The Commission Internationale d'Eclairage (CIE) chart defines a standardized color space.
The CIE chart is a useful tool for specifying colors.
It represents hue and saturation on a two dimensional chart.
Fully saturated hues lie along the outside edge with desaturated colors toward the center of the chart.
This CIE X-Y chart was later revised to better reflect human perception of color.
While the CIE chart is a very useful tool, the color space it defines does not provide an intuitive model for our color vision system or the devices we use to reproduce color.
The following diagram is an approximate representation of the CIE color chart.
|
|
Since our vision system uses three different sensors to selectively detect the visual spectrum, the color space they define is inherently three dimensional.
We can best visualize this three dimensional color space as a cube.
One corner represents zero excitation for all three sensors or the color we call black.
There is a sensor vector along each of the three edges which leave this zero excitation corner.
These vectors represent the extent of the stimulus for the Rho, Gamma and Beta sensors.
This cube has white at the corner directly opposite black.
It has a primary color (red, green or blue) in the corner opposite its complimentary color (cyan, magenta or yellow - the secondary colors).
Here is the visualization of the color space defined by the Rho, Gamma and Beta sensor stimulus vectors.
The RGB Color Cube |
|
There is a line connecting the black and white corners of the cube.
This is the line of neutral gradient or you might think of it as the 21-step stepwedge.
Neutral Gradient Line |
|
There are lines connecting each of the primary colors (RGB) with their corresponding secondary colors (CMY).
These are the lines of primary-secondary gradient.
These are the lines along which we make color correction judgments for prints of color images.
Primary-Seconday Gradient Lines |
|
There is a triangular plane connecting each of the primary colors (RGB).
Notice also that all the fully saturated colors live on the surface of the cube.
Plane of the Primary Colors |
|
There is also a triangular plane connecting each of the secondary colors (CMY).
Notice that this plane crosses the neutral gradient line at a point closer to white than black and that the plane of the primaries crosses at a point closer to black than white.
We generally expect secondary colors to reproduce lighter than primary colors in black and white images.
Unlike most other visualizations of color, this one based upon sensor sensitivity vectors meets our expectation.
The color cube, while perhaps a little more complex to visualize, is a very good model for gaining a better understanding of color.
Plane of the Secondary Colors |
|
Each of the primary and secondary colors have their own paths from black to white.
The RGB primaries move away from the black corner along three separate paths.
Whenever a vector moves along a corner of the cube, it is changing in a single variable - in this case, the RGB primary itself.
Once the RGB primaries reach their fully saturated corn of the cube, new vectors move diagonally across a cube side, toward the white corner.
When a vector moves diagonally across a cube side, it is changing in two variables.
To move from any one of the fully saturated primaries toward white, an equal amount of the other two primaries are added.
For example to move from the red corner to the white corner, green and blue are added.
Primary and Secondary Gradient Vectors |
| Secondary colors move from black to white corners in the opposite way from primary colors. The move from black to fully saturated as diagonals on a cube surface (two variable changes). In moving from fully saturated corners to the white corner, they travel along an edge (single variable change). |
|
The fully saturated outer edge of the CIE chart exists as path around the outside surface of the color cube.
Edge of Saturated Hues |
| The color cube defined by our color sensitivity vectors applies equally well to many of the systems we use to record color and to reproduce color because they are also three color systems. Below is an illustration of the 3 vector representation for an RGB value as used with 24-bit color on a computer. The RGB value of [102,140,166] represents 102/255 or 0.40 red, 140/255 or 0.55 green, and 166/255 or 0.65 blue. Three component color is easy to visualize as three vectors describing a location within the color cube. There is no equivalently intuitive description of RGB values on a CIE color chart. The cube is an over simplification, since this color space is as non linear as Einstein's warped time and space that it lives within. Still, the color cube is quite a useful first order approximation concept for understanding how we perceive color. |
| The 216 color palette used by web browsers to down color 24-bit images for 8-bit video cards is a real world realization of this color cube and provides another good way to visualize 3 sensor color space - Web Browser Color Space. |
| Previous |  |
|||
| Contents |  |
|||
| Next |  |
|||
Color Space