Phonatory dynamics and states - project background and first approaches
The human voice is a complex apparatus that transforms a DC pressure from the lungs into an AC acoustic volume velocity. Such a conversion is a result of a nonlinear process. The nonlinearity in the oscillations of the vocal folds is manifest in the variation of the waveform shape with its amplitude, as well as in bifurcations i.e. transitions between different vibrational modes. We are exploring the possibility of identifying and tracking such states and state transitions of the voice source while a subject vocalizes over the dynamic and frequency range, by utilizing non-invasive measurements and real-time phase-space analysis techniques. This will involve the combination of several techniques from non-linear dynamics, fourier descriptors and pattern recognition.