 |
ITRW on Non-Linear Speech Processing
|
 |
|
ITRW on Non-Linear Speech Processing
|
|
Voiced segments of speech are assumed to be composed of non-stationary voiced acoustic objects which are generated as stationary (secondary) response of a non-stationary drive oscillator and which are analysed by introducing a selfconsistent part-tone decomposition. The self-consistency implies that the part-tones (of voiced continuants) are suited to reconstruct a topologically equivalent image of the hidden drive (glottal master oscillator). As receiver side image the fundamental drive (FD) is suited to describe the broadband voiced excitation as entrained (synchronized) and/or modulated primary response and to serve as low frequency part of the basic time-scale separation of auditive perception, which separates phone or timbre specific processes from intonation and prosody. The self-consistent time-scale separation avoids the conventional assumption of stationary excitation and represents the basic decoding step of the phasemodulation transmission-protocol of self-consistent (voiced) acoustic objects. The present study is focussed on the adaptation of the contours of the centre frequency of the parttone filters to the chirp of the glottal master oscillator.
Bibliographic reference. Drepper, Friedhelm R. (2007): "Non-stationary self-consistent acoustic objects as atoms of voiced speech", In NOLISP-2007, 75-79.