Development of a music-information-system based on audio-features and music-theoretical principles

The steadily growing amount of multimedia content (e.g. online-music-collections) calls for the development of entirely new concepts of analysis that allow a systematic and clear representation of the available bulk of data.

One growing branch in the field of music-related signal processing is “Music Information Retrieval” (MIR). MIR is concerned with automatic extraction and analysis of information from music in digital formats.

Within the scope of this project, we will develop a system that is capable of analyzing audio-data based on both signal-theoretical and music-theoretical principles. The extracted information will be directly accessible to users in its pure form as well as in terms of abstract classes (“fingerprints”).

Therefore the core of our work will consists in the implementation and evaluation of current state of the art analysis and classification methods from the field of MIR in MATLAB. The chosen approach will discriminate between pure signal processing in a conventional sense (spectra, MFCCs, …) and an extended analysis of the data based on music-theoretical principles (tempo, key, etc. ).

In addition, we will develop a “Graphical User Interface” (GUI) that provides intuitive handling and task-specific editing of the computed data.