Assignment 1: Summary of Problem from Chapter 8
Arrangements and Duality
Supersampling and Ray Tracing
February 3, 2000
Table of Contents
An important topic in computer graphics today is the generation of
3-dimensional scenes. A technique that helps to do this is ray tracing.
A synthesis of what is involved in ray tracing:
We are given a 3-dimensional scene, a light source, and a view point. To
render this scene of the screen, we must determine
which object is visible to every pixel on the screen.
We must also find out the intensity of light emitted from the object in
the direction of the view point at the particular point which the pixel
'sees'. We must now take into account how much light that point is
receiving from the light source, either directly or via reflection off
A problem is that a pixel is not a point, but a small square area. Most
of the time, the pixel is completely covered by one object. On the borders
the pixel may see more than one object. We could assume falsely that the
whole pixel was covered, but this would give a jaggedy looking image. We
can improve this by taking into account the percentage of hit or miss, and
adjust the intensity of the pixel with this
information. If more than one object is visible inside the pixel, we
could mix the object intensities.
To incorporate all the different pixel intensities into the ray tracing
scheme, we shoot more than one ray per pixel. Distributing these rays
uniformly again gives us problems because the eye perceives regularity
easily. A random pattern improves on this, but we still want the sample
points to be distributed so that the number of hits is close to the
percentage of covered area.
render - generating an image
pixel - extremely small dot on screen that displays color
intensity - brightness
Computational Geometry Bibliography
The Geometry Center
Computational Geometry Software
Los Alamos e-print archive
This assignment took me 2.5 hours to complete.
Back to the main page