pyLeTalker 0.0.0

Wave-reflection voice synthesis framework

pyLeTalker - Wave-reflection voice synthesis framework (placeholder version)

pyLeTalker is a Python library to synthesize voice for research use, specifically aiming to enable the synethsis of pathologial voice. This library started as a repackaged version of LeTalker, a Matlab GUI demo by Dr. Brad Story but since has been evolved as a flexible general voice synthesis library built around the wave-reflection vocal tract model. The wave-reflection vocal tract model treats the vibrating air pressure as propagating waves through the vocal tract. Some portion of an incidental wave reflects when it encounters a change in the cross-sectional area of the vocal tract.

The pyLeTalker modularize the voice production system into 5 elements: lung, subglottal tract, vocal folds (glottis), supraglottal tract, and lips. Each subsystem is implemented as a Python class. The library comes with one or more classes for each voice production element. For example, there are two built-in vocal-fold models: LeTalkerVocalFolds (self-oscillating 3-mass model with muscle activation inputs) and KinematicVocalFolds (3D vocal fold model with preprogrammed oscillation pattern).

The other part of pyLeTalker is its function_generators subpackage to enable time-varying control of voice production models. Actual voice production is perpetually dynamic. The synthesis models accept both constant parameter values as well as a FunctionGenerator object, so that the user can better program the behaviors.

Both synthesis elements and function generators can be customized. Hence, pyLeTalker is suitable framework to evaluate new vocal fold models.