How many voices did you hear? Natural variability disrupts identity perception from unfamiliar voices

Abstract

Our voices sound different depending on the context (laughing vs. talking to a child vs. giving a speech), making within-person variability an inherent feature of human voices. When perceiving speaker identities, listeners therefore need to not only 'tell people apart' (perceiving exemplars from two different speakers as separate identities) but also 'tell people together' (perceiving different exemplars from the same speaker as a single identity). In the current study, we investigated how such natural within-person variability affects voice identity perception. Using voices from a popular TV show, listeners, who were either familiar or unfamiliar with this show, sorted naturally varying voice clips from two speakers into clusters to represent perceived identities. Across three independent participant samples, unfamiliar listeners perceived more identities than familiar listeners and frequently mistook exemplars from the same speaker to be different identities. These findings point towards a selective failure in 'telling people together'. Our study highlights within-person variability as a key feature of voices that has striking effects on (unfamiliar) voice identity perception. Our findings not only open up a new line of enquiry in the field of voice perception but also call for a re-evaluation of theoretical models to account for natural variability during identity perception.

Acoustic voice variation within and between speakers

“…(Future studies using spontaneous speech will test this possibility directly.) Finally, LDA and PCA differ in the nature of the questions they ask: LDA provides insight into the variables that maximally separate stimuli, while PCA can reveal the structure of the acoustic space in which the stimuli vary, somewhat analogous to "telling voices apart" versus "telling voices together" (Lavan et al, 2018). These different emphases may partially explain differences in the importance of F0 across experiments.…”

“…include two different kinds of variability: differences within talkers from their own prototype, and deviations of representations for individual speakers from a group prototype. Listeners who are unfamiliar with the voices should be adept at assessing the second kind of variability ("telling voices apart;" Lavan et al 2018), given that the same acoustic features appear to characterize both group and individual prototypes. However, listeners who are unfamiliar with a talker's voice should have difficulty in associating different tokens of a single talker's voice with each other ("telling voices together;" Lavan et al 2018), given their unfamiliarity with the specific idiosyncrasies that characterize that talker's overall acoustic variability.…”

“…Within-talker variability can also interfere with a listener's ability to judge that samples come from the same (rather than different) talkers. In a "telling voices together" task, listeners frequently judged that exemplars from a single talker came from multiple speakers when samples were drawn from different speaking situations with varied interlocutors (Lavan et al, 2018).…”

“…Although it is commonly assumed that prototypes are statistical averages derived from multiple samples of a given talker's voice (e.g., Latinus and Belin 2011a;Maguinness et al 2018), to our knowledge no data exist about how much detail (and what kind of detail) about quality is actually needed to specify the prototype, and how much is reserved as "deviations" from the prototype. Second, the nature (or even the existence) of similar reference patterns for individual talkers and the way in which within-talker variation affects formation of these patterns have not to our knowledge been addressed, although such patterns would seem to be essential for the formation of stable representations of voices and thus for voice recognition (Lavan et al, 2018).…”

A model of person perception from voices

“…", or continuously to "How much does this recording sound like Tom?" (see Lavan, Burston & Garrido, 2018;see Jenkins, White Van Montfort &Burton, 2011 andRitchie, Kramer &Burton, 2018 for faces).…”