| |
Abstract:
Voice quality generally refers to the extralinguistic features
of a speaker's voice that may provide cues to identity,
personality, health, and emotional state. Such features provide
much of the acoustic variability found in speech signals across
individuals. A broad description of voice quality would include
features contributed by all the subsystems of speech production;
i.e. respiratory, phonatory and articulatory systems. To narrow
the focus, this paper is concerned with modeling and simulating
only those aspects of voice quality that arise due to the shaping
of the vocal tract. This work was influenced by Laver [Laver, J.
(1980). The Phonetic Description of Voice Quality, Cambridge
University Press.], who proposed that long-term "settings" of the
vocal tract bias the resulting formant structure toward a
particular type of global timbre. He defined two broad categories
of vocal tract settings as longitudinal and latitudinal.
Longitudinal settings describe the state of the long axis of the
vocal tract such as larynx height and protrusion/retraction of
the lips. The latitudinal settings are "tendencies to maintain a
particular constrictive (or expansive) effect" within some region
located along the length of the vocal tract. In this study, the
idea is pursued that voice quality can be partially represented
by the underlying shape of a speaker's neutral vocal tract, while
more generic or canonical movement patterns are used to
superimpose the linguistically relevant deformations on the
neutral shape. Using an area function model, which allows direct
access to the neutral tract shape, four separate modifications
were made to one male speaker's vocal tract. The modifications
involve the pharyngeal and oral cavities as well as lip aperture
and the size of the epi-laryngeal tube. A single word utterance
and a sentence were first simulated with the original neutral
tract shape and then each was simulated again using the four
modified neutral shapes. The modifications are demonstrated with
sound files and resulting formant trajectories corresponding to
each modification are shown and discussed.
|
