Lawrence R. Brown
The Recording and Res. Ctr., The Denver Ctr. for the Performing Arts, 1245 Champa St., Denver, CO 80204
Four singers (soprano, alto, tenor, and bass) simultaneously recorded unison /a/ vowels on middle-C (261 Hz) in seven different conditions: No instruction, matching or aggregating vibratos, matching or aggregating a nonvibrato tone quality, matching vowel qualities with and without vibrato, and prescribed instructions for a ``target'' vowel with and without vibrato. A perceptual panel of choral directors, singers, and naive auditors rated the blend or homogeneous quality of randomized samples of the unison vowels. Spectral analyses of the unison /a/ samples indicated varying deviations of the actual frequency of the first 14 partials (0--4000 Hz) from the calculated integer multiples of the fundamental frequency. The tabulated results of the auditors were correlated with the percentage error between actual and calculated partial frequencies. [Work supported by NIH Grant No. R08 DC01150-03.]