Output / Input Hardware
Overview | Basic types | 3D Video
![]()
Visual display unit :
|
|
Colour depth
|
3D graphics technology is all about deception: how to fool the viewer into thinking that a 2-D image (the only type of image displayable on a "flat" screen) - also has depth. |
![]() |
To do this all the information required to create the 3D scene is
calculated and stored in system memory. This includes the coordinates of each object
(height, width and depth), surface characteristics, lighting effects and perspective
information (relationship to other objects). Then in a process called
"rendering" the 3D graphics card translates this data pixel by pixel to create
the 3D image for the flat screen. Since rendering makes such huge demands on computer hardware the manufacturers of the cards have to make compromises between maximum resolution ("blockness"), colour depths and refresh rates (30 frames per second is used for standard video). |
To take advantage of a cards 3D acceleration features software designers often write for specific chip sets. Unfortunately this leads to problems where a game written for one chipset will not work on another. To assist developers a range of API's (Applicaation programming interfaces) exist. An API is a layer of softwsare between the software and the chipset. The 2 leading APIs are Open GL (Silicon graphics designed for high end graphics as seen in many movie special effects) and Direct 3D (Microsoft designed for Windows 95)
Alpha blending
- Defines opacity of objects. Blending allows two textures to be mixed to create the illusion of transparency (eg. fog or glass door)
Anti-aliasing
- Lines that are on an angle tend to appear jagged. Anti-aliasing smudges the pixels on either side to smooth the jagged lines.
Fogging
- Blending the surface texture with a shade of white. When hardware based can speed up rendering considertably.
Gourand shading
- Manipulates the 3D surfaces to achieve subtle lighting effects.
Perspective correction
- Process by which textures are adjusted to compensate for changes in the viewers location.
Polygon throughput.
- 3D objects are composed of polygons - usually triangles. The greater the number of polygons the clearer the image. A higher polygon throughput means faster rendering.
Texture mapping.
- Applies a texture to a surface. A grainy texture for example can be used for a brick wall.
Z-buffering.
- Where two objects intersect, the processor must decide which of the objects surfaces are hidden and which are visible. This information is tracked in the z-buffer.
Ref: Clarkson, N. and Tsang, S. (1997,
Aug)
[Rev: 1/2/98] 14/10/97 © 1997-98 V/2-Com (Verhaart), P O Box 8415, Havelock North, New Zealand.