On a spectral scaling technique for pitch control in the high-quality text-to-speech type system

Speech synthesis technique is classified into three groups; waveform coding, source coding and hybrid coding. Waveform coding and hybrid coding are used for sentence-based-synthesis by synthesis-by-analysis method; the difficulty of pitch altering makes them inappropriate for synthesis-by-rule. However, if it is possible to alter the pitch period when the waveform coding is used, synthesis-by-rule is available maintaining good intelligibility and naturalness comparable to the original speech. In this paper, we propose a new pitch alteration method that can change the pitch period in waveform coding by scaling the time-axis and where phase compensation is performed by using the zero-inserting and pitch-halving method.