Needless to say, in the amazing 3D Modeling and animation world, each face and each vertex in a model require computer resources
to store and process. The more complex the model becomes the less RAM is available for rendering.

By reducing the face/vertex count, you can often cut the rendering time enough to equal the speed of adding a new computer
to the network. Just as too much information can overload your system's processor, it can overload yours, so work in manageable chunks.

Imagine, for instance, you are modeling a new school building to present to the school board on videotape. You want to show the school as it relates to the neighborhood; and you want to highlight the attractive front entry and show off the new chemistry lab and gym.

To see the exterior in the neighborhood, you have to be 2000 yards away from and above the building. At that distance, the resolution of the videotape used as output will only show details larger than three feet. It is unrealistic to work with a fully detailed model of the school with all the interior furniture and fixtures in place.

Instead, make one low detail model of the massing of the building, and add windows and doors using mapped materials instead of geometry. Trees and shrubs at this resolution can also be mapped onto flat planes rather than created as complex mesh objects.

Next, have a more detailed model of the front entry viewed from ground level at a distance of 200 yards. Such details as door handles and window and door trim details may still be too small to show in the tape. Keep the shapes true, but leave fine details out and use higher resolution mapped materials to simulate details like brick grout or shadows.

Next is an interior shot of the lab or classroom with the important details included, especially where the camera will get close. If the client decides later to see a classroom, it is usually easier to create the new model than to have it already in the scene "just in case."