Speech synthesis based on PSOLA algorithm and modified pitch parameters
Yuanyuan Wang, Yang Shun · International Conference on Computational Problem-Solving · 2010
Combined the prosodic features of Chinese speech, pitch is extracted accurately with autocorrelation function and three-level clipping technique based on PSOLA algorithm. The complexity of the algorithm is effectively reduced. Using the advantages of TD-PSOLA algorithm, the scale factors of pitch periods are analyzed and adjusted, after modifying the time length and fundamental frequency. Finally the shaking of sound is eliminated, and the synthesized speech is more clearly. The experiment shows that the synthesized speech has a high quality and a strong value through using this method.