Color Systems Defining Colors for Digital Image Processing Various models exist that attempt to describe color numerically. An ideal model should be able to record all theoretically visible colors in the form of numbers, thereby making every color exactly reproducible at any time. The fact that various color modelling systems exist is due in part to the existence of different sources of light (reflected or radiated), but also because it is possible to interpret the composition of color in different ways. Colors can be described in terms of a simple mixture of primary colors, or alternatively in terms of their own defining characteristics, such as hue, saturation, and brightness. The two most important approaches to the description of color are based on the technique of trichromatic synthesis, or, in other words, synthesis using three primary colors. Additive Color Synthesis Additive color synthesis produces colors by superimposing red, green and blue colored light sources on top of each other. The exact shade of the resulting color depends on the strength of the three sources relative to each other. Maximum brightness of all three sources produces an impression of white, and when none of the three sources shines at all, the result is black. Conventional CRT monitors, LCD displays, and the CCD chips in digital cameras work on this additive principle. These devices all use red, green and blue colored filters to control the behavior of the source light. The resulting system of color description is called RGB, due to the primary colors of red, green and blue, that it is based upon. While monitors use very small primary-colored filters to control and mix colors, the filter principle can also be used the other way around. A CCD registers Additive color synthesis: synthesis using incandescent light sources
Color Systems: Defining Colors for Digital Image Processing Page 2 of 7 light that has been filtered into the three primary colors and uses this filtered light to attribute a certain color to each pixel, (comparable also to the way the human eye works). Subtractive Color Synthesis Subtractive color synthesis is based on the principle of absorption and reflection. Colored particles that only reflect a very small band of wavelengths (and absorb all others) are applied to a white surface. The basic colors used here are the complementary colors of the three primary colors used in additive synthesis, namely: cyan, magenta and yellow. Each colored particle absorbs large parts of the visible spectrum, consequently, maximum absorption occurs where all three colored particles are applied. Here, the resulting point reflects no light and is therefore perceived as black. Subtractively synthesized colors are used in slide films, color photographic paper, and for all color printing processes, such as offset or inkjet printing, for example. Four-color printing, (the most widespread printing method,) uses black in addition to the three subtractive base colors, which increases contrast and optimizes the neutral tones in darker colors. This color system is called CMYK (K being short for Key plate, the plate traditionally used to print the detail of a picture using black ink). Modern inkjet photo printers also use the subtractive method to create their colors, and often use additional, subsidiary colored inks (in lighter cyan, magenta and black or in more intense red and green tones) in order to produce the widest possible range of colors. The RGB Color Model The RGB model interprets the colors of the spectrum as an additive mixture of the primary red, green and blue colors. The range of colors thus produced constitute the RGB color space, which can be visualized using various three-dimensional models. One such model, which is also suitable for the processing of color images Subtractive color synthesis: synthesis using colored particles
Color Systems: Defining Colors for Digital Image Processing Page 3 of 7 on a computer, is the color cube. In order to express colors numerically, the cube is described using a system of coordinates whose base point is black (i.e., no color - where all three base components have the value 0). The x-axis represents the red component, which reaches its maximum value of 255 (pure red) at its far end. The green and blue components are represented in the same way by the y- and z-axes, respectively. Color Depth The term color depth represents a gage for the number of possible colors contained in a particular color system. In the illustration, each axis is divided into eight segments, thus defining a color space with 8 x 8 x 8 (= 512) elements. This color system therefore contains 512 different color shades in the range between black and white. Computer calculations are based binary numbers (i.e., using only the digits 0 and 1), and numbers are specified using powers of 2 rather than the powers of 10 we are used to. The decimal digit The color cube model embodies and simultaneously delimits the RGB color space using the three color axes: blackred, black-green and black-blue. If we imagine the the RGB color space as a cube made up of many smaller cubes, each small cube represents a shade of the RGB spectrum. If we then put the cube into a system of coordinates spanned by red, green and blue axes, every color shade can be given a coordinate consisting of three color values. These spacial coordinates correspond to the proportions of the primary colors present in the shade in question. The relationship between the coordinates describes the RGB definition of a color.
