Waveform-based speech synthesis approach with a formant frequency modification

A novel approach to speech synthesis based on waveform segments is proposed. One novel point of this approach is its new formant frequency modification algorithm which makes it possible to change formant frequency flexibly and so reproduce the desired speech quality. The algorithm characterizes speech formants not only by formant frequencies and formant bandwidths, but also by spectral intensities of formant frequencies. The desirable formant structure, which is specified by the parameters, is obtained by iteratively modifying the formant bandwidths. Using the specified formant structure, the speech signal is synthesized by the FFT (fast Fourier transform). Evaluation by the acoustic distance measure and by listening tests confirms the good performance of the approach. As evaluated by listening tests, the proposed method was found to increase significantly the naturalness of speech and clearly to increase speech quality.<>