Detection of glottal closure instant and glottal open region from speech signals using spectral flatness measure

Sudarsana Reddy Kadiri, RaviShankar Prasad, B. Yegnanarayana

Abstract: Highlights•A new approach for determination of glottal closure instant and glottal open region from speech signals is proposed based on the changes in the spectral characteristics within a glottal cycle, using two signal processing methods, namely, zero time windowing (ZTW) and single frequency filtering (SFF).•The spectral flatness parameter derived from the SFF spectra highlights the impulse-like characteristics at the glottal closure instant (GCI).•The spectral flatness derived from the ZTW spectra highlights the glottal opening, as there are significant changes in the effective vocal tract length in the glottal opening region (GOR).•The proposed method of GCI detection is compared with several existing methods in clean speech and telephone quality speech scenarios. The results of the proposed method is comparable to the existing methods in clean speech and better in telephone quality speech.•The proposed method of GOR detection is shown to be comparable with the existing methods.