Zeiss AxioCam MRc5: no image / invalid parameters (110) in Color mode (MM 2.0), but works in Monochrome

[BoyuanGIT]

Description

I am using Micro-Manager 2.0 (nightly 2025-03-27) with a Zeiss AxioCam MRc5 (Zeiss AxioCam adapter) and a Zeiss Imager.M2M microscope (Zeiss CAN29 adapter, RS232 connection to microscope body, not the MCU controller).

Problem:
When the camera property Color mode is set to Color, I cannot acquire any image:

• Clicking Snap gives an error: Error in device "Zeiss AxioCam": Invalid parameters (110)
• Clicking Live shows no image (black window, or the Live button does nothing)
• Changing Binning after this triggers an additional message: Missing pixel type

If I change Color mode to Monochrome, everything works correctly (black-and-white images, exposure and white balance adjustable).

The same hardware configuration works fine in Micro-Manager 1.4.21 with Color mode = Color (color images are displayed without any error).

Steps to reproduce:

1. Load hardware configuration (Zeiss CAN29 + Zeiss AxioCam) in MM 2.0.
2. In Device Property Browser, set camera Color mode = Color (other settings default: Scan Mode = Interpolated, Exposure = 10, etc.)
3. Click Snap.
4. Observe the error popup.

Error logs (relevant excerpts):

[ERR,Core] Failed to snap image from current camera
[ERR,Core] Error occurred in device CMMCore::snapImage: Error in device "Zeiss AxioCam": Invalid parameters (110)
java.lang.Exception: Error in device "Zeiss AxioCam": Invalid parameters (110)
    at mmcorej.MMCoreJJNI.CMMCore_snapImage(Native Method)
    at mmcorej.CMMCore.snapImage(CMMCore.java:951)
    ...

Additionally, the log repeatedly shows a parse exception (may be unrelated, but occurs during GUI updates):

java.text.ParseException: Unparseable number: " 6"
    at org.micromanager.internal.utils.NumberUtils.coreStringToInt(NumberUtils.java:114)
    ...

Comparison table (properties):

Property	MM1.4 (works)	MM2.0 Color (fails)	MM2.0 Monochrome (works)
Color mode	Color	Color	Monochrome
Scan Mode	Interpolated	Interpolated	Progressive
Binning	1	1	2
Color processing	Off	Off	Off

All other white balance and exposure settings are identical.

Additional findings (tested with DLL replacement)

To further isolate the issue, I attempted to replace the MM 2.0 device adapter DLL with the one from MM 1.4 (which works correctly with color mode on my system). After copying mmgr_dal_AxioCam.dll from MM 1.4 into the MM 2.0 installation folder, Micro-Manager 2.0 fails to load the hardware configuration with the following error:

Error: Failed to load hardware configuration
Line 17: Device, ZeissAxioCam, AxioCam, ZeissAxioCam
Failed to load device adapter "AxioCam" from "D:/ProgramFiles\Micro-Manager-2.0\mmgr_dal_AxioCam.dll"
[Incompatible device interface version (MMCore requires 75; device adapter has 70)]

After this replacement, even when I try to create a new hardware configuration from scratch, the Zeiss AxioCam device is no longer selectable in the device list (presumably because the adapter fails the version check and is rejected by MMCore).

Testing environment:
I am debugging a custom plugin in IntelliJ IDEA. All logs shown in this issue were captured from the IDEA console during plugin execution (green triangle run). No other modifications were made to the system.

Thank you very much for taking the time to look into this issue. I am happy to provide any additional information, run further tests, or try experimental builds if that helps to identify the root cause of the color‑mode failure in MM 2.0.

Please let me know if there are specific diagnostic steps, log levels, or code patches you would like me to apply. I will respond promptly.

Best regards,
Ryan

[marktsuchida]

Thanks for the detailed report.

Can you try the build 6089f32 at https://download.micro-manager.org/ci/2.0/Windows/ (or the 2026-04-25 nightly if it's out)? It (hopefully) fixes some of the related issues and adds logging that might help identify the root cause. Logs from this version while snapping (with debug logging enabled, Tools > Options) would be helpful for finding out the root cause of the main problem.

[BoyuanGIT]

@marktsuchida Hi Mark, thanks for the suggestion. Here is my test result.

Test environment:

• Micro-Manager 2.0 (the new nightly MMSetup_64bit_6089f32_20260425.exe)
• OS: Windows

Steps and observations:

1. Started MM, color mode was initially Color
2. Clicked "Live" → nothing happened → stopped live
3. Clicked "Snap" → Error: "Failed to snap image: Missing pixel type"
4. Changed color mode to "Monochrome"
5. Clicked "Live" → preview window showed, everything fine (grayscale)
6. Stopped live, clicked "Snap" → success
7. Changed color mode to "Color"
8. Clicked "Live" → nothing happened → stopped live
9. Changed Binning to 2 (color mode still Color), clicked "Live" → failed, stopped live
10. Changed Scan Mode from Progressive to Interlaced, clicked "Live" →
    Error: "Couldn't start live mode sequence acquisition. Error in device 'Zeiss AxioCam': Camera busy error (111)"
11. Changed Scan Mode back to Progressive, clicked "Live" →
    same error: Camera busy error (111)
12. Changed color mode back to Monochrome →
    still same error: Camera busy error (111)

Once the camera enters “Camera busy error (111)”, it seems to latch and even switching back to Monochrome cannot recover (requires restart of MM).

I also saved a debug log (591 KB) during the test. Please find it attached.

Thank you for your time.
Best regards,
Ryan

CoreLog20260426T093019_pid984.txt

[marktsuchida]

Thanks, I think the original error 110 got fixed.

The Missing pixel type error is not from the camera but from the Micro-Manager UI. The camera is producing 16-bit-per-component RGB images under these settings, and MMStudio (the Java layer, 2.0) does not currently support displaying or saving those. If you are writing your own code, you might be able to access such images (you should get a short[] containing components in BGRa order, where the a is unused).

Unfortunately the AxioCam adapter doesn't currently support producing 8-bit RGB images. So for it to work in the user interface, we'd need to either add that, or wait for MMStudio to support RGB64 images (which would be nice — I sometimes want it, too — but I can't promise a timeline because it will need to work significantly differently from the RGB32 images that are natively handled by the ImageJ canvas we use).

The error 111 (camera busy) likely comes from failure to correctly clean up the previous acquisition. I've attempted a fix, which will (soon) be in CI build c4ba78b (or nightly build 20260426). It tries to ensure proper stopping of sequence acquisition and also resets to Interpolated and binning = 1 when switching to color mode (for MRc5 specifically). Let me know if this fixes that error.

[BoyuanGIT]

Thanks Mark – appreciate the fix and the explanation. I'll go with retrieving the raw short[] data myself and handle it outside the UI.