![]() |
ITRW on Non-Linear Speech Processing
|
![]() |
Voice source analysis is an important but difficult issue for speech processing. In this talk, three aspects of voice source analysis recently developed at LIMSI (Orsay) and FPMs (Mons) are discussed. In a first part, time domain and spectral domain modelling of glottal flow signals are presented (Doval & al., 2006). It is shown that the glottal flow can be modelled as an anticausal filter (maximum phase) before the glottal closure, and by as a causal filter (minimum phase) after the glottal closure. In a second part, taking advantage of this phase structure, causal and anticausal components of the speech signal are separated according to the location in the Z-plane of the zeros of the Z-Transform (ZZT) of the windowed signal (Bozkurt & al. 2005). Results of a comparative evaluation of the ZZT and linear prediction for source/tract separation are reported. This method is also useful for voice source parameters analysis. Voice open quotient and glottal flow asymmetry analyses using the ZZT and ElectroGlottoGraphy (EGG) are compared (Sturmel & al. 2006). In a third part, glottal closure instant detection using the phase of the wavelet transform is discussed. A method based on the lines of maximum phase in the time-scale plane is proposed. This method is compared to EGG for robust glottal closure instant analysis (Vu Ngoc Tuan & al, 2000).
Bibliographic reference. d'Alessandro, Christophe (2007): "Phase-based methods for voice source analysis", In NOLISP-2007 [abstract only]