If you examine a reel of traditional celluloid movie film you will immediately notice that it consists of a series of individual pictures. If you examine any one picture all by itself, it reveals nothing about the sequence of pictures from which it came. The mechanism of the movie camera simply captures a series of images, which the projector replays. When we view the movie we perceive a continuous flow of motion because our eyes cannot adapt to the speed at which the images change, so they 'run together'.
As Eadweard Muybridge demonstrated, with a sufficiently short exposure time, film becomes a technology by which real-world motion can be captured as a series of still images. Animation reverses the process; the illusion of motion is created artificially through the rapid presentation of a series of still images. It is necessary to alter each frame in such a way that when presented rapidly one after another, the viewer perceives continuous motion. Critical to this effect is the continuity of the motion: there would be no value to displaying a random collection of images at 30 frames per second, since we would be unable to extract any single image from the flow, and the sequence has no meaning in itself. (Obviously, certain cinematic effects do just this, but usually there is some common element to the geometry of each frame, which at least suggests meaning).
The Problem: Choreography
Change, especially motion, is central to animation. Three-dimensional motion over time becomes a four-dimensional problem. Organizing and orchestrating that change in an efficient and compact form is one of the central problems of animation (and film making in general).
On top of this, motion breaks down into two distinct types: camera animation (using terms from movie-making like dolly, pan, zoom) and model animation, in which parts of the geometry move over time such as when we study shadows cast by the sun as it moves through the sky, or when a door opens, a car drives through the scene, etc.) Both kinnds of motion offer rich opportunity for computer graphics, and present some particular challenges.
Besides the fundamental conceptual problems of animation (i.e., describing camera and model motions, especially complex motions such as those of human beings, with their joint linkages) there are substantial user-interface issues in the software. Consider actions such as editing the motions, resequencing images, or changing the time a particular sequence takes. Further, animation represents many events occuring over time. The animator is intimately aware of the sequence and has little need to experience it each time a modest change is made to the animation. How can the sequence be compressed or abstracted in such a fashion that we work on the compressed form, with consistent control of the finished form?
Finally, animation places substantial demands on our computing resources. An animation involves the rapid retrieval or generation and display of a large volume of information in a very short time. It is important to appreciate where the bottle-necks exist in this information processing.
Last updated: April, 2014