DocumentCode :
2575482
Title :
Voice conversion based on style and content separation with dual latent variable model
Author :
Sun, Xinjian ; Zhang, Xiongwei ; Sun, Jian
Author_Institution :
Inst. of Commun. Eng., PLA Univ. of Sci. & Tech., Nanjing, China
fYear :
2011
fDate :
9-11 Nov. 2011
Firstpage :
1
Lastpage :
5
Abstract :
This paper presents a novel method for voice conversion based on style and content separation, which is solved by using dual latent variable model (D-LVM). Based on D-LVM, the vocal tract spectrum of speech represented by line spectral frequencies (LSF) is explicitly decomposed into so-called style and content factors, which are used to represent the speech meaning and the speaker individuality respectively. On the basis of reasonable separation of style and content for speech, voice conversion is performed successfully by reproducing converted speech using the initial speech content and the target speaker style. The objective and subjective tests show that, under the condition of limited training dataset, the method proposed in the paper gets better conversion performance compared to the conventional mapping based GMM system and SVD based bilinear model.
Keywords :
content management; singular value decomposition; speaker recognition; GMM system; SVD based bilinear model; Xiongwei vocal tract spectrum; content factor; content separation; dual latent variable model; line spectral frequency; speaker individuality; speech meaning; speech representation; style based voice conversion; target speaker style; Data models; Fitting; Probabilistic logic; Speech; Speech recognition; Training; Training data; latent variable model; separation; style and content; voice conversion;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Wireless Communications and Signal Processing (WCSP), 2011 International Conference on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4577-1009-4
Electronic_ISBN :
978-1-4577-1008-7
Type :
conf
DOI :
10.1109/WCSP.2011.6096805
Filename :
6096805
Link To Document :
بازگشت