Thai speech synthesis based on Formant synthesis for home robot

A few years ago, the home robot has an important role for the human life. Human needs to operate the robot of the work instead for the humans, and some robots can talk with humans speech, such as English language or Japanese language. On the other hand, but Thai language is different from other languages because it is the tonal language, which one word can be pronounced five tones. The most Thai speech synthesis use concatenation technique, which does not flexible to adjust the parameters and emotional tone cannot be performed. Therefore, this paper proposes Thai speech synthesis based on Formant synthesis to synthesize Thai syllables for home robot. The parameters of the proposed system can be adaptively adjusted by the rule-based. The parameters will be adapted by the rules to synthesize the desired speech. The Fundamental frequency is a significant parameter. In proposed method, the main part of the proposed system focuses on development of the tonal pattern rules. These rules consist of the consonant group, vowel group, and spelling group. The system starts with replacing each phoneme by an input array with corresponded Formant frequencies. After that, phoneme array are adapted process by rules. We utilized 50 samples of Thai syllable. Each syllable is tested by 10 peoples to listen in the control room. The experimental results showed that the synthesized syllable correct in Thai language as 71.8 percent.