Colour Space

Colour spaces are generally 3-dimensional spaces that represent the Tristimulus spectral sensitvities.
The colours are specified by a set of three numbers (e.g. CIE coordinates X,Y and Z, or values such as Hue, Colourfulness and Lightness). There are also color spaces that are not 3-D (monochromatic, dichromatic)

Even though they are mostly 3-D, they are represented using 2D Chromaticity Diagrams for simplicity.

Types of Colour Spaces

Physical Palettes

• PANTONE

Commercial

• Munsell Colour System
• Natural Colour System (NCS)
• PANTONE

Mathematical

• CIE Colour Spaces
• RGB Primaries
• Cylindrical Colour Space Models
• YUV and YCbCr
• YIQ
• Munsell Colour System
• Helmholtz Colour Coordinates
  • Chromaticity is defined by dominant wavelength and purity

Not 3-D

Dichromatic

• RG Colour Model for early Technicolour films
• RGK for early colour printing
• rg chromaticity for Computer Vision

Special

• rg chromaticity for Computer Vision
• LMS Colour Space
• TSL Colour Space - for face and skin detection

Choosing Colour Spaces

There is no best colour space. Ideally you work in the same space as the destination device.

• For best performance: Use 8 bit sRGB D65, the default standard for web.
• For compositing: Use any linear RGB space with gamma (TRC) removed (gamma = 1.0), 16 bit float and white point set to D65, white mapped to 1.0 and black to 0.0
• For perceptual tasks (to feel natural): Use \(\text{L*a*b*}\), \(\text{L*u*v*}\), \(\text{LCh}_\text{uv}\), etc. with D65. Lab uses D65 normally in the textile, paint, and display device industries. They use floating point values and are also processor hogs, perhaps not as much as linear spaces.