Research has shown that different musical instruments have unique emotional characteristics; voice research has investigated the timbre and low-level acoustic features of the soprano voice and their correlation with emotional expressiveness. This thesis investigates how pitch, dynamics, and vowel influence the emotional characteristics of the western classical solo singing voices. Listening tests were conducted whereby listeners gave absolute judgments over ten emotional categories on the soprano, mezzo-soprano, alto, countertenor, tenor, baritone, and bass voice tones, comprised of two dynamic levels (loud, soft), three pitches (octave-steps within individual voices), and five vowels (A, E, I, O, U); the data were subsequently analyzed via logistic regression.

Regarding dynamics, loud...[ Read more ]

Research has shown that different musical instruments have unique emotional characteristics; voice research has investigated the timbre and low-level acoustic features of the soprano voice and their correlation with emotional expressiveness. This thesis investigates how pitch, dynamics, and vowel influence the emotional characteristics of the western classical solo singing voices. Listening tests were conducted whereby listeners gave absolute judgments over ten emotional categories on the soprano, mezzo-soprano, alto, countertenor, tenor, baritone, and bass voice tones, comprised of two dynamic levels (loud, soft), three pitches (octave-steps within individual voices), and five vowels (A, E, I, O, U); the data were subsequently analyzed via logistic regression.

Regarding dynamics, loud tones dominated the high-arousal categories (Happy, Heroic, Comic, Angry, Scary) whereas soft tones dominated the low-arousal categories (Romantic, Calm, Mysterious, Shy, Sad). Regarding pitch trends across the low-to-high pitch range, Happy, Heroic, Romantic, and Comic were overall upward; Mysterious, Shy, Scary, and Sad were undulating yet overall flat; Calm was asymmetric arch-shaped with peaks at A2 and E3; Angry was irregularly sawtooth-shaped. Regarding vowel, the low-arousal categories (Romantic, Calm, Mysterious, Shy, Sad) and overall had a similar vowel profile of A downward to E then upward to I, O, U; there were no discernible vowel profile similarities among the remaining high-arousal categories. The overall vowel strength-of-expressiveness ranking was U first, followed by O and A, with I and E last; synthesized vocal/choral instruments typically only supply A, O, and U vowel samples. Among the voices, the soprano and countertenor had comparatively lower significant differences between the vowels; vice versa for the baritone.

Overall, pitch had the strongest marginal effect on the emotional characteristics, closely followed by dynamics, with both effects approximately twice as strong as the vowel marginal effect. These results give a quantified preliminary perspective on how pitch, dynamics, and vowel shape emotional expression in the western classical solo singing voices.