Voice Quality: Functions, Analysis and Synthesis

August 27-29, 2003
Geneva, Switzerland

The Voice Source as a Causal/Anticausal Linear Filter

Boris Doval (1), Christophe d’Alessandro (1), Nathalie Henrich (2)

(1) LIMSI-CNRS, Université Paris XI, Orsay, France
(2) LAM, Paris, France

A new type of glottal flow model, namely a causal-anticausal linear filter model, is proposed. It is shown that the glottal flow signal can be considered as the impulse response of a linear filter. Then the source/filter speech model can be interpreted as an excitation/filter speech model, the "filter" comprising the glottal flow, vocal tract and radiation components. The spectral features of the voice source models are reviewed, both in the amplitude and phase domains. In the spectral amplitude domain the main features are a spectral maximum (the "glottal formant") and spectral tilt. Evidence for a mixed causal/anticausal phase behavior of the source is given. Then, a causal-anticausal linear filter voice source model is designed. In conclusion, applications of this new approach are discussed for voice quality modification, voice source estimation, voice quality perception.

Full Paper    Presentation (PDF)

Bibliographic reference.  Doval, Boris / d’Alessandro, Christophe / Henrich, Nathalie (2003): "The voice source as a causal/anticausal linear filter", In VOQUAL'03, 16-20.