Computer Music Modeling and Retrieval

Volume 2771 of the series Lecture Notes in Computer Science pp 42-53

Evolving Automatically High-Level Music Descriptors from Acoustic Signals

  • François PachetAffiliated withSony CSL Paris
  • , Aymeric ZilsAffiliated withSony CSL Paris

* Final gross prices may vary according to local VAT.

Get Access


High-Level music descriptors are key ingredients for music information retrieval systems. Although there is a long tradition in extracting information from acoustic signals, the field of music information extraction is largely heuristic in nature. We present here a heuristic-based generic approach for extracting automatically high-level music descriptors from acoustic signals. This approach is based on Genetic Programming, that is used to build extraction functions as compositions of basic mathematical and signal processing operators. The search is guided by specialized heuristics that embody knowledge about the signal processing functions built by the system. Signal processing patterns are used in order to control the general function extraction methods. Rewriting rules are introduced to simplify overly complex expressions. In addition, a caching system further reduces the computing cost of each cycle. In this paper, we describe the overall system and compare its results against traditional approaches in musical feature extraction à la Mpeg7.