The SOM-enhanced JukeBox: Organization and Visualization of Music Collections based on Perceptual Models

The availability of large music repositories calls for new ways of automatically organizing <br> and accessing them. While artist-based listings or title indexes may help in locating a specific <br> piece of music, a more intuitive, genre-based organization is required to allow users to browse <br> an archive and explore its contents. So far, however, these organizations following musical <br> styles have to be designed manually.<br> <br> With the SOM-enhanced JukeBox (SOMeJB) we propose an approach to automatically create an <br> organization of music archives following their perceived sound similarity. More specifically, <br> characteristics of frequency spectra are extracted and transformed according to psychoacoustic <br> models. The resulting psychoacoustic Rhythm Patterns are further organized using the <br> Growing Hierarchical Self-Organizing Map, an unsupervised neural network. On top of this <br> advanced visualizations including Islands of Music (IoM) and Weather Charts offer an interface <br> for interactive exploration of large music repositories.<br> <br> Keywords: Music Clustering, Rhythm Patterns, Self-Organizing Map (SOM), Interfaces