Category : Printer + Display Graphics
Archive   : RAYSH386.ZIP
Filename : LIGHTS.TEX

\chapter{Light Sources}

The lighting in a scene is determined by the number, type, and nature
of the light sources defined in the input file. Available light sources
range from simple directional sources to more realistic but computationally
costly quadrilateral area light sources. Typically, you will want to use
point or directional light sources while developing images. When
final renderings are made, these simple light sources may be replaced by
the more complex ones.

No matter what kind of light source you use, you will need to
specify its intensity.
In this chapter, an {\em Intensity\/} is either a red-green-blue triple
indicating the color of the light source, or a single value that is
interpreted as the intensity of a ``white'' light.
In the current version of {\rayshade}, the intensity of a light does
not decrease as
one moves farther from it.

If you do not define a light source, {\rayshade} will create a directional
light source of intensity 1.0 defined by the vector (1., -1., 1.).
This default light source is designed to work well when default
viewing parameters and surface values are being used.

You may define any number of light sources, but keep in mind that
it will require more time to render images that include many
light sources. It should also be noted that the light sources themselves
will not appear in the image, even if they are placed in frame.

\section{Light Source Types}

The amount of ambient light present in a scene is controlled by a
pseudo light source of type {\em ambient}.

\begin{defkey}{light}{{\em Intensity\/} ambient}
Define the amount of ambient light present in the entire
scene.
\end{defkey}

There is only one ambient light source; its default intensity is
{1, 1, 1}. If more than one ambient light source is defined,
only the last instance is used. A surface's ambient color
is multiplied by the intensity of the ambient source to give the
total ambient light reflected from the surface.

Directional sources are described by a direction alone, and are useful
for modeling light sources that are effectively infinitely far away
from the objects they illuminate.

\begin{defkey}{light}{{\em Intensity\/} {\tt directional} \evec{direction}}
Define a light source with the given intensity that is
defined to be in the given direction from every point
it illuminates. The direction need not be normalized.
\end{defkey}

Point sources are defined as a single point in space. They produce
shadows with sharp edges and are a good replacement for extended
and other computationally expensive light source.

\begin{defkey}{light}{{\em Intensity\/} {\tt point} \evec{pos}}
Place a point light source with the given intensity at the
given position.
\end{defkey}

Spotlights are useful for creating dramatic localized lighting effects.
They are defined by their position, the direction in which they
are pointing, and the width of the beam of light they produce.

\begin{defkey}{light}{{\em Intensity\/} {\tt spot} \evec{pos} \evec{to}
{$\alpha$} [ $\theta_{in}$ $\theta_{out}$ ]}
Place a spotlight at \evec{pos}, oriented as to be pointing at
\evec{to}. The intensity of the light falls off as
$(cosine \theta)^{\alpha}$, where $\theta$ is the angle between the
spotlight's main axis and the vector from the spotlight to the
point being illuminated. $\theta_{in}$ and
$\theta_{out}$ may be used to control the radius of the cone of light
produced by the spotlight.
\end{defkey}
$\theta_{in}$ is the the angle at which
the light source begins to be attenuated. At $\theta_{out}$,
the spotlight intensity is zero.
This affords control
over how ``fuzzy'' the edges of the spotlight are. If neither angle
is given, they both are effectively set to 180 degrees.

Extended sources are meant to model spherical light sources. Unlike
point sources, extended sources actually possess a radius, and as such
are capable or producing shadows with fuzzy edges ({\em penumbrae}). If
you do not specifically desire penumbrae in your image, use a point
source instead.

\begin{defkey}{light}{{\em Intensity\/} {\tt extended} {\em radius} \evec{pos} }
Create an extended light source at the given position and with
the given {\em radius}.
\end{defkey}
The shadows cast by
extended sources are modeled by taking samples of the source at
different locations on its surface. When the source is partially
hidden from a given point in space, that point is in partial shadow
with respect to the extended source, and the sampling process is
usually able to determine this fact.

Quadrilateral light sources are computationally more expensive than extended
light sources, but are more flexible and produce more realistic results.
This is due to the fact that an area source is approximated by a
number of point sources whose positions are jittered to reduce aliasing.
Because each of these point sources has shading calculations performed
individually, area sources may be placed relatively close to the
objects it illuminates, and a reasonable image will result.

\begin{defkey}{light}{{\em Intensity\/} {\tt area} \evec{p1} \evec{p2} {\em usamp}
\evec{p3} {\em vsamp}}
Create a quadrilateral area light source.
The $u$ axis
is defined by the vector from \evec{p1} to \evec{p2}. Along
this axis a total of {\em usamp} samples will be taken.
The $v$ axis of the light source is defined by the vector
from \evec{p1} to \evec{p3}. Along this axis a total of
{\em vsamp} samples will be taken.
\end{defkey}
The values of {\em usamp} and {\em vsamp} are usually chosen to be
proportional to the lengths of the $u$ and $v$ axes. Choosing a
relatively high number of samples will result in a good approximation
to a ``real'' quadrilateral source. However, because complete
lighting calculations are performed for each sample,
the computational cost is directly proportional to the product
of {\em usamp} and {\em vsamp}.

\section{Shadows}

In order to determine the color of a point on the surface
of any object, it is necessary
to determine if that point is in shadow with respect to each
defined light source. If the point is totally in shadow with respect to
a light source, then the light source makes no contribution to the
point's final color.

This shadowing determination is made by tracing rays from the point
of intersection to each light source. These ``shadow feeler'' rays
can add substantially to the overall rendering time. This is especially
true if extended or area light sources are used. If at any point you
wish to disable shadow determination on a global scale, there is
a command-line option ({\tt -n}) that allows you to do so. It is also
possible
to disable the casting of shadows onto given objects through the use
of the {\tt noshadow} keyword in surface descriptions. In addition,
the {\tt noshadow} keyword may be given following the definition
of a light source, causing the light source to cast no shadows onto
any surface.

Determining if a point is in shadow with respect to a light source
is relatively simple if all the objects in a scene are opaque. In
this case, one simply traces a ray from the point to the light source.
If the ray hits an object before it reaches the light source, then
the point is in shadow.

Shadow determination
becomes more complicated if there are one or more objects with
non-zero transparency between the point and the light source.
Transparent objects may not completely block the light from a source,
but merely attenuate it. In such cases, it is necessary to compute the
amount of attenuation at each intersection and to continue
the shadow ray until it either reaches the light source or until
the light is completely attenuated.

By default, {\rayshade} computes shadow attenuation by assuming
that the index of refraction of the transparent object is the
same as that of the medium through which the ray is traveling.
To disable
partial shadowing due to transparent objects, the {\tt shadowtransp}
keyword should be given somewhere in the input file.

\begin{defkey}{shadowtransp}{}
The intensity of light striking a point is {\em not} affected by
intervening transparent objects.
\end{defkey}
If you enclose an object behind a transparent surface, and you wish
the inner object to be illuminated, you must not use the {\tt shadowtransp}
keyword or the {\tt -o} option.