The RenderMan Interface - Paul Bourke

Section 8
INTRODUCTION TO THE SHADING LANGUAGE
Remarkably realistic images can be produced with a few fairly simple shapes made from
interesting materials and lighted in a natural way. Creating a photorealistic image requires
the specification of these material and lighting properties. This part of the document describes
the Shading Language, which is used to write custom shading and lighting procedures
called shaders. Providing a language allows a user to extend shading models or to
create totally new ones. Models of light sources with special lenses, concentrators, flaps or
diffusers can be created. The physics of materials can be simulated or special materials can
be created. This is done by modeling the interaction of light at the surface and in the interior
of a region of space. Material types can also be combined, simulating the many coats of
paint or finish applied to a surface. Providing a shading language also allows many of the
tricks and shortcuts commonly performed during production rendering to be accommodated
without destroying the conceptual integrity of the shading calculations. Visualizing
the results of scientific simulations is also easier because shaders can be written that produce
a surface color that is based directly on the results of a computation. For example, it
is possible to write a shader that sets the surface color based on temperature and surface
curvature. Shaders can also be used to modify the final pixel values before they are written
to the display.
The Shading Language is a C-like language with extensions for handling color and point
data types. A large number of trigonometric and mathematical functions, including interpolation
and noise functions, are provided. Color operators are provided that simulate the
mixing and filtering of light. Point and vector operators perform common geometric operations
such as dot and cross product. A collection of commonly used geometric functions
is also provided. These include functions to transform points to specific coordinate systems.
Common lighting and shading formulas, such as ambient, diffuse, specular, or phong,
are available as built-in functions. Built-in texture access functions return values from images
representing texture maps, environment maps, and shadow depth maps. The texture
coordinates given to these functions can be either the presupplied texture coordinates or
values computed in the Shading Language. Since texture map values are just like any other
value in the language, they can be used to control any aspect of the shading calculation.
There is in principle no limit to the number of texture maps per surface.
The Shading Language can be used for specifying surface displacement functions such as
ripples or nubs. Shading Language functions are also used for pixel operations. This type
of shader is referred to as an imager. Imagers are used to do special effects processing, to
compensate for non-linearities in display media, and to convert to device dependent color
107