A Speech Synthesis Device using Formant and Residual Information

A new speech synthesis device has been developed. This speech synthesizer is based on a formant synthesizer and uses residual information. Generally, synthesized speech using residual information has high quality. In this speech synthesizer, the residual information is introduced as FIR (Finite Impulse Response) filter coefficients. Great flexibility in bit rate can be obtained by this implementation. Speech synthesis experiment showed the flexibility and the high performance for this speech synthesizer. This speech synthesizer can be applied to a high quality text to speech synthesis system and many other systems.