We present a study on the importance of psycho-acoustic transformations for effective audio feature calculation. From the results, both crucial and problematic parts of the algorithm for Rhythm Patterns feature extraction are identified. We furthermore introduce two new feature representations in this context: Statistical Spectrum Descriptors and Rhythm Histogram features. Evaluation on both the individual and combined feature sets is accomplished through a music genre classification task, involving 3 reference audio collections. Results are compared to published measures on the same data sets. Experiments confirmed that in all settings the inclusion of psycho-acoustic transformations provides significant improvement of classification accuracy.

With increasing amounts of audio being stored and distributed electronically, intuitive and efficient access to large music collections is becoming crucial. To this end we are developing algorithms for audio feature extraction, allowing to compute acoustic similarity between pieces of music, as well as tools utilizing this information to support retrieval of as well as navigation in music repositories. This paper provides an overview of the Rhythm Patterns feature set and demonstrates its suitability for music genre recognition. Furthermore, it outlines the principles of organizing digital music repositories using Self-Organizing Maps and presents the novel PlaySOM interface and the PocketSOMPlayer for mobile devices, both providing intuitively explorable music information spaces.

We present a general framework for performing featurebased synthesis – that is, for producing audio characterized by arbitrarily specified sets of perceptually motivated, quantifiable acoustic features of the sort used in many music information retrieval systems. 1.

Abstract not found