AUTOMATIC TIMBRE CLASSIFICATION OF ETHNOMUSICOLOGICAL AUDIO RECORDINGS
Dominique Fourer, Jean-Luc Rouas, Pierre Hanna, Matthias Robine · 2014
Automatic timbre characterization of audio signals can help to measure similarities between sounds and is of in-terest for automatic or semi-automatic databases indexing. The most effective methods use machine learning approa-ches which require qualitative and diversified training data-bases to obtain accurate results. In this paper, we intro-duce a diversified database composed of worldwide non-western instruments audio recordings on which is evalu-ated an effective timbre classification method. A compar-ative evaluation based on the well studied Iowa musical instruments database shows results comparable with those of state-of-the-art methods. Thus, the proposed method offers a practical solution for automatic ethnomusicologi-cal indexing of a database composed of diversified sounds with various quality. The relevance of audio features for the timbre characterization is also discussed in the context of non-western instruments analysis. 1.