Lumix GX80 for Leica S9D

[lukego]

Hello!

I would like to upgrade the camera system on my Leica S9D microscope with a 4K video capability. I'd like high quality for reasonable price, particularly avoiding unnecessary purchase of expensive Leica optics (approx $500 for anything that is made out of glass.)

I have read the forum and it sounds like a Lumix GX80 (aka GX85) for ~$500 is the answer. I would need some kind of a thread adapter to mate this with the Leica microscope. I don't know if I would need to buy that from Leica, buy a third-party one, or whether I could 3D print it on my Prusa printer. I believe the Leica part number would be 10 450 317 ("1.0× C-mount adapter for HC") but I will need to check with a rep. I imagine it will cost about $200 even if it is just a small piece of plastic or steel "because Leica."

Questions for anyone who enjoys discussing such things

Is the Lumix GX80 indeed a good choice? I'd primarily be recording on a separate device via the HDMI output.

Any hot tip on mating the Lumix with the microscope without undue fuss?

How is the field of view likely to be with the 4/3" sensor on a standard microscope photo port? (Should I expect to see a rounded image with black fringes like in the eyepieces? Or a somewhat cropped view compared with the eyepiece?)

How is the low-light performance of the Lumix? I have had trouble with other microscope cameras needing very strong illumination, particularly IMX290 sensor which I found to be borderine useless. I hope the Lumix will have better sensor performance and better ISO...?

I am upgrading from a Raspberry Pi with HQ camera on 1/2 reducer lens. Currently I'm only recording in less than 1080p but that's partly due to a poor choice of reducer lens forcing me to use the full sensor in 2x2 binned mode. The RPi camera does make much better use of available light than the (cheap) dedicated C-mount cameras I have tested in the past though.

Thank you for your attention

[Scarodactyl]

For the adapter, see my thread over here viewtopic.php?p=118391#p118391
I used a 200mm doublet for my aps-c camera but you could definitely go with a shorter focal length for a m4/3 sensor. It's set to take a 1" thorlabs doublet but that could be modified of course.

That said it would help to know what you already have on there.

[lukego]

I attach a picture of my current setup:

scope.jpg (88.65 KiB) Viewed 6387 times

The stack from bottom to top is:

• Leica S9D
• Leica C-mount adapter with 0.5x lens (10450528 $389)
• C-mount to CS-mount adapter
• Raspberry Pi HQ camera
• Raspberry Pi 4

I like the idea of putting a Lumix camera on if this can be done economically thanks to the 4/3 sensor. If this is not especially the case then I might be tempted to use a full-frame Canon or Nikon camera instead but I have been quoted around $600 for the microscope optics (2.5x camera adapter.)

[Scarodactyl]

You would also need the 1x adapter to use the 2.5x, unless that has been updated. The 1x should cover m4/3, but you'd also need a 37mm OD adapter to hold the camera almost flat against the top of it.
I think making your own is the best/most economical approach but I do have a used 1x adapter if you end up wanting to go that route.

[lukego]

You would also need the 1x adapter to use the 2.5x, unless that has been updated. The 1x should cover m4/3, but you'd also need a 37mm OD adapter to hold the camera almost flat against the top of it.
I think making your own is the best/most economical approach but I do have a used 1x adapter if you end up wanting to go that route.

That's interesting!

Does the 1x adapter include glass/optics? Or is it just a mechanical piece? (Are they cheaper than relay lenses?)

I don't have a feeling for the Leica second-hand market. How much would you imagine selling your relay for? Do you know if it would work on a Leica S9D? (How much should I imagine selling my 0.5x relay lens for after this upgrade?)

[Scarodactyl]

With no adapter at all, just bare threads, the image coming out of the top of the microscope is focused at infinity. The adapters have various focal lengths to focus the image onto various size sensors, with longer focal lengths producing a larger image. The 1x lens has something like a 150mm lens in it--I am not 100% sure since my dslr can't sit low enough to focus on it (but a m4/3 camera probably could because of the short distance from shoulder to sensor).
Actually, this guy has his priced lower than I'd want to go, so that would be my first pick if I were you https://www.ebay.com/itm/LEICA-10445930 ... 4329381743
Well, that or finding a ~150mm fl achromatic doublet on surplus shed and using that in a printed adapter, but 75 bucks isn't too bad a premium to pay for one made with proper metal. You would still need to 3d print an adapter to hold your camera though--the 1x has a 37mm iner diameter with a clamp.
I have not worked with m4/3 either so I can't guarantee 100% the logistics will work out with mounting it on the 1x.
I imagine the 0.5x adapter could fetch 200ish depending on how long you want to take to sell it.

