• DocumentCode
    1697364
  • Title

    Effect of individually tailored spectral change enhancement on speech intelligibility and quality for hearing-impaired listeners

  • Author

    Jing Chen ; Moore, Brian C. J.

  • Author_Institution
    Dept. of Machine Intell., Peking Univ., Beijing, China
  • fYear
    2013
  • Firstpage
    8643
  • Lastpage
    8647
  • Abstract
    Most information in speech is carried by spectral changes over time. We determined if enhancing such changes improves the intelligibility of speech in background sounds for hearing-impaired listeners. The values of four parameters controlling the processing were selected for each subject based on a genetic algorithm. The amount of spectral-change enhancement at a given center frequency increased with increasing hearing loss at that center frequency. A control condition was included with no spectral-change enhancement. The intelligibility of speech was measured for sentences in a speech-shaped noise and in multi-talker babble, using two signal-to-masker ratios, specifically chosen for each subject to avoid floor and ceiling effects. The processing led to small but statistically significant improvements in speech intelligibility for both backgrounds and both signal-to-masker ratios.
  • Keywords
    genetic algorithms; hearing; spectral analysis; speech enhancement; speech intelligibility; genetic algorithm; hearing loss; hearing-impaired listeners; individually tailored spectral change enhancement; multitalker babble; signal to masker ratio; speech intelligibility; speech quality; speech shaped noise; Auditory system; Frequency control; Genetic algorithms; Noise; Speech; Speech enhancement; Spectral change enhancement; hearing loss; speech intelligibility;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on
  • Conference_Location
    Vancouver, BC
  • ISSN
    1520-6149
  • Type

    conf

  • DOI
    10.1109/ICASSP.2013.6639353
  • Filename
    6639353