Super compact HSV/RGB conversions for Arduino/C

hsv2rgb.ino

int redPin = 6;
int greenPin = 5;
int bluePin = 9;

float col[3];
float hue = 0.0;

void setup() {
  pinMode(redPin, OUTPUT);
  pinMode(greenPin, OUTPUT);
  pinMode(bluePin, OUTPUT);
}

void loop() {
  setColor(hsv2rgb(hue, 1.0, 1.0, col));
  delay(50);
  hue += 0.01;
  if (hue >= 1.0) hue = 0.0;
}

void setColor(float *rgb) {
  analogWrite(redPin, (int)((1.0 - rgb[0]) * 255));
  analogWrite(greenPin, (int)((1.0 - rgb[1]) * 255));
  analogWrite(bluePin, (int)((1.0 - rgb[2]) * 255));  
}

// HSV->RGB conversion based on GLSL version
// expects hsv channels defined in 0.0 .. 1.0 interval
float fract(float x) { return x - int(x); }

float mix(float a, float b, float t) { return a + (b - a) * t; }

float step(float e, float x) { return x < e ? 0.0 : 1.0; }

float* hsv2rgb(float h, float s, float b, float* rgb) {
  rgb[0] = b * mix(1.0, constrain(abs(fract(h + 1.0) * 6.0 - 3.0) - 1.0, 0.0, 1.0), s);
  rgb[1] = b * mix(1.0, constrain(abs(fract(h + 0.6666666) * 6.0 - 3.0) - 1.0, 0.0, 1.0), s);
  rgb[2] = b * mix(1.0, constrain(abs(fract(h + 0.3333333) * 6.0 - 3.0) - 1.0, 0.0, 1.0), s);
  return rgb;
}

float* rgb2hsv(float r, float g, float b, float* hsv) {
  float s = step(b, g);
  float px = mix(b, g, s);
  float py = mix(g, b, s);
  float pz = mix(-1.0, 0.0, s);
  float pw = mix(0.6666666, -0.3333333, s);
  s = step(px, r);
  float qx = mix(px, r, s);
  float qz = mix(pw, pz, s);
  float qw = mix(r, px, s);
  float d = qx - min(qw, py);
  hsv[0] = abs(qz + (qw - py) / (6.0 * d + 1e-10));
  hsv[1] = d / (qx + 1e-10);
  hsv[2] = qx;
  return hsv;
}

Yes I saw it .
I think this is causing the problem with brightness and saturation also:
If you change this:

void setColor(float *rgb) {
  analogWrite(pinRED, (int)((1.0 - rgb[0]) * 255));
  analogWrite(pinGRN, (int)((1.0 - rgb[1]) * 255));
  analogWrite(pinBLU, (int)((1.0 - rgb[2]) * 255));  
}

to this (removing the inversion):

void setColorNEW(float *rgb) {
  analogWrite(pinRED, (int)(rgb[0] * 255));
  analogWrite(pinGRN, (int)(rgb[1] * 255));
  analogWrite(pinBLU, (int)(rgb[2] * 255));
}

You get the right colours, but you also get the opposite behaviour with Saturation and Brightness (which is the correct behaviour BTW!)

So, it seems to me that to correct your code, you have to remove the inversion in the setColor function.

Hello, thanks for your code! I get weird RGB-values on an ESP32:
https://pastebin.com/zs44RqPb

Calculations don't yield fractional results, only one and zero. What could be the problem?

I have the exact same behavior with the ESP8266. I used your code for quite a few projects with several types of Arduino, which always worked for me, but not with any ESP microprocessor.

Ensure that all variables are float types, maybe you are writing a value into an int...

what kinda sensor u use?

There's no sensor at all, it's to control an RGB LED...

You are absolutely beautiful :)! Is there anywhere you give an in-depth explanation of how all of your code works? (the important stuff like where is the 0.666666 number coming from, not how passing a array by reference works). I would very much appreciate that! Thank you again

Problems with inverted / non-inverted math in SetColor() function are related to the RGB LED diode you are using. The inverted version is for the common anode (CA) diode, while the non-inverted is for the common cathode (CA) RGB diode.