Brightness and gamma

[crazymofo]

If you are struggling to get a matrix or megatree to look how it should (not washed out). Instead of turning down the brightness, turn brightness up to 50% and adjust gamma to 2.5.

My display was 30% on the controller, and tried turning down both matrix to 20% to try fix it. But this is where ive ended up and i am so happy with how it looks now (50%brightness, 2.5 gamma). Ive also pushed that setting across the whole of my display to keep it consistent.

 

[merryoncherry]

Unless it is a GS8208 pixel or something, gamma of 2.5 (or 2.2, or something in that range) will make an amazing difference to mixed colors.

 

[crazymofo]

I am super stoked with how it is now. And yes, these were ws2811 bullets. A lot of what i have read about the correlation between brightness and colour suddenly makes a lot more sense.

 

[Indigogyre]

I still struggle with this same issue. Even with trying some of the adjustments they just do not seem right.

 

[merryoncherry]

The biggest problem with WS2811 is that they just can't produce dim colors, certainly not very many. If you go from 0 - off, to 1 - the lowest setting, the light output is quite noticeable, and of course that only permits the 7 pure colors, no subtle hues or unsaturated shades. By the time you get many of those, the pixels have gotten noticeably bright. This argues for not turning the gamma up as much as you might otherwise want... a lot of dark colors will just come out black as they round closer to 0 than 1.

Running the pixels below 100% affects the issue noticeably. If you're running 30% brightness, you still have the same limited set of dim colors, but lose the top range to make things stand out. Or, if you are using an image or someone else's effect, it can squeeze even more of the dimmer colors down to black / completely off / down into the range where subtle shades are no longer possible.

I've seen, though, that the power consumption of WS2811 is not linear at the low end, and wonder if that is true of the light output also. If so, a more involved color curve may help a bit. Wondering if anyone has experience with this, or has tried to exploit it (if it is even true)?