TY - GEN
T1 - Genre specific dictionaries for harmonic/percussive source separation
AU - Laroche, Clément
AU - Papadopoulos, Hélène
AU - Kowalski, Matthieu
AU - Richard, Gaël
N1 - Publisher Copyright:
© Clément Laroche, Hélène Papadopoulos, Matthieu Kowalski, Gaël Richard.
PY - 2016/1/1
Y1 - 2016/1/1
N2 - Blind source separation usually obtains limited performance on real and polyphonic music signals. To overcome these limitations, it is common to rely on prior knowledge under the form of side information as in Informed Source Separation or on machine learning paradigms applied on a training database. In the context of source separation based on factorization models such as the Non-negative Matrix Factorization, this supervision can be introduced by learning specific dictionaries. However, due to the large diversity of musical signals it is not easy to build sufficiently compact and precise dictionaries that will well characterize the large array of audio sources. In this paper, we argue that it is relevant to construct genre-specific dictionaries. Indeed, we show on a task of harmonic/percussive source separation that the dictionaries built on genre-specific training subsets yield better performances than cross-genre dictionaries.
AB - Blind source separation usually obtains limited performance on real and polyphonic music signals. To overcome these limitations, it is common to rely on prior knowledge under the form of side information as in Informed Source Separation or on machine learning paradigms applied on a training database. In the context of source separation based on factorization models such as the Non-negative Matrix Factorization, this supervision can be introduced by learning specific dictionaries. However, due to the large diversity of musical signals it is not easy to build sufficiently compact and precise dictionaries that will well characterize the large array of audio sources. In this paper, we argue that it is relevant to construct genre-specific dictionaries. Indeed, we show on a task of harmonic/percussive source separation that the dictionaries built on genre-specific training subsets yield better performances than cross-genre dictionaries.
M3 - Conference contribution
AN - SCOPUS:85030224512
T3 - Proceedings of the 17th International Society for Music Information Retrieval Conference, ISMIR 2016
SP - 407
EP - 413
BT - Proceedings of the 17th International Society for Music Information Retrieval Conference, ISMIR 2016
A2 - Mandel, Michael I.
A2 - Devaney, Johanna
A2 - Turnbull, Douglas
A2 - Tzanetakis, George
PB - International Society for Music Information Retrieval
T2 - 17th International Society for Music Information Retrieval Conference, ISMIR 2016
Y2 - 7 August 2016 through 11 August 2016
ER -