The complication of clipping is computing the new vertices.
This process is greatly simplified by using a canonical clipping volume.
We mentioned last lecture that it is often desireable to introduce an intermediate coordinate frame
in-between eyespace and the final projection stage (i.e. the dividing through by w). In this space the viewable
region is mapped into a volume that ranges from -1 to +1 in all dimensions after projection.
This space has several advantages. It simplifies the clipping test (all dimensions are compared against the w
component of the vertex) and it is the perfect place in the pipeline to transistion from a floating-point to
a fixed-point representation.
It also simplifes the interpolation of the new vertex positions and triangle parameters as shown in the figure.
|
|