[Microscopy_is_fun]

Hi Lukego,
here is my setup on the S9D with a 1x the adapter for mft-cameras (I use a Panasonic G9) or 1"cameras. The cameras be attached to the c-mount of the adapter. For mft you will then need an additional c-mount to mft adapter, which is readily available at amazon/ebay.

[Scarodactyl]

Yeah, that model 1x adaptee might be more convenient if you can get it, though if you have a 3d printer it shouldn't be too hard to do.

[lukego]

Currently I am mostly using the 1" microscope camera. But the test shots with the G9 worked without problem, the field of view of the S9D is large enough to prevent vigneting when using mft. At the moment I am travelling, but next week I could make some test pictures, if that's helpful for you.

That's a very compact camera! If I may ask, what model is that and how do you like it? I suppose a full size camera body would take up more space (and obstruct access to my tool shelf behind the microscope) so it could be an alternative to consider.

Also, how is the field of view on the 1" compared with the 4/3"? I would like to avoid excessive cropping on my viewport. Better a little vignetting in a video than to entirely miss the area I that I am manipulating.

Thanks for the tips!

[lukego]

I have now ordered a Lumix GX80 camera with a C mount adapter. I plan to try it out with the 0.5x lens initially to see if I like the camera and otherwise to return it.

I have contacted my Leica rep about the recommended 1x adapter.

[Microscopy_is_fun]

That's a very compact camera! If I may ask, what model is that and how do you like it? I suppose a full size camera body would take up more space (and obstruct access to my tool shelf behind the microscope) so it could be an alternative to consider.

Also, how is the field of view on the 1" compared with the 4/3"? I would like to avoid excessive cropping on my viewport.

[jfiresto]

Also, how is the field of view on the 1" compared with the 4/3"? I would like to avoid excessive cropping on my viewport. Better a little vignetting in a video than to entirely miss the area I that I am manipulating....

The GX80/GX85 has a 2.2 crop factor (w.r.t. 35mm full frame) when shooting video. I bet it will not suffer vignetting from a c-mount.

So, what do you think of the S9D after using it for a year?

[Scarodactyl]

The diameter of a m4/3 sensor is 22.5mm, so it is theoretically possible for the c mount to cause a little vignetting. It may depend on the adapter.

[jfiresto]

The diameter of a m4/3 sensor is 22.5mm, so it is theoretically possible for the c mount to cause a little vignetting. It may depend on the adapter.

I think we will get lucky (I bought a GX85 for video as well). The crop factor reduces the camera's video diagonal to 22.5 / 2.2 * 2 = 20.5 mm.

[Scarodactyl]

Yeah, that should be safe. I am pretty sure I've seen complaints of vignetting on this type of setup without the crop but it's been a while. Mostly notable since it's really not a necessary optical evil so much as a mechanical issue, but it's definitely a lot easier logistically.

[lukego]

So, what do you think of the S9D after using it for a year?

That is the five thousand dollar question!

On the one hand, I love it. It's beautiful and ergonomic. The 9:1 zoom range means that I can fluidly move between placing parts onto circuit boards and then zoom in to soldering them in place, and then zoom in more to inspect solder joints. Then back out to work on the next part. I use the whole zoom range on a minute-to-minute basis.

On the other hand, it's expensive and the high-quality optics are surplus to my requirements. If there were a cheaper option with 9:1 zoom range, for example an upgrade to my old AliExpress AmScope clone or an affordable Meiji model, then I would probably have been content with that instead and saved some money. But I didn't find any such alternative.

I am really, really glad that I picked the S9D instead of the Leica A60, Zeiss Stemi 305, or Meiji microscopes that I was considering. I need both the extended zoom range and the trinocular camera port. If I'd bought a cheaper microscope that didn't have both features then I'd be regretting it by now and thinking about an upgrade.

I'm sure there's an alternative universe where I spent time learning enough about microscopy to acquire a bargain scope on eBay, but hey, I spent all of that time on other topics like studying mathematics and Bayesian data analysis instead.

In summary I love the S9D and regret nothing I hope to continue enjoying it for decades to come.

