CS891 Special Lecture: Music Information Retrieval (MIR)

Course Info: This course will cover various topics from low-level signal processing techniques for feature extraction from audio to machine learning algorithms for high-level musical metadata extraction in Music Information Retrieval systems. The program will focus on several tasks active in the MIR society such as tempo/beat tracking, melody/harmony extraction, music similarity and retrieval, cover song identification, and so on. We will also touch on structural analysis including music segmentation and summarization. The focus will be made on the raw audio processing, but symbolic data processing will also be discussed. Frequent references will be made to the rich literature found in the Proceedings of ISMIR (International Symposium on Music Information Retrieval), IEEE ICASSP (International Conference on Acoustics, Speech, and Signal Processing), and other related proceedings/journals.

The course will consist of half-day lectures followed by half-day supervised lab sessions, when the participants will be implementing algorithms in MATLAB.

*Prerequisites: familiarity with MATLAB, introductory knowledge of digital signal processing