IMPORTANT NOTE: Starting in the Fall of 2017, Fundamentals of Music Processing has a permanent course number: 21M.387. In the past it was taught under the course 21M.359.
Fundamentals of Music Processing is offered:
- Fall 2017. TR11-12:30. Room 24-033F
- Instructor: Eran Egozy
- TA: Larry Wang
Pre-requisites: 6.003, 21M.051, python programming
Fundamentals of Music Processing deals with music analysis in the audio domain as a signal-processing problem. In the same way that speech processing uses signal analysis to understand spoken words, music processing uses signal analysis on music waveforms to understand higher level musical structure.
We begin by introducing frequency analysis using the Discrete Fourier Transform, and its commonly used relative, the Short-Time Fourier Transform. From there we study features extraction methods such as chroma analysis and onset detection, and examine analysis tools like dynamic-time-warping and self-similarity matrices. These techniques serve to develop a variety of music analysis applications, including:
- Onset classification
- Temporal alignment of different renditions of the same music
- Automatic chord recognition and key detection
- Structural analysis of music
- Tempo and beat tracking
- Pitch detection
- Content-based audio retrieval
- Music-based audio decomposition
Students practice these techniques with in-class lab exercises and coding assignments in python.
The class uses a required text, Fundamentals of Music Processing, by Meinard Müller.