Is Zoom viable for sociophonetic research? A comparison of in-person and online recordings for sibilant analysis

This study is part of a larger project investigating whether Zoom is a viable data collection method for sociophonetic research, examining whether Zoom recordings yield different acoustic measurements than in-person recordings for the exact same speech for 18 speakers. In this article we analyze five spectral measures of sibilants (peak, center of gravity, standard deviation, skewness, and kurtosis) which have been shown to be conditioned by dimensions of identity like speaker gender and sexual orientation in much previous sociolinguistic research. We find that, overall, Zoom recordings yield significantly lower peak, center of gravity, and standard deviation measurements and significantly higher skewness and kurtosis values than in-person recordings for the same speech, likely due to a lower sampling rate on Zoom recordings. However, a preliminary analysis controlling for sampling rate across recording methods reveals the opposite patterns for nearly all measures, suggesting that Zoom stretches the spectral space when compared with the in-person recorder. Because the values of these measurements can lead analysts to draw social interpretations relating to a speaker's performance of gender and sexual identity, we caution against comparing across Zoom and in-person recordings, as differences in measurements may result from the recording method used to collect the data.

Is Zoom viable for sociophonetic research? A comparison of in-person and online recordings for vocalic analysis

In this study, we explore whether Zoom is a viable method for collecting data for sociophonetic research, focusing on vocalic analysis. We investigate whether recordings collected through Zoom yield different acoustic measurements than recordings collected through in-person recording equipment, for the exact same speech. We analyze vowel formant data from 18 speakers who recorded Zoom conversations at the same time as they recorded themselves with portable recording equipment. We find that, overall, Zoom recordings yield lower raw F1 values and higher F2 values than recording equipment. We also tested whether normalization affects discrepancies between recording methods and found that while discrepancies still appear after normalizing with the Watt and Fabricius modified method, Lobanov normalization largely minimizes discrepancies between recording methods. Discrepancies are also mitigated with a Zoom recording setup that involves the speaker wearing headphones and recording with an external microphone.

Comparing aerosol concentrations and particle size distributions generated by singing, speaking and breathing

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic has resulted in an unprecedented shutdown in social and economic activity, with the cultural sector particularly severely affected. Restrictions on musical performances have arisen from a perception that there is a significantly higher risk of aerosol production from singing than speaking, based upon high-profile examples of clusters of COVID-19 following choral rehearsals. However, comparing aerosol generation from different types of vocalization, including singing, across a range of volumes is a rapidly evolving area of research. Here, we measured aerosols from singing, speaking and breathing from a large cohort of 25 professional singers in a range of musical genres in a zero-background environment, allowing unequivocal attribution of aerosol production to specific vocalizations. We do not assess the relative volumes at which people speak and sing. However, both showed steep increases in mass concentration with increase in loudness (spanning a factor of 20–30 across the dynamic range measured, p < 0.001). At the quietest volume (50 to 60 dBA), neither singing (p = 0.19) nor speaking (p = 0.20) were significantly different to breathing. At the loudest volume (90 to 100 dBA), a statistically significant difference (p < 0.001) was observed between singing and speaking, but with singing only generating a factor of between 1.5 and 3.4 more aerosol mass. Guidelines for musical performances should be based on the loudness and duration of the vocalization, the number of participants and the environment in which the activity occurs, rather than the type of vocalization. Mitigations such as the use of amplification and increased attention to ventilation should be employed where practicable. Copyright © 2021 American Association for Aerosol Research. © 2021 American Association for Aerosol Research.