In the history of mankind, there are those who have led each era or have received great reception from the public who have made great achievements.These great men are not merely excellent in appearance, but also because their spirit, their actions, their horses, their achievements are superior to ordinary people. Today, the development of information and communication technology can record a person s appearance, voice, and actions, and descendants can store, transmit, and reproduce this information. However, it is difficult to restore the original sound even if you want to hear the voice of a great men or the voice of a great singer recorded in a harsh environment in the past. In addition, it is the more difficulty to reconstruct the voices of the great men who have no clue that might be the basis of restoration, even if they have excellent science and technology. In this way, not only can restored voices be communicated to descendants, it has the advantage of being able to convey the message of encouragement and love during the lifetime of parents who died early. These voices contain many verbal and non-verbal information. In this study, we have researched and synthesized high-quality voices by adjusting synthesis parameters in the time-frequency domain based on past data in relation to restoration of voices.