Parsing the spectral envelope: Toward a general theory of vocal tone color

The purpose of this study is to identify issues with the visual models of sung vowels currently used within singing voice pedagogy and voice science texts, and to propose a conceptual framework and new visual models that may more accurately characterize objective elements of timbre present in the singing voice. The timbre of the classical female (and countertenor) voice exposes blind spots in these spectrographic and schematic models, notably that they accommodate certain ambiguities present in speech, but problematic when applied to singing (especially high-pitched, melismatic singing). Essentially, above the treble staff, vowel clarity disappears entirely not because of vowel substitutions or modifications, but because the simplicity of the listener's percept is too distant from the timbral complexity of speech. The manner in which this vowel clarity changes as pitch ascends informs a meaningful discussion of the psychoacoustics of sung pitches throughout the range of both male and female voices, and suggests locating the source of timbre not within the singing body (at the point of production), but rather within the listener's paradoxically limited hearing mechanism (at the point of perception). The author hopes to point toward a still elusive general theory of vocal tone color by proposing the following five principles of singing voice perception currently absent in the voice science and vocal pedagogy literature: absolute spectral tone color, the multiple missing fundamentals, local spectral coherence, weak tone color bridging, and the obvious true fundamental. This thesis explores both immediate pedagogical implications of this framework for singers and voice teachers, and also points to substantial revisions to the models used in the singing voice pedagogy and voice science literature.