Perceptual and Acoustic Characteristics of Speech Clarity with and Without a Facemask

Abstract Some studies have found that the speech of speakers wearing facemasks has reduced intelligibility. Although it has been found that facemasks attenuated high-frequency energy, no study has examined the effects of masks on spectral characteristics of vowels or voiceless fricative consonants. The present study investigated auditory perceptual rating of speech clarity and acoustic-phonetic measures of vowels and voiceless fricative consonant production in 16 health care workers who produced standardized voice tasks without and with wearing either a standard surgical mask or a KN95 mask. Voice samples were perceptually rated for speech clarity and were acoustically analysed for root-mean-square amplitude (ARMS), spectral moments of two voiceless fricatives /s/ and /ʃ/, and ARMS and amplitude of the first three formants (A1, A2, and A3). Speech produced whilst wearing either a surgical or KN95 mask was significantly less clear than without a mask, with KN95 showing greater impact than surgical masks. In both fricatives, ARMS was lower in the surgical mask and KN95 mask conditions compared to the non-mask condition. None of the amplitude measures of vowels were affected by facemasks. Linear regression models indicated that perception of speech produced by mask users was mainly affected by modification of voiceless fricative consonant characteristics.