EDIT: Let me add that the camera is really what is letting me down lately. Making videos and sharing clips is one of the most fun parts of using my microscope. I've fallen out of the habit of doing this because the workflow I setup is too complicated. I have made myself a Linux system administrator of my microscope camera and that feels like work: ssh, raspivid, scp, ffmpeg, etc. I'm looking forward to streamlining this and doing more of the fun stuff!

[lukego]

I think we will get lucky (I bought a GX85 for video as well). The crop factor reduces the camera's video diagonal to 22.5 / 2.2 * 2 = 20.5 mm.

I am curious to try the GX80/GX85 with my current 0.5x reducer lens. That should have a lot of vignetting. Maybe in 1080p mode the camera could crop a region of interest though?

I don't expect that to be a serious alternative to the correct 1x lens but it will be an interesting diversion while I'm evaluating the camera and working out how to acquire the lens.

(Is there a better way to buy Leica parts in Europe than emailing a local sales rep? I bought the scope from Spectra Services in the US - who were great - but I don't want to pay international postage on individual parts.)

[Scarodactyl]

Buying dorect from Leica is almost never the right answer, even with international shipping. Their prices are aimed at institutions, not individuals.

[jfiresto]

EDIT: Let me add that the camera is really what is letting me down ... because the workflow I setup is too complicated. I have made myself a Linux system administrator of my microscope camera and that feels like work: ssh, raspivid, scp, ffmpeg, etc. I'm looking forward to streamlining this and doing more of the fun stuff!

Other than that, I get the impression you are not completely unhappy with the microscope. After all that product research, I am glad things worked out so well.

I am setting up my GX85 to appear as a network drive: to allow me (or a program) to quickly get into the camera and pull images and videos. You made me realize that I need a device to do that with two cameras. I think a Raspberry Pi and some Python would answer nicely.

[dtsh]

EDIT: Let me add that the camera is really what is letting me down lately. Making videos and sharing clips is one of the most fun parts of using my microscope. I've fallen out of the habit of doing this because the workflow I setup is too complicated. I have made myself a Linux system administrator of my microscope camera and that feels like work: ssh, raspivid, scp, ffmpeg, etc. I'm looking forward to streamlining this and doing more of the fun stuff!

As part of my usual automation I tend to make scripts to do a lot of the repetitive work. I suspect this is easier for me than most as these are long standing habits for me as I've been using unix and unix-like systems since the early 90's, but if you can describe your process I might be able to help. At one point I had my script setup to use a pushbutton connected to the GPIO pins to trigger the camera when pressed, but the current iteration just uses the [Enter] key. Currently I have a python script that uses the live view feature of the picamera python module to display to a monitor, but I would like to replace that with an HTTP interface so that I can dispense with the monitor; unfortunately I've been saying that for over a year now. In another iteration I was playing with I had a stepper motor controlling the focus to capture images for focus stacking, it worked well, but I'm stalled waiting on parts and motivation for a different, hopefully cleaner, approach. The RPi is a fantastic platform, but there's a whole lot there to learn to take advantage of it all. In 30 years of constant use, there are still a lot of areas where my knowledge is lacking.

[lukego]

As part of my usual automation I tend to make scripts to do a lot of the repetitive work.

I'm experimenting with going in the opposite direction: Today I acquired a Black Magic Mini Extreme Iso video mixing station. This has 8xHDMI inputs, which can all be recorded simultaneously in 1080p, and 2xHDMI outputs. My plan is to connect my stereo microscope, my autofocus inspection microscope, my GoPro, my DSLR to the inputs and my two large monitors to the outputs.

This way I can press a button to start a synchronized recording from all cameras. I can also press buttons to map cameras onto displays in various configurations (full screen, split-screen, picture-in-picture, etc.) It is even able to stream the mixed video directly to the internet (or save it in Davinci Resolve project file format.)

I think that it will make working in my lab much more social. I'll be able to show people what is going on in real time. I especially want to do this with my young kids who often don't know what is going on under the microscope.

I have wanted to do this for a long time but I thought it would be a major undertaking with PCIe capture cards and Linux-compatibility software problems and so on. I've already been through one round of that using USB cameras and fighting to get enough USB2 bandwidth into a server. I was pleasantly surprised to hear about the Black Magic appliance. I've been using it for a couple of hours now and so far I love it

I also plan to incorporate my Raspberry Pi cameras into this setup. My idea is to set them up as dedicated HDMI cameras instead of using the ethernet/wifi interface. I have printed a case for one and I'm waiting for mechanical parts to arrive to screw it together, based on this design https://medium.com/javarevisited/using- ... af84aafee2.

