Performance measurement of a novel pitch detection scheme based on weighted autocorrelation for speech signals

A novel pitch detection scheme (PDS) based on weighted autocorrelation function (WACF) is proposed. The proposed scheme has been simulated and then integrated in an analysis-synthesis system for speech signal. The simulation and real-time performance comparison of this scheme with two other existing schemes [ACF and weighted ACF (WACF)] has been carried out. The performance comparison results show that the proposed PDS outperforms (in terms of speech quality and intelligibility) for both clean as well as noisy environment as compared to the other conventional PDS schemes considered for the comparison. Moreover, simulation and real-time implementation results show that the time taken for computation and memory consumption for the proposed PDS is less as compared to the weighted ACF based PDS.