In this case, after TTS sentences including the running time, sex, age, speech speed, TTS characters, and prosody information are produced through the input sentences, a bit stream is generated according to M-TTS syntax, and the speech EPG is reproduced through an M-TTS decoder.