Effects of speaker's and listener's acoustic environments on speech intelligibility and annoyance

機関リポジトリ

説明

Speech signal modification to increase intelligibility in noisy environments may cause discomfort in quieter environments. Human speech production in noise is known to improve intelligibility in noise. The so-called Lombard effect might cause annoyance in quiet environments, However, the effect disappears in a quiet environments. This suggests that humans might modify their voices depending on the acoustic environment considering intelligibility and discomfort. The speech modification algorithm which use rules obtained from human speech production are expected to meet the demand for both of high intelligibility and reducing discomfort in various acoustic environments. This paper examined the influence of the coincidence between speaker’s and listener’s acoustic environments on speech intelligibility and the level of annoyance and to confirm the expectation. Speech produced by four Japanese speakers in quiet and noisy environments (at non, 66, 78, 90 dB) were presented to Japanese listeners in quiet and noisy environments (at non, 66, 78, 90dB) to measure intelligibility of words and the level of annoyance. The results showed that the speech is less intelligible or more annoying when the listener’s environment is different from the speaker’s one. This supports that mimicking human speech production would meet the demand for high intelligibility and less annoyance.

Proceedings of the INTER-NOISE 2016 : 45th International Congress and Exposition on Noise Control Engineering : Towards a Quieter Future, August 21 - 24, 2016, Hamburg, Germany

Poster - P05 - Psychoacoustics, Product Sound Quality, Soundscape

identifier:https://dspace.jaist.ac.jp/dspace/handle/10119/14762

収録刊行物

  • Inter-noise 2016

    Inter-noise 2016 171-176, 2016-08-21

    German Acoustical Society (DEGA)

キーワード

詳細情報 詳細情報について

  • CRID
    1050292572155338880
  • NII論文ID
    120006676387
  • ISSN
    0105175X
  • Web Site
    http://hdl.handle.net/10119/14762
  • 本文言語コード
    en
  • 資料種別
    conference paper
  • データソース種別
    • IRDB
    • CiNii Articles

問題の指摘

ページトップへ