Linux: using local attached USB camera bCNC 0.9.x is very slow/unusable. Partially resolved here ...

[AstxTyrol]

--- Camera.py   2025-08-18 22:11:44.196140193 +0200
+++ Camera.py   2025-08-18 22:07:18.531758889 +0200
@@ -146,6 +146,8 @@
         for prop_id, prop_value in self.props.items():
             self.camera.set(prop_id, prop_value)
 
+        self.camera.set(cv.CAP_PROP_FOURCC, cv.VideoWriter_fourcc(*'MJPG'))
+
         if self.camera is None:
             return
         s, self.image = self.camera.read()

Note: I like bCNC - a very fine tool! But I was never able to get the camera working. UI freezed. Till now:
Added above line and the camera can be used.
The fix is working on a x86_64 host and a raspberry pi 5.

It works for me ... but maybe I'm not able to add the correct parameters in the .bCNC config file. A detailed configuration help for Camera settings would be helpful.
Many thx for providing this CNC tool!

BR, Toni

[Harvie]

• Thanks for working on this!
• Can you please explain what this does?
• It seems that it switches camera to MJPG codec, what will happen when the camera hardware does not support MJPG?
• Also i prefer any code modifications being submitted as pull request (not a big deal for single line, but still preferable, also github will mention you as an bCNC author if you use pull request)

[AstxTyrol]

Think it switches to MJPG format on /dev/video0. Without this my bCNC gui freezes when using the usb camera.
Have not provided a pull request because it needs deeper knowledge of using streams from v4l2 devices which I haven't.
It may be a problem hard setting the codec to 'MJPG' if the camera would deliver only YUVY or other codecs.

The "Camery.py" module has a configuration parameter "codec" near "POSSIBLE_PROPERTIES":
Using

"codec": (Utils.getStr, cv.CAP_PROP_FOURCC),

it doesn't correctly construct "cv.CAP_PROP_FOURCC" using the cv.VideoWriter_fourcc() function (see link here)

If the parameter "codec" (or a new parameter) would do something like my one liner, more users may have success using the camera.

BR Toni

P.S.: my camera data:

root# v4l2-ctl --list-formats-ext -d /dev/video0
ioctl: VIDIOC_ENUM_FMT
        Index       : 0
        Type        : Video Capture
        Pixel Format: 'MJPG' (compressed)
        Name        : Motion-JPEG
                Size: Discrete 1920x1080
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 640x480
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 1024x768
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 1280x720
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 1280x1024
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 2048x1536
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 320x240
                        Interval: Discrete 0.033s (30.000 fps)

        Index       : 1
        Type        : Video Capture
        Pixel Format: 'YUYV'
        Name        : YUYV 4:2:2
                Size: Discrete 640x480
                        Interval: Discrete 0.040s (25.000 fps)
                Size: Discrete 1024x768
                        Interval: Discrete 0.100s (10.000 fps)
                Size: Discrete 1280x720
                        Interval: Discrete 0.100s (10.000 fps)
                Size: Discrete 1280x1024
                        Interval: Discrete 0.200s (5.000 fps)
                Size: Discrete 1920x1080
                        Interval: Discrete 0.200s (5.000 fps)
                Size: Discrete 320x240
                        Interval: Discrete 0.033s (30.000 fps)
                Size: Discrete 2048x1536
                        Interval: Discrete 0.200s (5.000 fps)

[Harvie]

can you please test what happens when self.camera.set(cv.CAP_PROP_FOURCC, cv.VideoWriter_fourcc(*'MJPG')) fails? (eg. when you put some nonsense like 'F4IL' instead of 'MJPG')? If it does not produce any error, please put some error message so we can easily see something went wrong when .set() returns error.

[statzern]

I have the same issue, activating the camera results in bcnc becoming unresponsive. It doesn't happen every time, but if I open and close it a few times or leave the camera active and do just about anything else in the program, it becomes unresponsive. The camera feed continues uninterrupted, but I can't click on anything else and have to force quit.

I tried dropping self.camera.set(cv.CAP_PROP_FOURCC, cv.VideoWriter_fourcc(*'MJPG')) into the camera.py just after the self.camera = cv.VideoCapture(self.idx) line in camera.py, which doesn't return any error but doesn't help. I know very little about python... So any guidance would help.

[statzern]

Just as an update, not sure if this makes any sense but I have been finding that when bCNC is not responding, switching to another window that obscures bCNC and then back seems to make it responsive again. Not every time, but most of the time. Very weird behavior. This is on an M1 Macbook Air running Mac OS 13.6.7.

On another note, the "orient" function does work incredibly well. While the hangups with the camera make the process take a little longer, the resulting accuracy is amazing! I tested by engraving 5 1mm diameter circles on a workpiece, then moving the entire workpiece to an arbitrary location, then orienting using the centers of 3 of the circles, and running the program again - only 1 path can be seen even under magnification! So the accuracy must be on the order of microns. Incredible!

[statzern]

Solved!!!

I am on an M1 Mac as stated, bcnc was running in Rosetta using the typical install. I fixed by installing in a virtual environment using Miniforge and thus forcing it to run in arm64 mode. To any newbs such as myself in the future:
If running
python3 -c "import platform; print(platform.machine())"
returns x86_64, you have the same problem.

Solution:
First install miniforge:
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-MacOSX-arm64.sh" bash Miniforge3-MacOSX-arm64.sh
Follow the Prompts:

Press Enter to view the license terms.

Type yes and press Enter to accept the license.

Press Enter to confirm the installation location (default is usually /Users/yourname/miniforge3).

Crucial Step: When asked "Do you wish the installer to initialize Miniforge3 by running conda init?", type yes and press Enter.

Restart terminal to apply changes.

Next, setup your environment:
conda create -n cnc python=3.10
Activate the environment
conda activate cnc
Install bCNC and opencv
pip install bCNC opencv-python
Again, confirm that you are running on the correct platform:
python3 -c "import platform; print(platform.machine())"
Run bCNC within the environment:
python -m bCNC

When you want to run bCNC again, just activate the environment and run bCNC as above:
conda activate cnc
python -m bCNC

Now CPU % has gone from 140% to 30% with the camera active and everything is smooth and responsive!!