Comparing identification of standardized and regionally valid vowels

In perception studies, it is common to use vowel stimuli from standardized recordings or synthetic stimuli created using values from well-known published research. Although the use of standardized stimuli is convenient, unconsidered dialect and regional accent differences may introduce confounding effects. The goal of this study was to examine the effect of regional accent variation on vowel identification.

The authors analyzed formant values of 8 monophthong vowels produced by 12 talkers from the region where the research took place and compared them with standardized vowels. Fifteen listeners with normal hearing identified synthesized vowels presented in varying levels of noise and at varying spectral distances from the local-dialect values.

Acoustically, local vowels differed from standardized vowels, and distance varied across vowels. Perceptually, there was a robust effect of accent similarity such that identification was reduced for vowels at greater distances from local values.

Researchers and clinicians should take care in choosing stimuli for perception experiments. It is recommended that regionally validated vowels be used instead of relying on standardized vowels in vowel perception tasks.