Signal Processing

Signal processing is a very basic mathematic tool that is crucial for computer graphics. Concepts from signal processing appear in an amazing array of places in graphics. Unfortunately, despite how fundamental image processing is, most graphics books only introduce signal processing in light of some of the more advanced 3D rendering topics. In Foley and van Dam, the discussion of signal processing is hidden away in Chapter 14 under the topic of realistic image generation.

I have written a tutorial on signal processing that was originally intended for people working on an animation problem called "motion editing." The discussion is far more general, although the examples all come from motion editing. With motion editing, we are mainly concerned with how an animated character moves over time. For example, you could imagine specifying the angle of the elbow (a "joint angle").

For graphics, we are generally interested in the 2D case (whereas basic signal processing is concerned with the 1D case). However, the intuitions are much easier to get in the 1D case, so I am asking you to learn them first by reading my chapter The Intutions of Signal Processing (for Motion Editing).

Pat Hanrahan, a graphics professor at Stanford (and Wisconsin alumni) has written a similar set of notes that are a little more mathematical, but deal more directly with images. I recommend these Notes on Signal Processing after you make it through mine.