Individuality and malleability in the human voice: Articulatory setting and its implications for forensic speaker comparison.

点击翻译按钮获取中文摘要

Voice is fundamental to human identity. It conveys extensive social information about a speaker, such as age, gender identity, and regional background, while also providing a means by which a person can be recognised as an individual. This hypothesised individuality of the human voice is highly significant within a forensic context; it is not uncommon for forensic phoneticians to be called upon in criminal investigations to assess whether a recorded speech sample is similar to that of a suspect (Nolan, 1991). While voice can be a robust indicator of identity, it is also incredibly malleable. Individuals vary their speech across different contexts (Sharma,2011) and some speakers can imitate the speech of others to a high degree of accuracy (Leemann & Kolly,2015). This dynamic between the voice as individual and malleable is problematic for forensic casework, which favours contexts of high between-speaker and low within-speaker variability. It also raises questions about whether there is anything fundamentally unique to an individual's voice. Determining whether voices are unique and, if so, in which dimensions, is essential in establishing the reliability of forensic speaker comparison