Matrix (3×3 channel mix)

A matrix, in color, is a 3×3 grid of numbers that mixes the channels: each output channel (R, G, B) becomes a weighted sum of the three input channels. That’s what the RGB mixer does, what a gamut conversion is under the hood, and what “3D” color operations mean. The contrast is with a curve, which touches each channel alone (a “1D” operation) — a curve can’t let one channel influence another, but a matrix can. In a DCTL you write a matrix as nine multiply-and-add terms (or with a built-in helper). Hold the three basic operations — gain, curve and matrix — and you can read the guts of most color transforms.

First used in: 3.4 · DCTL