(A) study on speech excitation signal

Abstract

In this thesis, excitation of the LPC vocoder has been studied in three areas. First, a high speed pitch estimation algorithm that is based on peak detection and average magnitude difference function (AMDF) is developed. A few pitch candidates are first estimated from the low-pass filtered (800 Hz) speech by peak detection algorithm. AMDF values of the pitch candidate that yields the minimum AMDF value is chosen as the desired pitch period. The new method requires far less computation time than other pitch estimation algorithms, yet it yields reasonably accurate results. Second, an improved 4800 bps LPC vocoder system that virtually eliminates the buzzy effect from synthetic speech is developed. Excitation signal in the new system is formed by adding pitch pulse or random noise to the baseband residual (0-600 Hz) that has been coded by pitch prediction DPCM. According to out informal listening tests the synthetic speech of the new system does not have the buzzy effect. As a result the vocoder speech quality is more natural than that of a conventional LPC vocoder. Third, a pitch extraction algorithm based on LPC inverse filtering and AMDF is hardware implemented using a bit slice micro-computer. The pitch extraction method that is modified to reduce the computation, its hardware architecture, and the system development procedure are described.