2

Motion capture is a special case of the kinematic approach in which the joint angles and/or velocity

data are measured from a real motion and then re-used on an animated character.

The most

common way of capturing a motion at present is to attach a series of markers to various points on

the subject's body and to use multiple video cameras or other sensory devices to record the motion

of the markers.

The subject's motions are mapped directly onto the animated character, thereby

ensuring that the animated motion will be realistic.

The ability to modify, blend and transition

between pre-recorded motions is important to provide the animator with sufficient control over the

final motion. However, results based on modifications of captured motions are not guaranteed to

remain realistic.

1. 2

Dynamics

An alternate approach toward providing realism is the use of physically-based animation.

In this

scheme, motions are the result of physical simulations, which include detailed modeling

of

internal and external forces and torques, the creature's mass and moments of inertia, and its

interaction with the environment. All these parameters affect the final result, as they would in the

real world.

The essence of this approach is to ensure realism by constraining the motion of the

system to abide by the laws of physics. Dynamics-based animation has the advantage that the task

of ensuring that motion is physically realistic has been automated.

The animator is, in principle,

free to apply his or her abilities to the more artistic aspects of the animation process.

Note that

"realism" in this context refers to behaviour consistent with a simulated model of the real world.

Similarity to the real world depends completely on the fidelity of this model.

This approach introduces new and challenging problems to be solved in order to be of practical

use. First, incorporating dynamics effects involves the integration of the equations of motion over

time, and has historically been computationally expensive for all but the most simplistic problems.

While

it

seems

that

no

amount

of

computing

power

is

truly

enough,

efficient

simulation

algorithms and faster hardware are beginning to bring the simulation of complex systems of