Junya Yang | Academic Homepage

Appearance

PostData Analysis / Data Visualization

DV-05-Color Vision

2025-12-29
Back to Blog

DV-05-Color Vision ​

Visual spectrum ​

Visible light is the small part within the electromagnetic spectrum

波长范围大约：400nm~700nm

  • 紫光：~400 nm（波长短）
  • 红光：~700 nm（波长长） 经典顺序：ROYGBIV
  • Phong illumination model
    • Ambient Light
    • Diffuse Light
    • Specular Light

  • Trichromatic theory

    • Three types of cones, each sensitive to different band of wavelengths
    • Long wavelength: red
    • Medium wavelength: green
    • Short wavelength: blue

  • Rods are extremely sensitive, and can be triggered by a single photon.

  • Cones require significantly brighter light to produce a signal.

  • Retinal ganglion cells

Color space ​

  • A color space is the set of colors which can be displayed or reproduced in a medium.
  • Dependent on devices and displays
  • Colors of the same RGB values may look differently on different color spaces.

Color space defines the range of colors

  • sRGB
  • Adobe RGB
  • SWOP CMYK

sRGB vs. Adobe RGB

  • sRGB: display correctly on Web
  • Adobe RGB: usually better for printing, but wont display correctly on web without conversion.

Color Model ​

A color model is the mathematical way of describing colors, independent from physical devices.

RGB ​

  • Primary colors: Red, Green, Blue
  • Additive color mixing
    • Nothing = black
    • Red + green = yellow
    • Green + blue = cyan
    • Blue + red = magenta
    • Red + green + blue = white
  • Usage
    • Web
    • TV, Computer Display

CMYK ​

  • Primary colors: Cyan, Magenta, Yellow, blacK

  • Subtractive color mixing

    • Nothing = white
    • Cyan + magenta = blue
    • Magenta + yellow = red
    • Yellow + cyan = green
    • Cyan + magenta + yellow = black

  • Usage

    • Color printing

RYB ​

  • Red, Yellow, Blue are not primary colors
  • Pigments absorb light
  • Gamut is smaller than CMYK

HSL ​

More closely align with the way human vision perceives color-making attributes

  • Hue:human perceived colors
  • Saturation (饱和度)：colorfulness, distance to gray
  • Lightness (亮度):relative degree of black and white

RGB, CMYK, HSL are perceptually non-uniform

Lab ​

a perceptual uniform color model - Lab

Lab approximate human vision

  • Luminance: human perception to lightness (black to white)
  • a: green (-a) to red (+a)
  • b: blue (-b) to yellow (+b)

Color map ​

Colormaps are used to characterize the mapping from the data domain D to color scales C in visualization.

f:D→C

Based on data attribute types, different colormaps shall be applied.

Categorical Colormap ​

There is no implied relationship between different levels of a variable

TK1G1C

Try to order the colors

  • Not able to order
  • Hue is not a perceptual cue for order

Quantitative colormap ​

Quantitative colormaps are used for presenting quantitative data values

  • Sequential colormap for sequential ordering
    • Temperature, population density, rainfall
  • Diverging colormap for diverging ordering
    • Temperature relative to average (cooler vs. warmer)
    • Above or below average (+Δ vs. −Δ)
    • Profit increase or decrease (loss vs. profit)

Sequential Colormap ​

Try to order the colors

  • Luminance is a strong perceptual cue for order

Suitable for ordered data attributes

  • Single-hue

For the Sequential plots, the lightness value increases monotonically through the colormaps.

Diverging Colormap ​

For the Diverging maps, we want to have monotonically increasing L values up to a maximum, which should be close to L=100, followed by monotonically decreasing L values.

  • Suitable for ordered data with diverging values

Semantics of colors ​

Color selection shall consider semantics

  • Stock markets at US vs. China

  • Dreaded rainbow colormap

  • Saturation affects area perception