Objective: To describe acoustic and communicative characteristics of voices perceived as neutral by Brazilian listeners.
Methods: A descriptive study analyzed 14 voices perceived as neutral by 101 cisgender judges on a visual analog scale (VAS) ranging from -50 (very masculine) to +50 (very feminine). Neutrality was defined as within ±1 standard deviation from 0 (neutral point). Sixteen acoustic measures were extracted using the Prosodic Descriptor Extractor script in Praat software, and two speech therapists assessed the expressiveness of a poem reading using the Speech-Language Pathology Observation of Expressiveness Protocol (RoFOE), which includes 28 parameters across initial impact and expressiveness. Spearman correlations were conducted for acoustic and communicative measures (p>0.05).
Results: Among the neutral voices, 7 (50%) were from transgender women, 5 (35.71%) from transgender men, and 2 (14.29%) from non-binary individuals. The average fundamental frequency (fo) for vowels was 150.53 Hz. In poem readings, the average HNR was 13.37 dB, shimmer was 7.59 dB and jitter 2.17%, and maximum and minimum fo was 185.32 Hz and 107.45 Hz, respectively. The average difference between the first two harmonics was 1.14 dB, and the ABI averaged 2.02. Regarding communicative perception, 11 (78.57%) had natural speech, 6 (42.86%) were perceived as pleasant, and 5 (35.71%) as captivating. Balanced resonance was observed in 5 participants (35.71%), with 6 (42.86%) noted for laryngopharyngeal resonance and 4 (28.57%) for hypernasality. Spearman correlations showed associations between fo standard deviation and a more pleasant (p=0.004; R=0.716) or engaging voice (p<0.001; R=0.832); laryngopharyngeal resonance and mean fo (p=0.048; R=-0.537) and minimum fo (p=0.013; R=-0.645); and hypernasal resonance and fo standard deviation (p=0.042; R=-0.549). Correlations were also found between precise articulation and both fo standard deviation (p=0.029; R=0.583) and minimum fo (p=0.029; R=0.583), as well as maximum fo (p=0.017; R=0.626). Imprecise articulation correlated with mean (p=0.004; R=-0.713), minimum (p=0.004; R=-0.713), and maximum fo (p=0.009; R=-0.669).
Conclusion: Acoustic characteristics of voices perceived as neutral appear to intersect with standardized values for both masculine and feminine voices. Most participants had natural speech quality, though few were rated as pleasant or engaging. Correlations indicate resonance and articulation are associated with fo parameters and that the standard deviation of fo correlates with the perception of more pleasant or engaging communication in these voices.