🔬 Colour Mixing — Additive RGB & Subtractive CMY

RGB (Red, Green, Blue) is an additive model used for self-luminous displays: colours are created by adding light at three wavelengths. CMYK (Cyan, Magenta, Yellow, Key/Black) is a subtractive model used in printing: inks absorb light rather than emitting it. Converting between the two is non-trivial because the gamuts (ranges of reproducible colours) are different, with CMYK typically unable to reproduce highly saturated blues and greens that an RGB display can show.

The choice of RGB primaries is rooted in the trichromatic nature of human colour vision: our three cone types can be independently stimulated to create the perception of any colour within the colour triangle formed by the primaries. Red, green, and blue LEDs or phosphors are chosen because they span a wide gamut while being efficiently producible. The exact chromaticities are defined by standards such as sRGB (Rec. 709) and the wider Rec. 2020 for HDR displays.

The gamut of a colour device is the range of colours it can reproduce. On a CIE chromaticity diagram, the human visual gamut is the horseshoe-shaped region; any display or printer's gamut is a triangle or polygon within it. sRGB covers about 35% of visible colours; DCI-P3 (used in cinema) covers about 45%; Rec. 2020 covers about 76%. Colours outside a device's gamut must be "gamut-mapped" to the nearest reproducible colour.

Watercolour is a transparent medium: pigment layers transmit light, which reflects off the white paper and passes back through the layers — true subtractive mixing. Oil paint is semi-opaque: thickly applied colours can physically blend and also scatter light, mixing both subtractively and via Kubelka-Munk light-scattering theory. This is why mixing rule predictions are less accurate for opaque paints than for inks or filters.

Metamerism is the phenomenon where two surfaces with different spectral reflectance profiles appear identical in colour under one illuminant but different under another. It arises because human colour perception depends only on the integrated responses of three cone types, not on the full spectrum. Industrial colour matching must account for metamerism by specifying the illuminant (e.g., D65 daylight) under which two samples should match.

Rayleigh scattering — the scattering of light by particles much smaller than the wavelength — is proportional to 1/λ⁴, scattering blue light (450 nm) roughly 5.5 times more strongly than red light (700 nm). The sky appears blue because scattered sunlight reaching us comes predominantly from the blue end of the spectrum. At sunset, sunlight travels through more atmosphere, so blue is scattered away and only red and orange wavelengths survive to reach the eye directly.

The Munsell system organises colours along three perceptually uniform axes: Hue (the colour wheel position, 100 steps), Value (lightness, 0–10), and Chroma (saturation). Unlike the RGB cube, Munsell spacing is designed so that equal numerical differences correspond to equal perceived differences — a property called perceptual uniformity. Modern colour science uses the CIE L*a*b* space, which is derived from similar principles and is widely used for colour difference measurement (ΔE).