Title :
Modeling glottal effect on the spectral envelop of STRAIGHT using mixture of Gaussians
Author :
Ling, Zhen-Huu ; Wang, Yu-Ping ; Hu, Yu ; Wang, Ren-Hua
Author_Institution :
iFlytek Speech Lab., Univ. of Sci. & Technol. of China, Hefei, China
Abstract :
This paper presents a method to model the influence of glottal excitation on the STRAIGHT (speech transformation and representation using adaptive interpolation of weighted spectrum) spectrum by fitting the spectral envelop with a mixture of Gaussians (MOG). The first Gaussian component is used as the estimation for the glottal formant in the STRAIGHT spectrum because analysis results show that it has an obviously stronger correlation with fundamental frequency than other spectral components and has similar characteristics to the glottal formant. Then linear regression is carried out to measure the relationship between F0 and the parameters of the first Gaussian component. This model is applied to the STRAIGHT synthesis process and proved to be effective in compensating the voice quality variation caused by pitch modification.
Keywords :
Gaussian distribution; adaptive signal processing; frequency estimation; interpolation; regression analysis; spectral analysis; speech coding; speech synthesis; vocoders; STRAIGHT spectrum; STRAIGHT synthesis process; fundamental frequency; glottal excitation; glottal formant estimation; linear regression; mixture of Gaussians; pitch modification; spectral envelop fitting; speech transformation and representation using adaptive interpolation of weighted spectrum; voice quality variation compensation; Frequency estimation; Gaussian processes; Interpolation; Laboratories; Linear regression; Roentgenium; Smoothing methods; Speech analysis; Speech synthesis; Time domain analysis;
Conference_Titel :
Chinese Spoken Language Processing, 2004 International Symposium on
Print_ISBN :
0-7803-8678-7
DOI :
10.1109/CHINSL.2004.1409589