answered
2014-03-18 18:37:01 +0200
- List item
At the moment there are 5 brightness levels that can be selected with the brightness slider.
When the "adjust automatically" is not selected, it just chooses between 20%, 40%, 60%, 80% and 100% points.
With "adjust automatically" it chooses between Minimum, Economy, Normal, Bright and Maximum ambient light profiles.
Due to directional nature of the ambient light sensor it has trouble telling apart truly dark environment from "there are lamps, but none of them are directly over the sensor" situations. To compensate for this the minimum applicable brightness grows with each step from Minimum to Maximum profile. Which also means there is not too much dynamic range left at the Maximum setting.
At "office lightning" conditions there is just a small difference between the profiles, which easily makes it look like the setting does nothing. If it gets darker, the Minimum and Economy will reach darker brightness values than the rest. If it gets brighter, the Maximum and Bright will reach full brightness earlier than the rest.
The profiles are configured in /etc/mce/20als-defaults.ini, at the [BrightnessDisplay] section. If you wish to edit them,
make a copy with higher number in front, say "cp 20als-defaults.ini 90als-custom.ini". The files are processed in alnum
order, latter files can override settings from earlier. Using a copy means you can easily get back to defaults and upgrades will not overwrite the custom data you want to keep. Also note that mce needs to be restarted (systemctl restart mce.service) before the modifications are taken into use.
But before starting to tinkering with that you might want to choose: the lowest automatic setting you feel comfortable with in normal lightning conditions.
csd hardware test app. For me in this test app, when running the Single Test -> L-Sensor (light sensor) test, the test fails as the difference is only 15 when the Pass criteria is 50. This is while testing in a well lit room, so maybe the L-Sensor requires calibrating - if I shine a torch on the L-Sensor it passes with flying colours (a difference of 18761!)
Milhouse ( 2013-12-29 00:19:59 +0200 )editThank you mentioning that hardware test app.
I'm getting similiar results with that setup. Test fails in a well lit room with the difference of 18, and when using a torch, the difference is 18823. So yeah the L-sensor is definitely working...
Milo ( 2013-12-29 01:58:19 +0200 )editRaw ambient light sensor values seem to be here, in als_ch0 and als_ch1, populated only while the display is active (on):
The proximity sensor is event10, but this seems to be active whether the display is on or off. Not sure why it needs to be on when display is inactive/locked... future flip-to-silence feature, maybe? But then the proximity sensor would only need to be enabled during an alarm or incoming phone call... possibly another unnecessary battery drain?
Milhouse ( 2013-12-29 03:51:38 +0200 )editL-sensor definitely reacts to light sources.. Covered the raw data was 5 an 3 on ch0 and ch1. on well lit room they were 2678 and 1310 and when torched values peaked at 8393 and 1100. Still no visible difference in brightness when brigthness is set to max at different lightning conditions.
Milo ( 2013-12-29 13:05:37 +0200 )editThe proximity sensor might be enabled by default because it was necessary in N9. The clock on its lock screen was shown when proximity sensor could breath freely, but when covered, the clock disappeared.
Milo ( 2013-12-29 13:06:55 +0200 )edit