![]() |
There are a number of possible methods: |
Bones
This is the method used by most 3D editors. This models the shape and appearance of the creature only, not the internal muscles, tendons, etc.
The outside 'skin' of the creature is modeled with a single polygonal mesh.
A hierarchical tree is created to define the relationship between the bones (see Kinematics) In this way when a joint is bent, say the shoulder, then all the bones below it will be moved relative to it so that the whole arm moves.
The vertices of the mesh are then related to the bones, so that when a joint is bent, the mesh will deform with it. Most of the polygons move relative to a particular bone. However, close to a joint the polygon will move somewhere between the movements of the adjacent bones, in this way the mesh flexes as the joint is bent.
Layered construction
This method models the internal structure of the body, so it should be more accurate. It has a skeleton, and a skin like the last model, but a middle layer is added to model all the squidgy bits:
- Articulated skeleton, made from rigid shapes connected by hinge or ball joints. These bones may not need to be modeled accurately and might be reduced to line segments.(See Kinematics)
- The rigid skeleton is covered by deformable shapes (grouped volume primitives) to model fat, muscle, tendons etc. These can be attached to bones and deformed to bend joints. This might be modeled by blobs (metaballs) but this may not be good enough for the more complex parts such as, head, hands and feet.
- Skin surface generated from layer 2
Framework
This method builds a framework which is used for both the internal structure and the external triangulation used for rendering, for more information see blobs
Including physics in human animation
Some motion, such as a row of dominoes, can be setup and then run from the physical parameters supplied. However things like a human walking have active processing (a brain) in the loop. Therefore scripting can be used to simulate this processing.
One way to model a person walking might be to use the C-of-M and weight of each part to calculate the position required to keep the body balanced in an upright position. However this is a complex job normally done by the human brain, so some shortcuts will probably be needed. One possibility would be to use canned or pre-calculated movements.









