Speech recognition error patterns for steady-state noise and interrupted speech

Kimberly G. Smith, Daniel Fogerty

Research output: Contribution to journalArticlepeer-review


Listening in various types of adverse listening conditions may lead to different errors in speech recognition. Young adults repeated sentences degraded by steady-state noise or periodically interrupted by noise preserved at varying proportions. Recognition errors were analyzed according to the noise type and speech proportion. Across noise types, as word recognition decreased, the occurrence of phonemic substitutions and whole word omissions increased. Listeners made more whole word omission and substitution errors during steady-state noise. Part word errors occurred most frequently when listening to speech presented in steady-state noise or interrupted by noise with the smallest speech proportion preserved.

Original languageEnglish (US)
Pages (from-to)EL306-EL312
JournalJournal of the Acoustical Society of America
Issue number3
StatePublished - Sep 1 2017
Externally publishedYes

ASJC Scopus subject areas

  • Arts and Humanities (miscellaneous)
  • Acoustics and Ultrasonics


Dive into the research topics of 'Speech recognition error patterns for steady-state noise and interrupted speech'. Together they form a unique fingerprint.

Cite this