Abstract

   Estimation of fundamental frequency is an area of interest in communications and speech processing, as well as in the investigation of various physical phenomena. We make in this paper a comparative analysis between the basic time domain approaches and those based on the reconstructed state space representation. We present the philosophies of these two directions and we propose a fundamental frequency estimation method based on quantification of diagonal lines in the recurrence matrix. We show that this method is in fact an extension of the difference function method. We also comparatively show results obtained by using these methods for different signals and for different SNRs (signal-to-noise ratios).