Unmasking the truf: impact of community masks on the perception of voiceless fricatives in English {M Canzi, T Rathcke}

August 8, 2023 - 3 minutes
acoustics perception phonetics

Output

Description

Abstract. The current study aims at quantifying the effects of wearing a face mask on speech perception, by investigating performance of native English listeners in a phoneme monitoring task with monosyllabic words containing voiceless fricatives. Previous experimental work on the topic has mainly focussed on the effects of acoustic filtering caused by the use of face coverings with mixed results and weak effects of mask wearing on speech perception. In this experiment, we explore the interplay of acoustic filtering with other potentially relevant factors such as the presence of visual cues, lexical frequency and listenerspecific background. We provide evidence that suggests the impact of face coverings (esp. FFP-2 face mask) on speech perception is not directly moderated by the acoustic properties of masked speech. Rather, it is inked to an interplay of audiovisual integration, the absence of visual cues for (some) target fricatives, and the listener-specific sociolinguistic background.

Introduction. Since the outbreak of the SARS-CoV-2 pandemic in March 2020, wearing face masks became obligatory in public places around the world. In Germany, masks ought to be worn on public transport only (FFP-2 and surgical masks). In many other countries (e.g. Spain, Italy), wearing of any type of a face mask is recommended in all public places. There exist many anecdotal reports of speech perception being negatively affected by the presence of a face mask cf. [1]. In theory, such negative impact of a face mask might arise for several reasons. Speech perception is multimodal and it relies on a number of cues [2], including visual cues which are lacking in masked speech. The visual modality is particularly important for the perception of labial and dental sounds whose articulation has more readily visible articulatory movements (e.g. [3]). Such visual cues are arguably less crucial for the perception of alveolar or post-alveolar sounds [4]. Moreover, a face mask can be considered a filter to the acoustic properties of speech, due to the air stream passing through the fabric [5] [4].

There has been a long-standing interest in the impact of culturally motivated face coverings on speech acoustics and perception in forensic contexts [4] [6] [7]. For example, the study by [7] analysed recordings of 9 speakers (F = 5) with seven types of face coverings, including surgical masks, motorcycle helmets, ski and cycling masks. The speakers produced words of the CVC structure, with either onsets or codas containing /f/, /T/ or /S/. Presence of a face covering was shown to affect the centre of gravity (COG) in all fricatives, especially (labio)dentals. Spectral skewness and kurtosis were also affected in all fricatives while only dentals additionally showed a large modification of the spectral peak.

…