MCIS 625: Computer Graphics
Winter 2004

WEEK 4

Instructor: Dr. Michael Laszlo


Assignment

  1. Review my slides on rasterization.
  2. Optional reading assignment: Sections 5.1 through 5.8 of Foley et al. Note that Section 5.1 covers mathematical preliminaries some of which you already may be familiar with and some of which not.You may wish to begin with Section 5.2 and turn back to Section 5.1 as necessary during your reading.

Rasterization
Modern computer graphics is based on the raster display, which represents graphical primitives such as lines, polygons, and curves in terms of its component pixels. The primitives are stored in a memory buffer known as a frame buffer, whose memory cells correspond one-to-one with the matrix of pixels on the screen. The screen pixels are grouped into horizontal rows known as scan lines; display proceeds one scan line at a time.

In CRT technology, each screen pixel is composed of three dots of phosphor for the colors red, green, and blue, whose intensities combine to form the pixel's color at any given time. To display the image stored in the frame buffer, a video controller uses the values stored in the frame buffer to control the intensity of three electron guns (one each for red, green, and blue). At each pixel, the intensities of the electron guns cause the dots of red, green, and blue phosphors to glow at specific brightness levels, thereby producing the pixel's color. The intensity of the electron guns change as their beams are deflected across the screen, scan line by scan line, according to the contents of the frame buffer. The screen is refreshed in this way about 30 times per second, to avoid flicker and to ensure that it accurately reflects the frame buffer's changing content.

There are two principal color modes that determine how the values in the frame buffer are used. Under the RGB color mode, each frame buffer cell stores a color represented by an RGB value. This value determines the brightness level for the red, green, and blue components for the cooresponding screen pixel. (We will look at color models later in this course.) Under the color look-up table mode, each frame buffer cell stores a numeric index into a table—the index references an entry in this table. The table, known as a color look-up table (or CLUT), stores RGB values representing colors. Thus color look-up table mode introduces a level of indirection: The value in each frame buffer cell references an entry in the color-lookup table, and the table entry specifies the color for the corresponding screen pixel.

The disadvantage of raster systems arises from the discrete nature of the pixel representation. Graphical primitives, such as line segments, circles, curves, and polygons, are described mathematically. A line segment is specified by the position of its two endpoints, a circle by its radius and the position of its center. Curves may be described in a number of ways; one approach specifies a curve's endpoints and control points. A polygon may be specified by the position of its vertices. In each of these cases, it is necessary to translate the primitive's mathematical description to a set of pixels—the pixels that the primitive covers. The primitive can then be "painted" into the frame buffer. This process of translation to pixel format is known as rasterization. It is also sometimes called scan conversion, which suggests the idea of conversion from a random scan (or mathematical) representation.


[ Home | Course | Syllabus ]