A Novel Iterative Speaker Model Alignment Method from Non-Parallel Speech for Voice Conversion

Peng SONG  Wenming ZHENG  Xinran ZHANG  Yun JIN  Cheng ZHA  Minghai XIN  

Publication
IEICE TRANSACTIONS on Fundamentals of Electronics, Communications and Computer Sciences   Vol.E98-A   No.10   pp.2178-2181
Publication Date: 2015/10/01
Online ISSN: 1745-1337
DOI: 10.1587/transfun.E98.A.2178
Type of Manuscript: LETTER
Category: Speech and Hearing
Keyword: 
non-parallel speech,  voice conversion,  iterative speaker model alignment,  Gaussian mixture model,  

Full Text: PDF(721KB)>>
Buy this Article




Summary: 
Most of the current voice conversion methods are conducted based on parallel speech, which is not easily obtained in practice. In this letter, a novel iterative speaker model alignment (ISMA) method is proposed to address this problem. First, the source and target speaker models are each trained from the background model by adopting maximum a posteriori (MAP) algorithm. Then, a novel ISMA method is presented for alignment and transformation of spectral features. Finally, the proposed ISMA approach is further combined with a Gaussian mixture model (GMM) to improve the conversion performance. A series of objective and subjective experiments are carried out on CMU ARCTIC dataset, and the results demonstrate that the proposed method significantly outperforms the state-of-the-art approach.