Controlling the convergence of vocal tract resonances to generate apparent multiphonic effects

A talker or singer typically controls the configuration of their vocal tract to produce a pattern of resonances sequentially ordered in frequency. This pattern carries information regarding the identity and quality of vowels and consonants by enhancing some parts of the voice source spectrum while suppressing others. In some forms of human sound making, however, the vocal tract is configured such that two or more resonances converge at essentially the same frequency, producing a particularly effective enhancement of a small band of the voice source spectrum. In various types of overtone singing, for example, the convergent resonances allow the singer to elevate a single harmonic of the voice spectrum to an amplitude high enough that listeners may perceive it as a sound source decoupled from the primary voice source. This presentation will review and expand previous work on how vocal tract constrictions and expansions can be systematically controlled to generate convergent resonances and to move them upward and downward in frequency for purposes of producing sounds with spectra that are distinctly different from more typical speaking and singing.