An effective speech preprocessing technique for normalized cross-correlation pitch extractor

A noise robust method for pitch estimation based on a new speech preprocessing technique is presented in this paper. The rectified dominant harmonic of the pre-filtered signal is used for preprocessing. The normalized crosscorrelation function and circular average magnitude difference function are used for pitch extraction. The performance of the proposed method is tested and compared with a recent method using the Keele pitch extraction reference database. Comparative results show that the proposed method can detect pitch with better accuracy for a wide range of signal-to-noise ratios (SNRs).