picframe-colepd/notes.txt

19 lines
895 B
Text

For PCB:
XR1151 - 800'ish mA, 65uA quiescent -> 3.5 year
ESP32C3 - need to test if works with same code, otherwise NP
Article:
Color Floyd-Steinberg is a PITA. Difference colors? LAB colors, euclidian thingy. That sucks,
pink white and blue-ish white is the same difference as red and blue. Initial modification:
multiply A and B differences (=hue difference) with sin(L). This makes those differences less
important. Great results. But WtF is up with the shitty 'official' way? Seems there's multiple
ways: latest and greatest is E2000. Actually not that much different than my hacky solution.
Non-linear color space: one pixel of (0,0,0) and one of (64,64,64) don't look the same
(are not as bright) as two of (32,32,32). Luckily (linearized) RGB works.
Stucki dithering?
https://mathematica.stackexchange.com/questions/7483/how-to-calculate-mix-of-4-colors-defined-in-cielab-lab-model