Music Processing Synthesis Course
Prof. Dr.-Ing. Sebastian J. Schlecht Friedrich-Alexander-Universität Erlangen-Nürnberg
Music processing and synthesis is choosing a representation of sound, controlling it, and trading off built-in structure against what is learned from data. Every lecture ends on four questions: how is the sound represented, what are the controllable parameters, what structure is built in, and what is fixed by design versus learned? See StudOn for organisation and the exam.
Lectures
| # | Deck | Topic |
|---|---|---|
| 1 | A History of Sound Synthesis | From tonewheels to diffusion models |
| 2–6 | Filters and effects | Lecture notes on StudOn |
| 7 | Oscillators and Spectra | Additive, subtractive, FM, and waveshaping |
| 8 | Samples and Grains | Sampling, granular synthesis, and the phase vocoder |
| 9 | Physical Models | From the wave equation to modes, waveguides, and instruments · notebook |
| 10 | Differentiable DSP | Keep the structure, learn the parameters |
| 11 | Artificial Reverberation | The capstone: every method of the course in one room |
| 12 | Diffusion models in audio | Guest lecture |
Interactive notebooks
Browser-side marimo notebooks: they run entirely in your browser, no install needed (the first load takes a few seconds). Each one follows the course pattern: listen, inspect, manipulate, play, match.
- A string, two ways (lecture 9): break the FDTD grid at the Courant limit, then tune a modal string until it plays the course’s guitar riff.