Capturing vocal fry: Understanding the acoustic and perceptual differences in elicited and naturally produced vocal fry

Abstract

Vocal fry, used frequently by both men and women, is often associated with negative stereotypes of young women. This study examines how different ways of eliciting vocal fry affects acoustic properties as well as listener perceptions of speech. Stimuli in previous studies have been obtained from actors who were presented with obvious examples of fry from media sources and were then instructed to produce speech with vocal fry. This method of eliciting fry may introduce other vocal characteristics that activate stereotypes. We obtained natural stimuli by asking speakers to read passages with no instructions about vocal fry and created a corpus of acoustic and prosodic data about naturalistic fry. We then recorded actors who listened to naturally-produced examples of vocal fry. We compared the acoustic characteristics of low and high fry speech from our naturally-produced utterances, our actors, and stimuli collected by Anderson et al. (2014). Elicited fry (from both studies) differed when compared to naturally-produced fry. Additionally, elicited fry in our study differed from that of Anderson et al on jitter and other measures. The three sets of stimuli were used to replicate and extend Anderson and colleagues’ study on listener perceptions of men and women using vocal fry.