We study the topology of several music recommendation networks, which arise from relationships between artist, co-occurrence of songs in play lists or experts' recommendation. The analysis uncovers the emergence of complex network phenomena in these kinds of recommendation networks, built considering artists as nodes and their resemblance as links. We observe structural properties that provide some hints on navigation and possible optimizations on the design of music recommendation systems. Finally, the analysis derived from existing music knowledge sources provides a deeper understanding of the human music similarity perception.

We present the MusicSurfer, a metadata free system for the interaction with massive collections of music. MusicSurfer automatically extracts descriptions related to instrumentation, rhythm and harmony from music audio signals. Together with efficient similarity metrics, the descriptions allow navigation of multimillion track music collections in a flexible and efficient way without the need for metadata nor human ratings.

We report experiments on the use of standard natural language processing (NLP) tools for the analysis of music lyrics. A significant amount of music audio has lyrics. Lyrics encode an important part of the semantics of a song, therefore their analysis complements that of acoustic and cultural metadata and is fundamental for the development of complete music information retrieval systems. Moreover, a textual analysis of a song can generate ground truth data that can be used to validate results from purely acoustic methods. Preliminary results on language identification, structure extraction, categorization and similarity searches suggests that a lot of profit can be gained from the analysis of lyrics.

In this paper, we analyze two social network datasets of contemporary musicians constructed from allmusic.com (AMG), a music and artists' information database: one is the collaboration network in which two musicians are connected if they have performed or produced an album together, and the other is the similarity network in which they are connected if they were musically similar according to the music experts. We find that, while both networks exhibit typical features of social networks such as high transitivity (clustering), we find that they differ significantly in some key network features such as the degree and the betweenness distributions. We believe that this highlights the fundamental differences in the construction mechanism (self-organized collaboration and human-perceived similarity) of the new networks.

No abstract