[lukego]

Other than that, I get the impression you are not completely unhappy with the microscope. After all that product research, I am glad things worked out so well.

Indeed! And thank you for all your patient help in choosing the right microscope at the time. It is my prize possession.

I am also finding new uses for the microscope, most recently while learning to sharpen knives. I wonder how I ever lived without one. Maybe people here would also appreciate the microscope-aided sharpening tutorial that I am following on Youtube now: https://www.youtube.com/watch?v=ZTuZB8uGbaU.

[Microscopy_is_fun]

I'm experimenting with going in the opposite direction: Today I acquired a Black Magic Mini Extreme Iso video mixing station.

This looks like a very interesting gadget for various purposes in video-microscopy!

[lukego]

Just received the Lumix GX80. Installed without any problems at all. Here is a quick first 4K video!

https://www.youtube.com/watch?v=0CrN3Q8QF-U

[Scarodactyl]

That's an impressive diagonal for a c mount adapter. Glad it's working!

[lukego]

That's an impressive diagonal for a c mount adapter. Glad it's working!

Thanks! What does "diagonal" mean in this context?

I am actually pleasantly surprised at the field of view with this 0.5x camera lens. On the one hand I don't use all the pixels due to vignetting. On the other hand I capture almost the entire FOV.

This seems like a good trade-off for me because I use the camera port for observing and documenting my work. I use the whole FOV through the eyepieces when I am working so it has been a frequent problem that the camera crops out the most important areas. I'd have a lot of potential for cropping/scaling/curating the 4K video down to 1080p in post-processing if I want to aswell.

And it doesn't hurt that it saves me hundreds of dollars on a new camera adapter, and makes me much feel more satisfied with my 0.5x adapter than I was before!

[Scarodactyl]

In this context diameter would probably be more appropriate. Most of these c mount adapters assume a 16mm diagonal at a 1x magnification factor, so even if the image field is bigger they may crop it. For a 0.5x anything you get beyond the 8mm diameter it takes to cover the diagonal of a 1/2" sensor is a bonus.

[lukego]

In this context diameter would probably be more appropriate. Most of these c mount adapters assume a 16mm diagonal at a 1x magnification factor, so even if the image field is bigger they may crop it. For a 0.5x anything you get beyond the 8mm diameter it takes to cover the diagonal of a 1/2" sensor is a bonus.

Thanks for this explanation. I feel very fortunate to get so much image quality on the out-of-spec parts of the FOV. I am feeling better and better about spending money on the Leica gear.

It's not as sharp as the FOV seen through the eyepiece, as captured by this other person's video (https://www.youtube.com/watch?v=ugbNumCa6jI), but for now I think that it's great.

Just as a point of comparison, here is an old video from my AliExpress AmScope-clone with its 0.5x camera adapter and a generic AliExpress microscope camera: https://www.youtube.com/watch?v=umkNrDrfWNA. I believe that on this one I did manual exposure correction in post-processing because the original image was too dark (and camera ISO support very low quality.)

[Microscopy_is_fun]

When I am back home, I will make a quick comparison regarding differences in field of view, that's actually of interest for me, too,

Here is the comparison of a 1"-microscope camera with c-mount attached to the Leica S9D without intermediate optics. Magnification was 10x (1x zoom with 10x eyepiece). The edges start to get slightly blurry, but are still ok in my opinion. I cannot see any vignetting.

[lukego]

Here is the comparison of a 1"-microscope camera with c-mount attached to the Leica S9D without intermediate optics.

That is very interesting, thank you for sharing! The "1x" FOV looks like a much better choice for crisp high-resolution photography/videography of specimens.

I think the 0.5x will suit me better for documenting electronics work done via the eyepieces. Capturing a larger FOV will reduce the risk of "missing the action" and I'll often want 1080p which means I can crop the FOV without loss too (e.g. via "2x digital zoom" feature of the camera.)

So many options! I am surprised at how much the right choice depends on specific applications.

EDIT: Oh, you said "without intermediate optics," meaning this is with no glass at all instead of the "1x" camera lens? I need to reread your posts when I am in less of a morning-rush