Synthesis

Creating an audio signal from control inputs

·   Issues:

- fidelity / richness
- flexibility / control `knobs'
- cost in complexity (CPU) & data size (store)

·   Mimic real signal, or just make a new one?

- abstract level of correspondence

·   Techniques:

signal models
- sampling
- sinusoid (plus...) models
- nonlinear algorithms e.g. FM
source models:
- source-filter & LPC
- waveguide & other physical models