The current understanding of listener sensitivity to regional accents comes from examination of speech processing in quiet and in noise. This study had two aims: 1) to examine the intelligibility of regional accents in a multitalker environment, and 2) to explore a methodological question of whether systematic regional features can be detected in the productions of only one representative talker or whether several talkers are necessary to provide the suitable sample. Two American English dialects, General American English and Southern American English, were systematically varied both in the target speech and in the masking babble at three sound-to-noise ratios. The results showed that regional accents did influence listeners' performance in a multitalker environment. Intelligibility was hampered when the target and the masker shared common dialect features or when listeners' heard their own dialect in the masking babble. Southern American was a more intelligible variety than General American, which can be attributable to a set of specific acoustic phonetic features. The study found that systematic regional features can be reliably detected in the production of only one representative talker.
Bibliographic reference. Jacewicz, Ewa / Fox, Robert Allen (2013): "Regional accents affect speech intelligibility in a multitalker environment", In INTERSPEECH-2013, 2081-2085.