Color Systems: Defining Colors for Digital Image Processing Page 4 of 7 Visualizing the RGB color cube To illustrate the model clearly, the diagram above shows the color cube unfolded like a paper cut-out. The other two diagrams show the RGB color cube viewed from different angles in three-dimensional space.
Color Systems: Defining Colors for Digital Image Processing Page 5 of 7 8 is equal to the binary expression 2 3 (2 x 2 x 2). In this system, a triple-digit (or 3-bit) number is required to generate eight elements. The color space in the example has a color depth of 3 bits per axis, or a total color depth of 9 bits. The color depth usually used to display colors on a computer monitor is 24-bit, which corresponds to 16.7 million different colors, or 256 (8-bit) color hues per primary color. These values also correspond to the values 0 to 255 in the RGB model. This system is called True Color, because the interval between two color hues is so small that a person with normal vision cannot tell the difference between two neighboring shades, even in delicate color runs or subtle nuances in the image. Quality digital cameras digitize their pictures using even finer mechanisms. The Nikon D200 produces RGB images with 12 bits (or 4,096 hues) per color. As well as providing a higher degree of differentiation within the image itself, this color depth also allows images to be manipulated with little or no loss of quality. If similar processes were to be applied to This RGB color cube is divided into 8 segments along each axis. The color space thus defined contains 8 x 8 x 8 = 512 different color hues. This represents a total color depth of 3 x 3 = 9-bit for all three color channels counted together. The red-green plane of the color cube, shown with different color depths: at the 8-bit level, the naked eye can no longer recognize any clear gradation between the individual color hues.
Color Systems: Defining Colors for Digital Image Processing Page 6 of 7 images with 8-bit color depth, the result would be visible loss of image quality in the form of jagged color runs or uneven rendering of tones. The HSB Color Model The RGB system for mixing colors is based on the perceptive mechanism of the human eye: the retina contains receptors that are sensitive to either red, green, or blue light and the brain assembles the signals it receives into a multicolor image. In spite of this similarity to a biological model, the RGB model itself is not very intuitive. Can anybody really define a color in terms of RGB without having to think very hard about it? Irrespective of this bio-technical color perception, an evaluation of color on a practical basis can be more intuitive. The HSB color model, for example, is a clear and intuitive model. While the RGB model represents colors according to their primary color components, the HSB model represents and describes colors using three quite different characteristics, namely: hue, saturation, and brightness. The HSB model can be visualized three-dimensionally as a cylinder stood on its end. The color hue is represented by the angle between the "ground" and the axis of the cylinder; the saturation (or intensity) of the hue is represented by the horizontal distance to the cylinder axis; and brightness corresponds to the vertical position of the color within the cylinder. Although the HSB color model is not often used to actually define colors, it is commonly used for choosing and refining colors in computer programs and is the basis of many DTP color correction tools. Similarly, the RGB model is the most utilized model when it comes to describing colors emitted by incandescent light sources, and the CMYK model is often employed to describe reflected colors. The colors of the RGB color space can be effectively depicted using the HSB color cylinder.
Color Systems: Defining Colors for Digital Image Processing Page 7 of 7 The options H, S, and B in the color picker dialog in Adobe Photoshop allow you to choose your colors according to the HSB model. Here, the HSB cylinder is shown as a two-dimensional plane. Option H represents a longitudinal section through the cylinder and allocates a color depending on its saturation (position on the horizontal axis) and brightness (position on the vertical axis). Option S represents the entire color spectrum for the same level of saturation (equivalent to cutting a theoretical barrel out of the cylinder). Option B shows the entire color spectrum for a constant level of brightness (equivalent to a horizontal slice of the cylinder). Related Topics: > Color Spaces