白色ガウス雑音を用いた日本語音声の合成
スポンサーリンク
概要
- 論文の詳細を見る
In this experiment, by using white Gaussian noise and Discrete Fourier Transform (DFT), spoken Japanese language was synthesized. Firstly, the spectrum of voice signals was calculated by DFT and multiplied to spectrum of white Gaussian noise. Secondly, through inverse DFT, synthesized voice signals was obtained. Furthermore, to improve this signal performance, several bandpath filters were used. Synthetic speech was evaluated by the human auditory sense and quantification theory I.
- 明石工業高等専門学校の論文