TAPESTRY: The Art of Representation and Abstraction

Rendering Overview

What's it all about?

Imagine sitting at a table in a darkened room. We can feel the table, and on the table we feel a number of cube-shaped objects. We can know they are there, we can feel their surfaces, but we can't see them. Now, imagine striking a match. We can now see the cubes. We realize that some cubes are multi-colored, each face painted with a different color. Others are polished, or glass, or metal. Some have square corners, others are rounded like giant dice. We can tell this because the redish-yellow light from the match is reflected in different ways from the surfaces of the cubes, and because we have had a lifetime of looking at objects under different lighting conditions.

The cubes each cast a shadow on the table too. If someone behind us switches on a light, we don't have to look at the light to tell whether it is a flash-light, or a candle, or a flourescent lamp. Some of this information comes from the color of the light, some from the character of the shadows. Reality is a complex and amazing place!

The challenge for the computer graphics programmer is to re-create the visual effect, without the actual materials and innumerable photons of the real world.

We see objects because of the light reflected from or transmitted through their surfaces. Differences in surface quality, such as color, roughness, and transparency have an effect on the appearance of the shape. The light which falls on the surfaces is emitted by one or more light sources. Sometimes the light is blocked, causing a shadow to fall on the surface. Finally, if the light sources and objects are moved around, the surfaces of the object will change in appearance (getting darker or lighter, for example). What we see may depend on

The material/surface quality of the object
The light distribution qualities of the light source
The relationship between object, light source(s) and viewer

Each of these issues is subject to a variety of approaches, ranging from ignoring the phenomenon, to simple approximations, to complicated simulations of the "real" phenomena. Furthermore, the developers may make different committments to the questions, so that we could have a simple approach to material definition, but a sophisticated definition of light sources. The number of combinations is obviously quite large. However, there tends to be a "linking" of the different issues. After all, why use a sophisticated description of a light source if the actual rendering algorithm uses a simple approach to calculating what that light does within the scene?

Last updated: April, 2014

[report bug]