Minimization of Error in Pitch Detection algorithm using Discrete Fractional Cosine Transform

Abstract

A novel Discrete Fractional Cosine Transform implementation for pitch estimation of noisy speech using dominant harmonic is introduced in this paper. The basic idea is to preprocess the speech signal with discrete fractional cosine transform before using the rectified dominant harmonic for signal reshaping. The performance of the proposed method is tested and compared with the latest previous method using Keele pitch extraction reference database. Comparative results show that the proposed method can detect pitch with better accuracy in terms of gross and fine pitch error as compared to the latest technique.