Low level processing in foreign language speech perception in quiet and noise

Abstract

It has been recognized that speech recognition by L2 learners may suffer disproportionally in adverse listening conditions compared to native listeners. The factors that are likely to be involved in this performance disparity are: (i) interference from the native language; (ii) incomplete acquisition (lack of phonetic detail) of L2 categories; and (iii) presumed universal confusions due to inherent maskability of certain sounds. It is important to tease apart these factors in order to gain insights into processesa nd representationuss edb y native and non-nativel isteners.T he currentw ork investigatesa phoneticc orelate which has a phonologicalr ole (vowel durationa s a cue to word-final consonant voicing in English), its realization in different speech styles and its perception by native and non-native listeners. We present production studies that report on the changes occurring in intelligibility enhancing speech modes, mainly Lombard and foreigner-directed speech, and aim to determine how noise-induced speecha nd speechd irectedt o non-native speakersa ffect duration-basedp honological contrasts, namely vowel shortening. We also present the perception study that investigatesth e way native (English) and non-native( Polish and Spanish)l istenersu se vowel duration as a cue to word-final consonant voicing and the benefit of intelligibility enhancing speech modes in quiet and noise conditions.