Speech recognition with cochlear implant simulations (Version 1)

  • Research type

    Research Study

  • Full title

    Speech recognition with cochlear implant simulations: The effects of background speech and music.

  • IRAS ID

    161815

  • Contact name

    Stephanie Corkill

  • Contact email

    sjc5v07@soton.ac.uk

  • Sponsor organisation

    University Of Southampton

  • Research summary

    TITLE: Speech recognition with cochlear implant simulations: The effects of background speech and music.
    SUMMARY:Cochlear implants (CI’s) enable users to perform at levels comparative to their normal-hearing (NH) peers, when testing word and sentence recognition in quiet environments.
    Listening in the presence of a competing masker (a masker is an interferring sound source which can affect a person's ability to hear the original sound source), such as babble (the continuous sound of several people talking at once, it is not intelligible, only jargon like speech) or music, is
    however a limitation of the CI, with users displaying poorer speech perception scores than their NH peers.
    Currently, the majority of studies involving NH and hearing impaired listeners suggest that speech is a more effective masker than music; this project will test this theory using CI simulations (a cochlear implant simulation is where the sound that will be heard through the headphones will be manipulated to replicate the sound if listening through a CI).
    This project’s objectives are to: investigate the effects of background speech and
    music on speech perception; test the hypothesis that speech is a more effective
    masker than music with CI simulations; and explore whether different types of music have different masking effectiveness.
    Speech recognition in the presence of multi-talker speech babble, and different
    music maskers, will be tested as NH subjects listen through CI simulations. The
    signal-to-noise ratio that produces a 80% correct response rate will be used as the
    outcome measure.
    Findings will add to existing knowledge on speech-in-noise perception by CI users,
    and may have implications for clinical testing of speech recognition in background
    music and the recognition of lyrics.

  • REC name

    HSC REC B

  • REC reference

    14/NI/1080

  • Date of REC Opinion

    13 Oct 2014

  • REC opinion

    Further Information Favourable Opinion