The mismatch between the production and perception of mean frequency in New Zealand English ethnolects was investigated. The perception study was a part of larger research project using innovative techniques to isolate the precise prosodic features that listeners might tune into in ethnic dialect identification in the New Zealand context. 107 listeners performed a forced-choice dialect identification task. Seven speech conditions were created, based on the speech conditions of 20 speakers. Each conditions kept different suprasegmental information in the speech signal. Condition Three was created as a monotonous speech rhythm only condition. Each consonant and pause was replaced by silence, while vowels were replaced by a tone complex created in Praat as a sum of a number of cosine waves with equidistant frequencies at a sampling frequency of 8000Hz. It was observed that the listeners do rely on the mean frequency characteristics to identify speaker ethnicity.
|Number of pages||2|
|Journal||Canadian Acoustics - Acoustique Canadienne|
|Publication status||Published - Sep 2008|