• DocumentCode
    3368101
  • Title

    Automatic Vocal Segments Detection in Popular Music

  • Author

    Liming Song ; Ming Li ; Yonghong Yan

  • Author_Institution
    Key Lab. of Speech Acoust. & Content Understanding, Inst. of Acoust., Beijing, China
  • fYear
    2013
  • fDate
    14-15 Dec. 2013
  • Firstpage
    349
  • Lastpage
    352
  • Abstract
    We propose a technique for the automatic vocal segments detection in an acoustical polyphonic music signal. We use a combination of several characteristics specific to singing voice as the feature and employ a Gaussian Mixture Model (GMM) classifier for vocal and non-vocal classification. We have employed a pre-processing of spectral whitening and archived a performance of 81.3% over the RWC popular music dataset.
  • Keywords
    Gaussian processes; acoustic signal detection; mixture models; music; signal classification; speech recognition; GMM classifier; Gaussian mixture model classifier; RWC popular music dataset; acoustical polyphonic music signal; automatic vocal segment detection; nonvocal classification; singing voice; spectral whitening preprocessing; vocal classification; Feature extraction; Harmonic analysis; Mel frequency cepstral coefficient; Music; Speech; GMM; LFPC; MFCC; Singing Voice Detection; Spectral Whitening; Subband Energy Variance; Vocal Segments Detection;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Computational Intelligence and Security (CIS), 2013 9th International Conference on
  • Conference_Location
    Leshan
  • Print_ISBN
    978-1-4799-2548-3
  • Type

    conf

  • DOI
    10.1109/CIS.2013.80
  • Filename
    6746416