ABSTRACT
INTRODUCTION: The use of vocal fry is common in running speech and has potential psychosocial and vocal health consequences. Determining the different patterns of vocal fry is relevant to differentiating phonatory function, understanding cultural and linguistic use of vocal fry, and clinical diagnostics and intervention. The purpose of this project was to study and categorize patterns of vocal fry in running speech. METHODOLOGY: Analyses were performed on recordings of the Rainbow Passage read out loud by five healthy females 18-21 years old. Praat was used to create audio files with text grids. The audio recordings were examined visually (the audio signal and its spectrogram) and perceptually in order to determine different patterns of vocal fry seen in the audio signal. Criteria for detecting vocal fry were (1) the presence of an acoustic transient (a relatively large and fast dip in acoustic pressure, presumably near glottal closure) with a relatively long cycle period compared to normal phonation periods, or (2) a frequency at or lower than approximately an octave below the nearby normal speaking fundamental frequency. RESULTS: There were 174 total vocal fry samples obtained from the recordings. Six vocal fry patterns were observed. The patterns were: single pulse fry (FRY1), double pulse fry (FRY2), multiple pulse fry (FRY3), period doubling fry (FRY4), inaudible fry (FRY5), and indeterminate vocal fry (FRY6). Single pulse fry was divided into a single (one) pulse fry cycle (FRY1a), a series of even single pulse fry cycles (FRY1b), and a series of uneven single pulse fry cycles (FRY1c). Double pulse fry was divided into a primary then another primary pulse fry cycle (FRY2a), a secondary then a primary pulse fry cycle (FRY2b), and a primary then a secondary pulse fry cycle (FRY2c). Multiple pulse fry, where a higher frequency was modulated by a lower frequency, was divided into the higher frequency being near the speaking fundamental frequency (FRY3a) and the higher frequency being inconsistent or well below the expected speaking fundamental frequency (FRY3b). The category single pulse fry had the most samples, with 76% of the total occurrences, followed by period doubling 13%, and the rest 11%. Relative to where the fry patterns occurred within syllables, 36% occurred at the onset of the syllable, 26% early in the syllable, 25% later in the syllable, and 13% at the end of the syllable. These tallies did not include the sixth category, indeterminate vocal fry (FRY6), which was not included in the study proper but recognized as indicating complicated patterns that did not fall within the first five categories. CONCLUSIONS: Vocal fry is a complex, multifaceted phenomenon. The results of this study suggest that there are identifiable patterns of vocal fry. These patterns need to be differentiated especially regarding the glottal adductory nature and phonatory function of each pattern, glottal closure appearing to be the primary physiological causative factor of the salient negative pressure dips (the adduction behavior will be reported in a study in progress). Further research is necessary to determine other potential categories of vocal fry, determine if there are individual idiosyncratic patterns of vocal fry, determine possible differences in vocal fry produced by individuals of different ages and gender expression and other factors, and research the physiologic, acoustic, aerodynamic, and perceptual reality of each pattern.