Abstract
OBJECTIVES/HYPOTHESIS
A multifactorial phenomenon, vibrato exists in a variety of musical styles and genre contexts. Current vocal vibrato analysis methods using average metrics are applicable only if the vibrato is uniform, consistent, persistent, and omnipresent; features belonging predominantly to the Western Classical Opera esthetic. Historically, vocal vibrato has been analyzed with tools presuming this lens, disregarding significant stylistic characteristics of many other genres with nonnormative, naturally occurring vibrato features. Therefore, a new system of vibrato parameters considering vibrato regularity, variability, and stability over time in more genres is essential.
METHODS
Fifteen professional Operatic, Musical Theater, and Jazz mezzo-sopranos and sopranos recorded a task list of two cross-genre songs and one vocal exercise. Sixteen pitch segments from each singer were subjected to sinusoidal extraction, fo band-pass filtering, and a fast fourier transform long term average spectrum in Praat. Correlations in mean half-extent (in cents), pitch, vowel, and style/singer subject were analyzed for each sample and assessed using standard deviation, coefficient of variation (CV), linear and polynomial regression, and non-linear regression techniques in R. A subsequent perceptual survey using samples most representative of each genre's average CV was distributed to seven vocal pedagogue judges.
RESULTS
The acoustic analysis results indicated that vibrato variability predictably distinguished performed genres. The CV well-characterized vibrato variability and was higher in the samples of Musical Theater and Jazz singers. A 4-parameter logistic regression model is proposed as a novel application and more accurate representation of such multiphasic vibrato with complex shapes. The perceptual survey results confirmed that genre may be accurately distinguished and classified based on the most representative vibrato variability for each group, though Musical Theater singers' vibrato was more challenging to categorize compared with Opera and Jazz singers' vibrato.
DISCUSSION
The novel application of perceptually correlated vibrato models and time-varying parameters proposed in this two-part study may be employed to examine and evaluate complex vibrato patterns and style-specific performance. In turn, this promotes more genre-inclusive voice training in the vocal studio and contributes ecologically valid normative thresholds for vibrato habilitation and rehabilitation in the voice clinic.
