Speech synthesis method with a harmonic plus noise model

The requirements of text-to-speech synthesis are increasing, since the speech interface is an appropriate user-friendly interface for consumer applications. However, in general, the quality of synthetic speech is not satisfactory, and speech synthesis requires a large amount of computation and memory. In this paper, we propose a high quality speech synthesis method for text-to-speech systems and report its implementation using a FPGA. The experimental results show the proposal method can generate high quality synthetic speech, and we can realize a low-power FPGA-based hardware for consumer applications.