Can you hear my body size?: Estimating speaker height and body build from speech*

This study aims to assess listeners’ ability to estimate speakers’ body size from speech. Twelve acoustic correlates related to resonance (F1, F2, F3, F4, and formant dispersion), pitch (f0), and breathiness/roughness (H1*-H2*, HNR05, HNR15, HNR25, HNR35, and CPP) were measured on /a/s in sentences spoken by 33 male and 37 female French and Korean speakers. Subsequently, 35 male and 70 female French and Korean listeners evaluated speakers’ body height and build from speech, each on a 6-point scale. Results indicated that French listeners generally outperformed Korean listeners in estimating body height from speech, although overall accuracy was low to modest. Acoustic analyses showed that lower values of f0, formant structure (F1-F4), and harmonic measures (HNR05, CPP) are associated with larger body size; however, the direction and strength of these associations varied across speaker sex as well as speaker and listener language. These patterns suggest that listeners are sensitive to pitch and spectral information linked to apparent vocal fold and vocal tract size, yet such cues provide only coarse and often inconsistent inferences about actual height and build. These findings are interpreted as evidence that speech–body associations are shaped by sex-dependent and language-specific perceptual stereotypes, rather than serving as a reliable vocal indicator of physical dominance.