luminance is not linear in reality

[tabascoeye]

The luminance accepts values from 0-255, but the percieved "brightness" isn't linear, so we should compensate for that.

e.g. using a smaller amount of "brightness levels" which map to different values in the 0-255 range.
Then the "fading" alogrithm should increase its steps via this replacement matrix instead of only increasing/decreasing the dmx values by 1.

The luminance/brightness curve has to be defined but could also be ballparked by checking how far apart different percieved brightness levels are.

I would guess that in the lower value areas, the difference between each "brightness level" is smaller than in higher values.

e.g.:
"brightness" 1 may be value 5
"brightness" 2 may be value 10
but
"brightness" 9 may be value 200
"brightness" 10 may be value 255