cell phone sensor, a one-inch Sony sensor, or an APS-C sensor is often entirely satisfactory.

PeteM wrote: ↑

Tue Jun 06, 2023 5:46 pm

To add to the discussion, it may seem intuitively wrong, but high-magnification images require far fewer pixels to capture the image than low magnifications.

The (photography) lens testing website DxO Mark has exhaustive technical discussions about this very thing. In addition to evaluating lenses, they evaluate lens/sensor pairings. In many cases, high performance lenses must be married to specific sensors (and pixel densities). Higher pixel counts alone do not automatically yield better images.

In fact, producing the best point spread function is also related to pixels size: this is why an image taken with a 16Mp APSC sensor is far superior to 1/3" format 16Mp mobile phone sensor.

We have a lot better image since having a FF camera (12MP), 35mm film was what the manufacturer had in mind when building the microscope.
FF isn't about more pixels but about sensor size.

Well, sure, APS-C cameras can be very good, but that's missing the point of the post : if you use those on a BH2, you're going to deal with substantial FOV crop. You want full frame because that is the format these scopes were designed to project on to.

Regarding pixel density and the size of Airy disks, well, I have a lot to say on that subject, but I will leave it for another time

josmann wrote: ↑

Wed Jun 07, 2023 4:51 am

Well, sure, APS-C cameras can be very good, but that's missing the point of the post : if you use those on a BH2, you're going to deal with substantial FOV crop. You want full frame because that is the format these scopes were designed to project on to....

How was Olympus's support of contemporary one-inch, c-mount video cameras? That was the image out of the photo ports of the Wild microscopes I am familiar with. To project on to a 35mm or Polaroid camera, Wild had you add more optics.

Remember too, that unless you're talking about a monochrome sensor, the number of effective pixels is less it would seem due to the Bayer color filter array. And "acquiring information" is not the same as acquiring a nice looking image.

Tom Jones wrote: ↑

Wed Jun 07, 2023 7:54 pm

"acquiring information" is not the same as acquiring a nice looking image.

^^^^^^^^^^^^^^^^^^^^^^^
Mr. Jones gets it!

jfiresto wrote: ↑

Wed Jun 07, 2023 7:03 pm

How was Olympus's support of contemporary one-inch, c-mount video cameras?

I haven't played around with them much personally, but there was a .3x reduction lens attachment called the MTV-3 that bolted on top of a photo eyepiece and had a C-mount thread. Not sure which photo eyepiece was used with them for which sensor format. I do have one and a few photo eyepieces so I can try it out some time.