So most of us want to be able to implement dimming curves in one way or another to take care of the pesky problem that almost no light turns on at 1% and gets brighter up until 100%.
The typical way to do this is to map the 255 values to different values. So if the lights don't turn on until 10% you have just lost 25 values of resolution. Then if you max your lights out at 80% or 90%, you've just lost another 10% to 20%, Now you're down to about 70%.
My idea was for hardware dimming curves (programmed into controllers) the output resolution would be upped from 8 to 9, 10, (some other number) of bits, and that way you don't lose the resolution.
What do you all think?
Does losing the 30% really matter?
Do you think it's worth it?
The hardware and communications protocols would also need to support nine or ten bits as well, otherwise there isn't much point in this.
Nm, I missed the critical part of your comment.
My main discussion is it really worth the effort for that extra resolution. Was wondering what others had to say.. You'd obviously be wasting quite a bit more data space and processor overhead in doing it.
Anyway, it's doable with Renard, although using an external clock may be required. Not that I'm promising anything, as there is still the question of what the curves should be, how they get into the controller, and how they are selected.
