Formant detail needed for identifying, rating, and discriminating vowels in Wisconsin English

Abstract

Neel [(2004). Acoust. Res. Lett. Online 5, 125–131] asked how much time-varying formant detail is needed for vowel identification. In that study, multiple stimuli were synthesized for each vowel: 1-point (monophthongal with midpoint frequencies), 2-point (linear from onset to offset), 3-point, 5-point, and 11-point. Results suggested that a 3-point model was optimal. This conflicted with the dual-target hypothesis of vowel inherent spectral change research, which has found that two targets are sufficient to model vowel identification. The present study replicates and expands upon the work of Neel. Ten English monophthongs were chosen for synthesis. One-, two-, three-, and five-point vowels were created as described above, and another 1-point stimulus was created with onset frequencies rather than midpoint frequencies. Three experiments were administered ( n = 18 for each): vowel identification, goodness rating, and discrimination. The results ultimately align with the dual-target hypothesis, consistent with most vowel inherent spectral change studies.