Game Development Reference
In-Depth Information
The information described usually consists of vertex positions and surface
normals. Object space is also known as local space and, especially in the
context of graphics, model space.
From model space, the vertices are transformed into world space (see
Section 3.2.1). The transformation from modeling space to world space is
often called the model transform. Typically, lighting for the scene is spec-
ified in world space, although, as we see in
Section 10.11,
it doesn't really
matter what coordinate space is used to perform the lighting calculations
provided that the geometry and the lights can be expressed in the same
space.
From world space, vertices are transformed by the view transform into
camera space (see Section 3.2.3), also known as eye space and view space
(not to be confused with canonical view volume space, discussed later).
Camera space is a 3D coordinate space in which the origin is at the cen-
ter of projection, one is axis parallel to the direction the camera is facing
(perpendicular to the projection plane), one axis is the intersection of the
top and bottom clip planes, and the other axis is the intersection of the left
and right clip planes. If we assume the perspective of the camera, then one
axis will be “horizontal” and one will be “vertical.”
In a left-handed world, the most common convention is to point +z in
the direction that the camera is facing, with +x and +y pointing “right”
and “up” (again, from the perspective from the camera). This is fairly
intuitive, as shown in Figure 10.7. The typical right-handed convention is
to have −z point in the direction that the camera is facing. We assume the
left-handed conventions for the remainder of this chapter
Figure 10.7
Typical camera-space conventions
for left-handed coordinate systems
Search WWH ::
Custom Search