M. A. Bartsch, G. H. Wakefield, M. M. Gross, M. Turnblom, and G. I. Shirley, "Singer, hear thyself: Matching of acoustic targets in a singer’s own voice," presented at 32th Annual Voice Foundation Symposium, Philadelphia, PA, 2003.

Abstract: One of the primary goals of vocal training is the development of an internal representation of vocal production. Commonly, this representation is reinforced through a teacher’s feedback. When a student has done well, the teacher says “Precisely so… now do it again.” The student may then repeat the performance by recreating the dynamic variations in posture, breath support, and vocal resonance that were in place while producing the original target. Alternatively, the student may recreate the same acoustic target using different physiological patterns. In either case, the teacher provides a perceptual judgment of success and provides further feedback. Ultimately, the student must return to the exercise and attempt “do it again” outside of the studio. Here, “success” depends entirely on the accuracy and appropriateness of the student’s internal representation. One might expect that an auditory cue from a recording of the lesson would serve to “shore up” this representation and, perhaps, improve the student’s ability to recreate the original performance.

The present paper asks two related questions. First, how does a singer recreate a particular performance, both immediately and after some delay? Second, what are the consequences of providing auditory cues from the student’s own voice when attempting to “do it again”? To address these questions, we compare measurements obtained of vocal exercises from three male singers of varying levels of training. Each singer was asked to sing a set of vocal exercises several times each. For each exercise, one of the instances was selected as a “target” and the singers were instructed to reproduce that target several times both with and without auditory feedback, either immediately or a day later. Objective measurements included center-of-pressure using a force-plate, thoracic and abdominal movement using a 3-D video imaging system, and acoustic features extracted from time-frequency surfaces [Mellody, Wakefield, and Herseth, Journal of Voice, 15(4), 469-482]. Subjective measurements included vocal assessment by a vocal pedagogue. In addition to the primary questions, we discuss issues of vocal variability, reproducibility, and correlations among the various measures.

[Work supported by the curriculum development funds of the Life Sciences Institute of the University of Michigan for Stradivarius as Biologist and by a National Science Foundation Graduate Research Fellowship awarded to the first author.]
 

Back to publications.