"We can therefore have the best of both worlds: a wider range of colors to enhance discriminability, while still being able to use color to know which point is from which category. The constraints are the same as with positional jitter: too little and the points all look the same, too much and we negate the encoding entirely." – Color Jitter: Using Randomness to Augment Categorical Visualizations
Digital painting software does not mix colours in a way that mimics real-life paint mixing. Specifically, colours like blue and yellow combine to form grey instead of green due to the reliance on the RGB colour model, which simulates the additive mixing of light rather than the subtractive mixing of pigments. The solution is to use the Kubelka–Munk (K–M) model, as it accurately predicts the behaviour of pigment mixtures. – Mixbox: Practical Pigment Mixing for Digital Painting