Speech bandwidth extension method using speech recognition and speech synthesis
作者： Masashi Takashina , Shingo Kuroiwa , Satoru Tsuge , Fuji Ren
关键词: Speech processing 、 Intelligibility (communication) 、 Speech technology 、 Speech synthesis 、 Linear predictive coding 、 Acoustic model 、 Speech recognition 、 Voice activity detection 、 Bandwidth extension 、 Codec2 、 Computer science 、 PSQM 、 Speech coding 、 Signal processing
摘要: In this paper, we propose two kinds of novel speech bandwidth extension methods based on a recognition, synthesis and signal processing technique. the proposed methods, lost component by bandlimiting is generated with technique synthesis, using bandlimited information. For evaluating both conducted subjective objective evaluation experiments. The experimental results shows effectiveness about feeling bandwidth. However, quality became not good result. Our problem an improvement method in future.
sci-hub.se 参考文章(5)
Masanobu Abe, Yuki Yoshida, An algorithm to reconstruct wideband speech from narrowband speech based on codebook mapping conference of the international speech communication association. pp. 1591- 1594 ,(1994)
Yoshihisa Nakatoh, Mineo Tsushima, Takeshi Norimatsu, Generation of broadband speech from narrowband speech based on linear mapping Electronics and Communications in Japan (Part II: Electronics). ,vol. 85, pp. 44- 53 ,(2002) , 10.1002/ECJB.10065
Nami Hatazoe, Yoshinori Tanaka, Reconstruction of Wideband Speech From Telephone-band Speech by Multi-layer Neural Networks 日本音響学会研究発表会講演論文集. ,vol. 1995, pp. 255- 256 ,(1995)
哲司 塔下, Shinji Wakizaka, Kazuyoshi Ishiwatari, 新路 脇坂, Tetsuji Toushita, 一嘉 石渡, Koji Ito, 功二 伊東, Makoto Tanaka, 田中 誠, Speech recognition system ,(1996)
Jia-Jang Tu, Shih-Chien Chien, Sen-Chia Chang, Speech recognition system ,(2002)