In this talk the development of a new parameterization technique for the harmonic phase
of the speech signal will be described. The harmonic phase is derived from the classical
Harmonic plus Stocastic model, and simple manipulations over the original phase are
performed, leading to a nice and meaningful representation named Relative Phase Shift.
These values have been used in several speech processing applications, such as polarity
detection, Speech and Speaker recognition and detection of synthetic speech signals.
Some experiments and results will be described in the talk.
Additionally, the history and research activities at the Aholab Signal Processing
Laboratory of the University of the Basque Country will be presented.