• DocumentCode
    2450001
  • Title

    Mel domain objective speech quality evaluation measure performance

  • Author

    Wenlian Zhan ; Zhulin Shen

  • Author_Institution
    Sch. of Foreign Languages, Hunan Int. Econ. Univ., Changsha, China
  • fYear
    2012
  • fDate
    16-18 July 2012
  • Firstpage
    459
  • Lastpage
    464
  • Abstract
    Effectively the objective evaluation of voice quality, comparative analysis of the MFSC as characteristic parameter of Mel-SD and MFCC for the characteristic parameters of Mel-CD, feature extraction filter structure change of the two measure, and Mel-SD compression factor to be studied. Test studies have shown that, Mel-SD performance is better than Mel-CD also has the robustness of the structural change of the filter banks; Mel-CD is more sensitive to changes in filter structure, with the number of filters in the number of filters for over 13 the increase in performance degradation. Mel-SD in the case given the number of filters, the best compression factor. Within a certain range, the compression factor is not serious. The best compression factor in line with psychoacoustic static measurements of the experimental conclusions of the approximate expression. Parameter Optimization of Mel-CD and Mel-SD for the objective evaluation of the quality of voice communication systems in interference conditions, the results show that the Mel-SD performance is better than Mel-CD and the PESQ performance of the Mel-CD is quite PESQ.
  • Keywords
    feature extraction; filtering theory; interference (signal); performance evaluation; speech coding; speech processing; voice communication; MFSC; Mel-SD compression factor; Mel-SD performance; characteristic parameter; comparative analysis; feature extraction filter structure; interference conditions; mel domain objective speech quality evaluation measure performance; parameter optimization; performance degradation; psychoacoustic static measurements; voice communication systems; voice quality evaluation; Correlation; Filter banks; Frequency domain analysis; Mel frequency cepstral coefficient; Nonlinear distortion; Optimization; Speech;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Audio, Language and Image Processing (ICALIP), 2012 International Conference on
  • Conference_Location
    Shanghai
  • Print_ISBN
    978-1-4673-0173-2
  • Type

    conf

  • DOI
    10.1109/ICALIP.2012.6376661
  • Filename
    6376661