A quasi-glottogram signal

We propose a novel signal processing technique that reliably shows the strength of glottal oscillations. The Pseudo-glottogram (PGG) signal can improve voicing estimates, and can quantify partial voicing. The technique is well adapted to the generation of text-to-speech systems, merely requiring extra microphones near the speaker's throat. For prosody studies, it can be used to provide a robust, noninvasive estimate of the subglottal pressure, or the speech effort. The PGG signal provides a more stable amplitude measurement than either total acoustic power or the inverse filtered mouth signal.