DocumentCode :
1798641
Title :
A novel speech reconstruction algorithm for DSR back-end
Author :
Jiang Wenbin ; Ying Rendong ; Liu Peilin
Author_Institution :
Sch. of Electron. Inf. & Electr. Eng., Shanghai Jiao Tong Univ., Shanghai, China
fYear :
2014
fDate :
7-9 July 2014
Firstpage :
367
Lastpage :
371
Abstract :
In this paper, a novel speech reconstruction algorithm for DSR back-end is proposed. The algorithm is based on the classic least-squares estimate, inverse short-time Fourier transform magnitude (LSE-ISTFTM) algorithm. Unlike the classic LSE-ISTFTM algorithm, initializing speech waveform with white noise, the proposed method reconstructs voiced and unvoiced speech waveform separately, initializing with a specific signal. Furthermore, the magnitude spectrum is inversed from MFCC with Moore-Penrose pseudo-inverse by Mel-scale weighting functions. The algorithm evaluation results show that the proposed Extended LSE-ISTFTM algorithm converges faster and more stable than the classic algorithm. The speech reconstruction results demonstrate that PESQ score of the proposed algorithm is higher than the classic LSE-ISTFTM algorithm and the DSR back-end method.
Keywords :
Fourier transforms; least squares approximations; speech processing; DSR back-end method; LSE-ISTFTM algorithm; MFCC; Mel-scale weighting functions; Moore-Penrose pseudo-inverse; PESQ score; least-square estimate inverse short-time Fourier transform magnitude algorithm; magnitude spectrum; speech waveform; white noise; Barium; MFCC; cepstral analysis; minimum phase; speech reconstruction; speech synthesis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Audio, Language and Image Processing (ICALIP), 2014 International Conference on
Conference_Location :
Shanghai
Print_ISBN :
978-1-4799-3902-2
Type :
conf
DOI :
10.1109/ICALIP.2014.7009817
Filename :
7009817
Link To Document :
بازگشت