Operational Models of Music Similarity for MIR

The rapidly growing amount of music available in digital form via internet or digital libraries calls for entirely new computer-based methods for analysing, describing, distributing, and presenting music. The currently emerging research and application field known as Music Information Retrieval (MIR) is a direct response to that need. Over the past years, our research group has accumulated substantial expertise in intelligent music processing. The goal of this project is to develop our know-how and methods further along three specific lines, to the point where they can be used as a basis for commercially relevant application projects. In particular, the research goals are - to develop computational models and metrics of music (audio) similarity that permit the computer to effectively `understand` which pieces of music may be considered similar by human listeners; - to develop new ways of using such similarity metrics to automatically structure large digital music collections according to musical criteria, into richly structured `music spaces`; - and to develop new methods for visualising such music spaces and permitting users to explore and browse through music collections structured in this way. The result of this will be a set of methods that can be used as a basis for computer systems that provide a rich variety of intelligent music services, such as content-based music collection organisation, search and retrieval of music files, automatic playlist generation, music recommendation -- services for which there is and will be a great demand in the currently developing era of digital music.