Name the colors of the rainbow in the correct order...
1. The order is not intuitive to everyone.
2. Categorization is intuitive but unwanted and not consistent.
3. The lightness is not linear.
Colors are mixable.
Any 3 colors can span a color space.
What is the most useful coordinate system?
Color Matching Functions (CMF):
spectral colors <-> color space coordinates.
Color spaces allow only either + or -
Any primary colors have partly negative CMF.
RGB covers most of the spectrum accurately.
'Darker' primaries (RGB) for + spaces,
'Brighter' primaries (CMY) for - spaces.
RGB is not uniquely defined!
avoid when possible
$Y \sim$ luminance, $XYZ$ is additive
$x=\frac{X}{X+Y+Z}, \; y=\frac{Y}{X+Y+Z}, \;z=1-x-y$.
There is no pink in the rainbow!
CIE 1931 xyY is (nearly) perceptually uniform.
Calibrated for "Standard Observer" (2°, fovea).
Standard Daylight: D65 ($T=6504 K$)
Lightness and saturation is more important than hue in guiding attention. But hue helps to indentify and distinguish values.
For single-hue palettes nonlinear lightness is not problem.
For the multi-hue 'Hot' palette nonlinear lightness is a problem.
Bezier curves in the color space can smoothen lightness.
The lightness can simply be linearized.
Brightness: how much light a colored area seems to emit.
Lightness: brightness relative to a reference white (~relative luminance)
The color spectrum is not perceptually uniform.
Words for colors shapes out perception.
Sequential: ordered hues, linear lightness
Diverging: ordered hues with center, (bi)linear lightness
Cyclical: cyclical hues, cyclical lightness
Qualitative: distinctive but uniform hues, distinct color names, varying lightness
The colors need to fit the data, and be accessible.
The colors need to fit the application.
The colors should guide the attention and interpretation.
The end product should be pleasant to look at.
Data ~ Colors
Data Type ~ Palette Type
Context ~ Number of Hues
Data Values ~ Lightness Values
Value Types ~ Continous/Categorical
Missing Data Points ~ Extra Color
Value Distinctions ~ Color Name Distinctions
Accessibility
Color deficiencies (8% in men, 0.5% in women)
RGB screens
CMYK prints, black/white prints
Low contrast beamer
Color perception is strongly influenced by context.
Look out for influences by
background color (dark-is-more, opaque-is-more)
amount of 'ink' (alpha, linewidth, bar height, ...)
neighbouring and overlapping colors within figure
existing color schemes in related figures
Pysiological Basis: The Opponent Process
Colors themselves have meaning.
Colors can evoke emotions.
Colors are associated to concepts.
-> highlight, discriminate, guide visual flow
depend on context and cultural background
Beware of conventions in the field.
one Color - many Concepts
many Colors - one Concept
Aestetics are directly related to Understanding and Engagement.
Aesthetics vs. Discrimination
Color Harmony?
Color preferences
~ preferences of associatiated concepts
Taste is subjective.
Less is more.
Use colors sparingly. Use subtle colors.
Use color combinations from nature.
Copy from your favorite artists.