Suitability of online and offline remote recording for vowel reduction analysis

Increasing demand for remote research techniques has led to several initial investigations of the suitability of remote recording for vowel production research. An initial consensus has emerged that remote recording may provide broadly accurate data for broad-strokes analysis, such as analyzing the relative positions of a speaker’s vowels, although some minor distortions may occur; fine-grained analysis of sociolinguistic variation suffers greatly from remote data collection (Freeman and De Decker, 2021). This study investigates the quality of data obtainable through two remote recording methods (online recording through Gorilla and offline recording via smartphone) for a vocalic phenomenon requiring a medium level of sensitivity of analysis, namely vowel reduction. Ten subjects completed a shadowing task eliciting target words containing one of five English vowels in reduced or unreduced form, taking simultaneous online and offline remote recordings. To determine whether the variability introduced through the remote recording process was sufficient to alter the findings of the study, results were compared to data from a second group of subjects who completed the task in person in a traditional laboratory setting. This comparison to in-person data allows for an assessment of the suitability of online and offline remote recordings for vowel reduction research.