On Prosodic Values by Sex and Their Prediction in a Sociolinguistic Sample of the PRESEEA-Valencia Corpus

Abstract

This study analyzes how the sociolinguistic variable of sex influences the prosody of speakers in the PRESEEA-Valencia corpus, particularly in acoustic variables such as pitch range, intensity, and speech rate. Using a quantitative approach and general linear models, the study found that men generally exhibit lower mean pitch and a greater intensity range, while women display higher mean pitch and a wider pitch range. These findings can be explained by both anatomical differences and sociocultural factors. The automatic classification achieved positive identification, although lower effectiveness was observed in identifying female voices, suggesting the need to explore subgroups and additional factors.