Overview

In this course we will primarily talk about generating images on a computer screen from a physical model. But first we need to understand the display devices prevalent today.

Displays

The most commonly used device is CRT (cathode ray tube). It is the same technology that is used in most commercial television sets today. The primary components of a CRT are: When an electron beam strikes the phosphorescent compound (deposited on the face of the CRT), it emits light. The beam "lights up" different spots on the screen along a horizontal line, also called a scan line. Once all spots or pixels on a line have been lit, the beam retraces and starts working on the next scan line, repeating until it covers the whole screen. We call this one frame. The pattern that is formed by actually lighting some pixels while leaving the others "dark", by turning the beam off is what we call an image. Such display is called scan-line display or raster display. (Sometimes a frame may be broken into two fields: In the first one, the beam scans only the odd lines and in the second field it scans only the even lines. When the two fields display the same image, flickering* can be reduced.)

Unfortunately, once a spot is "lit", it does not continue to be lit. It dissipates energy. Persistence is defined as the time it takes for a spot to decay to 10% of its original intensity. Most CRTs have persistence varying from 10 to 60 microseconds. If we want to keep a spot lit, we must re-light it, or "refresh" it. It turns out that for most humans refreshing the whole screen 60-70 times a second allows their brain to integrate the temporally distinct spots into one continuous image. The frequency at which we do not perceive any flicker is termed the critical fusion frequency. Most television screens are refreshed about 30 times a second. (*) However, in each frame two fields are drawn, which is almost like drawing the same image twice. Thus an effective refresh rate of approximately 60Hz is achieved.

The pixel positions that we light do not have a well defined boundary. The center is the brightest and as we move away the intensity goes down. The intensity along any cross-section of the spot follows a Gaussian curve. The resolution of a display is measured in terms of the number of discernibles pairs of black and white lines that can be drawn per inch. Typically, adjacent pixels are distinguishable if their 60% radii abut, i.e. the inter-spot distance is twice the spot-radius at which the intensity falls to approximately 60% of the maximum intensity.

By turning the electron beam on or off when pointed at a certain pixel position, we can get a monochrome image (which may be described by 1 bit per pixel: which we also call bitmap.) It is possible to actually vary the intensity of the spot, thus allowing us to generate a gray scale image. Most good quality graphics monitors can display 256 different levels. (Some very high quality monitors display more) In order to generate color, we typically use a combination of Red, Green and Blue colors. Three different materials are deposited at each pixel position. These emits Red, Green and Blue light, respectively, when "lit". We use three separate beams to strike at this triad of materials. A metal mesh close to the screen ensures that each material is struck only be its corresponding beam. There are two common arrangements for the R G B dots, a triangular arrangement (delta-delta) or a linear pattern (in-line).

From the software point of view, most displays in use today, including hardcopies, are similar-- they display or print a rasterized image. An array of pixels, or a pixmap. Hence it is possible to generate a pixmap independent of the technology. This pixmap may later be expanded ("blurred") or contracted ("sharpened") to display the final image on a variety of different platforms and screen sizes. Note, however, that typically for interactive graphics, we generate a pixmap of the size appropriate for the target window.

Other common display technologies include LCD (liquid crystal display), Plasma based display, Electroluminescent displays etc. The basic principle in all these devices is the same: light up an array of pixel positions on the screen. LCDs do it by exploiting the spiral crystalline structure of some compounds. In their natural state lights passing through such compound is "twisted": the direction of polarization of light is changed (by approximately 90 degrees). If horizontally polarized light is projected onto the crystals, and a vertical polarizer is placed at the back-end, the light passes through. However, by applying charge to these crystals, they can be "straightened" out. Now a horizontally polarized light is not twisted and hence is not "let through" the vertical polarizer. By using a set of horizontal and a set of vertical conductors (on the opposite sides of the crystal layer) we can finely control the charge at the grid points -- which form our pixels. Thus at each pixel the light can be passed or stopped. Color is again obtained by placing a triad of materials at each pixel position with fine and independent control of charge across each material. Oblique angles of viewing can cause these displays to get "washed out". Currently they do not produce as good quality visual perception and resolution as achieved by CRTs, but they are less bulky and more portable at similar price.

Plasma panel displays use a charge across gas molecules (e.g. neon) to cause the molecules to break down into ions and cause a spark. Thin film electroluminescent displays use phosphorescent material instead of gaseous material to generate precisely positioned light-emissions.

Stereo Display

Interaction Devices

  • Keyboard
  • Mouse
  • Joysticks
  • TrackBall
  • Light pens
  • Touch Panels
  • Digitizer tablets
  • Force feedback arm
  • Data gloves
  • Trackers
  • Gyrometers etc.