Hello - just wanted to start by saying thank you to the team for Asahi Linux! When I got the most recent progress report from Open Collective, I had to give the firmware rebuild process a try and test out the ambient light sensor.
I've noticed what I think can be broken down into two small issues:
- The adjustments happen very sharply instead of being smoothed or ramped.
- The sensor is susceptible to changes in luminance/ambient light produced by content on the screen itself.
Together these can result in a slightly harsh and distracting user experience, primarily in dimly-lit or dark environments. As an example, if watching a movie that switches from darker to lighter content (or vice-versa), the ALS picks up the change in luminance and quickly adjusts the brightness - this affects the luminance of the display, which the ALS compensates for again, leading to a brief feedback loop of sudden brightness changes.
There may be plans to do so already, but smoothing out the brightness transitions should help. I'd imagine there may also be some method by which the sensor can be biased to at least partially ignore light from the screen, but I'm not sure how feasible it would be to implement such a thing, if it's already in the firmware or part of macOS, etc.
For reference, my test machine is a base M1 Air running Asahi Fedora Remix 44 on kernel 6.19.13.
Thank you!
Hello - just wanted to start by saying thank you to the team for Asahi Linux! When I got the most recent progress report from Open Collective, I had to give the firmware rebuild process a try and test out the ambient light sensor.
I've noticed what I think can be broken down into two small issues:
Together these can result in a slightly harsh and distracting user experience, primarily in dimly-lit or dark environments. As an example, if watching a movie that switches from darker to lighter content (or vice-versa), the ALS picks up the change in luminance and quickly adjusts the brightness - this affects the luminance of the display, which the ALS compensates for again, leading to a brief feedback loop of sudden brightness changes.
There may be plans to do so already, but smoothing out the brightness transitions should help. I'd imagine there may also be some method by which the sensor can be biased to at least partially ignore light from the screen, but I'm not sure how feasible it would be to implement such a thing, if it's already in the firmware or part of macOS, etc.
For reference, my test machine is a base M1 Air running Asahi Fedora Remix 44 on kernel 6.19.13.
Thank you!