ObjectivesProtective facial coverings (PFCs) such as surgical masks attenuate speech transmission and affect speech intelligibility, which is reported in languages such as English and German. The present study intended to verify the detrimental impacts on production of tonal languages such as Cantonese, by examining realization of speech correlates in Cantonese under PFCs including facial masks and shields. MethodsWe recorded scripted speech in Hong Kong Cantonese produced by three adult speakers who wore various PFCs, including surgical masks, KF94 masks, and face shields (with and without surgical masks). Spectral and temporal parameters were measured, including mean intensity, speaking rate, long-term amplitude spectrum, formant frequencies of vowels, and duration and fundamental frequency (F0) of tone-bearing parts. ResultsSignificant changes were observed in all acoustic correlates of Cantonese speech under PFCs. Sound pressure levels were attenuated more intensely at ranges of higher frequencies in speech through face masks, whereas sound transmission was affected more at ranges of lower frequencies in speech under face shields. Vowel spaces derived from formant frequencies shrank under all PFCs, with the vowel /aa/ demonstrating largest changes in the first two formants. All tone-bearing parts were shortened and showed increments of F0 means in speech through PFCs. The decrease of tone duration was statistically significant in High-level and Low-level tones, while the increment of F0 means was significant in High-level tone only. ConclusionsGeneral filtering effect of PFCs is observed in Cantonese speech data, confirming language-universal patterns in acoustic attenuation by PFCs. The various coverings lower overall intensity levels of speech and degrade speech signal in higher frequency regions. Modification patterns specific to Hong Kong Cantonese are also identified. Vowel space area is reduced and found associated with increased speaking rates. Tones are produced with higher F0s under PFCs, which may be attributed to vocal tension caused by tightened vocal tract during speaking through facial coverings.
Read full abstract