• DocumentCode
    231567
  • Title

    Comparison of performance between normal and whispered speech in Chinese isolated word recognition

  • Author

    Jun Sha ; Xueqin Chen ; Yibiao Yu

  • Author_Institution
    Sch. of Electron. & Inf. Eng., Soochow Univ., Suzhou, China
  • fYear
    2014
  • fDate
    19-23 Oct. 2014
  • Firstpage
    545
  • Lastpage
    548
  • Abstract
    Whisper plays an important role on special occasions. Due to differences between normal and whispered speech, this paper firstly compares them in Chinese acoustic characters and MFCC feature distance. Besides, the performance in recognition is compared based on an isolated word recognition system. The experiment result shows that the performance of whisper is worse than normal speech whether in the degree of distinction among isolated word features or the recognition rate by using the same feature parameter and recognition model.
  • Keywords
    cepstral analysis; feature extraction; speech recognition; Chinese acoustic character; Chinese isolated word recognition; MFCC feature distance; Mel-frequency cepstral coefficient feature; whispered speech recognition; Databases; Hidden Markov models; Mel frequency cepstral coefficient; Speech; Speech recognition; Training; MFCC; Speech recognition; feature distance; whispered speech;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal Processing (ICSP), 2014 12th International Conference on
  • Conference_Location
    Hangzhou
  • ISSN
    2164-5221
  • Print_ISBN
    978-1-4799-2188-1
  • Type

    conf

  • DOI
    10.1109/ICOSP.2014.7015064
  • Filename
    7015064