Effects of forensically-relevant facial concealment on acoustic and perceptual properties of consonants

This thesis offers a thorough investigation into the effects of forensically-relevant facial concealment on speech acoustics and perception. Specifically, it explores the extent to which selected acoustic-phonetic and auditory-perceptual properties of consonants are affected when the talker is wearing ‘facewear’ while speaking. In this context, the term ‘facewear’ refers to the various types of face-concealing garments and headgear that are worn by people in common daily communication situations; for work and leisure, or as an expression of religious, social and cultural affiliation (e.g. surgical masks, motorcycle helmets, ski and cycling masks, or full-face veils such as the niqāb). It also denotes the face or head coverings that are typically used as deliberate (visual) disguises during the commission of crimes and in situations of public disorder (e.g. balaclavas, hooded sweatshirts, or scarves). The present research centres on the question: does facewear influence the way that consonants are produced, transmitted, and perceived? To examine the effects of facewear on the acoustic speech signal, various intensity, spectral, and temporal properties of spoken English consonants were measured. It was found that facewear can considerably alter the acoustic-phonetic characteristics of consonants. This was likely to be the result of both deliberate and involuntary changes to the talker’s speech productions, and of sound energy absorption by the facewear material. The perceptual consequences of the acoustic modifications to speech were assessed by way of a consonant identification study and a talker discrimination study. The results of these studies showed that auditory-only and auditory-visual consonant intelligibility, as well as the discrimination of unfamiliar talkers, may be greatly compromised when the observer’s judgements are based on ‘facewear speech’. The findings reported in this thesis contribute to our understanding of how auditory and visual information interact during natural speech processing. Furthermore, the results have important practical implications for legal cases in which speech produced through facewear is of pivotal importance. Forensic speech scientists are therefore advised to take the possible effects of facewear on speech into account when interpreting the outcome of their acoustic and auditory analyses of evidential speech recordings, and when evaluating the reliability of earwitness testimony.