Game Development Reference
want as little data in the file as possible and to let the transformation and
rendering engine look after the creation of the different resolution meshes.
Another problem of storing the data within the file will be the use of a
single mesh deformation engine. The approach we will adopt for the
movement of characters is to have a key resolution that is deformed using
some kind of bones system, as outlined earlier. Having deformed the
mesh we then create a display resolution version of this mesh. If the
character is close to the camera, then we may use subdivision to smooth
the mesh. If the character is further away, then we can use polygon
The options for polygon reduction
When reducing the number of polygons in a mesh there are fundamen-
tally three alternative routes.
An algorithm is adopted to scan a mesh and determine the vertex which,
if removed, would make the least difference to the mesh. This vertex is
removed leaving a hole. The hole is re-triangulated and the iteration
process continues until a target number of polygons is in the mesh or a
limiting error is reached. Most algorithms that take this approach cannot
work on arbitrary meshes, so this approach is not really suited to the multi-
resolution meshes we intend to create.
Figure 16.3 Removing a vertex and re-triangulating the hole.