Synthesized speech quality measurement of an Improved Fundamental Frequency (Pitch) Detection Algorithm

The fundamental frequency (pitch) is essential in determining speech quality. This paper proposes a modified circular average magnitude difference function (CAMDF) based algorithm for estimating the pitch of the speech signal. The performance of the proposed algorithm has been compared with autocorrelation function (ACF), average magnitude difference function (AMDF), and CAMDF based pitch detection methods by incorporating them into an analysis-synthesis system for speech signal. The proposed method has a lower gross error of 5.33%, 4.7%, and 0.83 % compared to AMDF, ACF, and CAMDF methods. Also, the synthesized speech quality of the proposed algorithm in respect of subjective and objective tests is superior with respect to other algorithms.