ARTICULATORY AND ACOUSTIC PHONETICS OF VOICE ACTING

Abstract

Professional voice actors are an excellent, but fairly untapped population for linguistic study. These actors are linguistically naïve but perform complex linguistic tasks throughout the course of their profession, and they perform these tasks safely, reliably, and consistently. While some previous studies used auditory and instrumental acoustic methods to investigate the speech produced by voice actors (Teshigawara, 2003; Teshigawara et. al., 2007; Teshigawara, 2009, cited in Teshigawara, 2011; Starr, 2015), in contrast there has been very little research on their articulation (Teshigawara & Murano, 2004). This work uses a combination of 3D/4D ultrasound, electroglottography (EGG), audio recordings, and webcam footage to observe the articulations of six voice actors (3 professionals and 3 amateurs) and compare those articulations to the acoustic outputs. Each actor read a syllable list and two stories from the Goldman Fristoe Test of Articulation 3rd ed. in both their regular adult speaking voice and a simulated child voice. Some general findings of this study indicate that actors do produce different acoustic outputs for a child voice compared to their adult voice including higher fundamental frequency (F0), different averages in acoustic measurement of vowel formants, and shorter estimates of vocal tract length. The ultrasound data indicate that actors also alter the place of lingual constriction to be more anterior, farther forward in the mouth, when imitating a child voice. EGG data, while limited to only two subjects due to the 2020 pandemic, indicate that actors may also raise their larynx to shorten the vocal tract while speaking in an imitated child voice. This work opens a new angle for linguistic study while also providing an empirical basis upon which new pedagogical methods to teach voice acting can be developed. Looking at voice actors speaking with intentionally contorted tongue positions could help us learn more about over which muscles we have volitional control within the vocal tract. This could help to improve interventions in speech therapy and even test and update acoustic-articulatory models of speech.