On the temporal aspect of post-consonantal fundamental frequency perturbation in English voicing perception

The fundamental frequency (F0) after English voiceless obstruents is generally higher than their voiced counterparts. This difference is usually largest at the vowel onset and diminishes before the vowel ends. Past research has established that the F0 perturbation influences English listeners' voicing perception. While many studies have highlighted the effect of F0 onset height, the temporal extent of the F0 perturbation is less investigated. In the present study, listeners categorized bilabial stops as voiced or voiceless. We manipulated the voice onset time (VOT), the F0 contour, and the presence of background noise. Except for one level contour at the baseline F0 value, all other F0 contours started at 20 Hz higher than the baseline. Critically, these contours fell to reach baseline at different times, rendering them different slopes. The experimental results revealed an interaction between VOT, F0, and noise. Specifically, when the stimuli were presented in noise and had VOTs near the category boundary, listeners were more likely to categorize F0 falling at slower rates as voiceless than those at faster rates. These results suggested that the duration of F0 perturbation also affects voicing perception in English and might play a role in developing tonal languages.