• DocumentCode
    3143891
  • Title

    Combining a harmonic-based NMF decomposition with transient analysis for instantaneous percussion separation

  • Author

    Janer, Jordi ; Marxer, Ricard ; Arimoto, Keita

  • Author_Institution
    Music Technol. Group, Univ. Pompeu Fabra, Barcelona, Spain
  • fYear
    2012
  • fDate
    25-30 March 2012
  • Firstpage
    281
  • Lastpage
    284
  • Abstract
    Many recent approaches on musical source separation rely on model-based inference methods that take into account the signal´s harmonic structure. To address the particular case of instantaneous percussion separation, we propose a method that combines a harmonic-based decomposition using a Non-negative Matrix Factorization (NMF) algorithm, with the transient analysis of spectral peaks from a single audio frame. The signal model allows the estimation of harmonic and non-harmonic sources. Later, as shown in the evaluation, adding transient peak information improves the Signal-to-Distortion Ratio (SDR). Compared to other existing methods, this approach achieves a comparable performance, being suitable at the same time for low-latency conditions.
  • Keywords
    audio signal processing; harmonic analysis; matrix decomposition; music; source separation; transient analysis; harmonic based NMF decomposition; instantaneous percussion separation; musical source separation; nonnegative matrix factorization algorithm; signal to distortion ratio; single audio frame; spectral peak; transient analysis; Estimation; Harmonic analysis; Instruments; Music; Source separation; Speech; Transient analysis; NMF; harmonic analysis; source separation; transient analysis;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on
  • Conference_Location
    Kyoto
  • ISSN
    1520-6149
  • Print_ISBN
    978-1-4673-0045-2
  • Electronic_ISBN
    1520-6149
  • Type

    conf

  • DOI
    10.1109/ICASSP.2012.6287872
  • Filename
    6287872