Fundamentals of Computer Science I: Media Computing (CS151.01 2008S)
In a related reading, you learned about raster graphics and ways to set and get colors at various locations in a raster image. However, in your work, you were limited to either the colors already in the image or the colors that someone has named for you.
One of the great advantages of computational image making is that it is possible to describe colors that do not have a name. In fact, it is often better to use a more precise definition that is possible with name. After all, we may not agree on what precisely something like “blood orange” or “Oregon salmon” means. (Do you know how the color “Seattle salmon” relates to the color “Oregon salmon”?)
In fact, it may not only be more accurate to represent colors non-textually, it may also be more efficient, since we will often need to look up color names in a table somewhere.
The most popular scheme for representing colors for display on the computer screen is RGB. In this scheme, we build each color by combining varying amounts of the three primary colors, red, green, and blue. (What, you think that red, yellow, and blue are the primary colors? It turns out that primary works differently when you're transmitting light, as on the computer screen, than when you're reflecting light, as when you color with crayons on paper.)
So, for example, purple is created by combining a lot of red, a lot of blue, and essentially no green. You get different purple-like colors by using different amounts of red and blue, and even different ratios of red and blue.
When we describe the amount of red, green, and blue, we traditionally use integers between 0 and 255 to describe each component color. Why do we start with 0? Because we might not want any contribution from that color. Why do we stop with 255? Because 255 is one less than 28 (256), and it turns out that numbers between 0 and 255 are therefore easy to represent on computers. (For those who learned binary in high school or elsewhere, if you have exactly eight binary digits, you can represent exactly the integers from 0 to 255.)
If there are 256 possible values for each component, there are therefore 16,777,216 different colors that we can represent in standard RGB. Can the eye distinguish all of them? Mine certainly can't. Nonetheless, it is useful to know that this variety is available, and many eyes can make very fine distinctions between nearby colors.
Let us now turn to the primary procedures we will use to work with RGB colors.
We build a new color with the
) procedure. For example,
we might set the center of a 9x5 image to a purple-like color with
(image-set-pixel! canvas 4 2 (rgb-new 128 0 128))
If we have color created by
rgb-new or extracted with
image-get-pixel (or computed in one of the ways you'll
learn over the next few weeks), we can extract the red, green, and blue
(rgb-green , and
(rgb-blue . For example,
we can remove the green component of the upper-right pixel of a 9x5
sample image with
(define color (image-get-pixel sample-image 8 4))
(image-set-pixel! sample-image 8 4 (rgb-new (rgb-red color) 0 (rgb-blue color))
These procedures can also be useful for telling us a bit about the named colors. For example, let's figure out the components of the color one of DrFu's developers named “blood orange”.
(define blood-orange (cname->rgb "blood orange"))
At times, we simply want to find out the three components of the
color so that we can think about the meaning of those components.
(That is, we will not compute with the values, but simply look at them.)
While we could use each of
rgb-blue, as in the sample code above, that requires
a bit more effort than seems necessary. Hence, DrFu also provides
( that let us view the three
(204 17 0)
(rgb->list (cname->rgb "cornflower blue"))
(66 66 111)
(rgb->string (cname->rgb "cornflower blue"))
In creating works, many artists and visual designers consider the applications of complementary colors. A pair of colors is complementary if the sum of the two colors is a kind of grey (including black or white). What does it mean to sum two colors? Well, it turns out that complementarity is really defined only for a different representation of colors (hue, saturation, and value, or HSV). Nonetheless, we can come close to simulating it in RGB, so we will call complementary colors defined using their RGB values pseudo-complementary colors.
In RGB, we can add the colors by adding the corresponding components (capping the sum at 255) or by averaging the corresponding components. We'll use the former technique, because it can be a bit easier to analyze capping.
For example, the pseudo-complement of green (0/255/0) is magenta (255/0/255) because when we add them together, we get 255/255/255, which is white.
Depending on what you accept as the definition of “grey”, colors can have many pseudo-complements. For example, consider the color 128/0/0, which is similar to blood red. One logical pseudo-complement to that color is 127/255/255 (similar to medium turquoise), since when we add the two colors together, we get 255/255/255, which is still white. However, one might also consider 0/128/128 (a color with no good name, but which DrFu thinks is close to steel blue) as a pseudo-complement, since when we add the two together, we get 128/128/128, a nice medium grey.
In general, when we say “pseudo-complementary color”, we mean the one which, when we add the RGB components to those of the first color, we get white. When we ask for multiple pseudo-complements for the same color, we'll mean those that, when added, give us a color in which all three components are the same (that is, a version of grey).
Copyright (c) 2007-8 Janet Davis, Matthew Kluber, and Samuel A. Rebelsky. (Selected materials copyright by John David Stone and Henry Walker and used by permission.)
This material is based upon work partially supported by the National Science Foundation under Grant No. CCLI-0633090. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
This work is licensed under a Creative Commons
Attribution-NonCommercial 2.5 License. To view a copy of this
or send a letter to Creative Commons, 543 Howard Street, 5th Floor,
San Francisco, California, 94105, USA.