After a few months of remote work I got fed up with my computer's terrible webcam. Without the perfect lighting everything's a blurry mess, and it constantly failed to actually be detected by Zoom correctly. If you're similarly frustrated, and happen to have a DSLR or Mirrorless camera laying around that supports USB previews, and use Arch Linux, and have a bit of patience to poke around in the command-line, then wow, this is the perfect guide for you! :)

I'll quickly detail here how I got my Canon 200D working as a "webcam" on my Arch Linux setup. After the initial tinkering, it's been a pretty flawless experience, and besides a few short mentions, I couldn't find any comprehensive walkthrough of these steps online.

Make sure you've got a relatively recent kernel installed - I've tested this setup with major versions above 5.15.x and they all seem to work fine.

First up, install the linux-headers package using pacman or your favourite AUR helper, if you don't already have it. These are needed in order to build kernel modules; we'll need them for v4l2. Note that if you're using a custom kernel, such as Xanmod or Zen, you'll need to install the headers for it instead (e.g. linux-xanmod-headers).

After the headers are installed, install the v4l2loopback-dkms package from the AUR. Then, ensure it's loaded using modprobe - A good idea is to also add an entry to /etc/modules-load.d so it's always loaded automatically at boot.

Next, we can list all video devices using v4l2-ctl - You should see something like 'Dummy video device (0x0000)', and a /dev/ path:

Yay! That means the module's loaded properly. Next up, we need to install gphoto2 in order to snag the camera's preview output and pipe it to that /dev/video0/ path, using ffmpeg.

After installing the gphoto2 package from the AUR, connect your camera via USB, and check the output of gphoto2 --auto-detect - it should look something like this:

The final step is to put together a one-liner script to take gphoto2's preview and pipe that to the dummy video device v4l2 has created for us - make sure to check the exact /dev/ path from earlier, yours may not be video0. Something like this:

Since ffmpeg is handling the stream's encoding, we can take advantage of any of its CLI options, e.g., limiting the frame rate to lower CPU usage, or applying filters to the video.

So, what's a good way to permanently use this setup? I personally don't mind a bit of jank in my setup, so I currently just have a script living in /usr/bin that runs the initial command to start piping the video output to /dev/video0, and to attempt to restart on any failures. I'll leave a link to it in my dotfiles github repo . This script is then added to my bspwmrc's autostart section to run every login, although you could also run it manually in a new shell only when you want to use your webcam.

That's all there is to it, though! Feel free to reach out if you have any issues getting this working, I've spent a while before troubleshooting so may have an idea of where to start.