These are chat archives for Makuna/NeoPixelBus

19th
Dec 2016
MrTrick
@MrTrick
Dec 19 2016 14:40
Hi @Makuna , sorry that was a little vague. Here's an example graph of some max raw r,g,b (r:blue, g:orange, b:grey) values at each point in time; https://i.imgur.com/k75bs9x.png
MrTrick
@MrTrick
Dec 19 2016 14:47
so the transformation 256*(1-exp(-value)) , maps any range into 8-bit RGB.
Michael Miller
@Makuna
Dec 19 2016 19:17
Adding white (light) to a color doesn't give you more brightness to that color; it will sort of washes it out and provide more overall brightness. Think of the white LED as just turning on all the RGB leds with equal values. More RGB will just blend the blue away not really add to it.
MrTrick
@MrTrick
Dec 19 2016 19:20
yeah. I'm considering tweaking it so that if the raw R or G channels are high enough (just that early peak) it'll switch the (warm white colour) W channel on progressively more.
because for raw values above 2 the above transform is pretty close to 0xFF
Michael Miller
@Makuna
Dec 19 2016 19:22
To be true to the color, shouldn't the calculation be done with the highest value not just the current channel value, so in the above, you would use the blue channel value for calculating the red and green? Or is this what you are already doing?
As you want to scale them all the same, otherwise the colors just wash out above 1.0.
MrTrick
@MrTrick
Dec 19 2016 19:24
well it's a bit tricky to visualise - I'll see how it looks with different approaches. :-)
and I can "cheat" in that I know (or can simulate ahead of time) what the values will be.
Michael Miller
@Makuna
Dec 19 2016 19:25
I write little sample apps all the time that I can just plug values into it (using the serial read and serial monitor) to play with such things and give me a visualizer ;-)
MrTrick
@MrTrick
Dec 19 2016 19:25
and per the graph above I know for example that the blue channel (gray) and green channel (orange) in that early peak time is always lower than the other colours.