完整後設資料紀錄
DC 欄位語言
dc.contributor.authorHwang, Hsin-Teen_US
dc.contributor.authorTsao, Yuen_US
dc.contributor.authorWang, Hsin-Minen_US
dc.contributor.authorWang, Yih-Ruen_US
dc.contributor.authorChen, Sin-Horngen_US
dc.date.accessioned2018-08-21T05:56:37Z-
dc.date.available2018-08-21T05:56:37Z-
dc.date.issued2013-01-01en_US
dc.identifier.issn2308-457Xen_US
dc.identifier.urihttp://hdl.handle.net/11536/146417-
dc.description.abstractIn this paper, we propose a discriminative training (DT) method to alleviate the muffled sound effect caused by over smoothing in the Gaussian mixture model (GMM)-based voice conversion (VC). For the conventional GMM-based VC, we often observed a large degree of ambiguities among acoustic classes (generative classes), determined by the source feature vectors for generating the converted feature vectors, causing the "muffled sound" effect on the converted voice. The proposed DT method is applied to refine the parameters in the maximum likelihood (ML)-trained joint density GMM (JDGMM) in the training stage to reduce the ambiguities among acoustic classes (generative classes) to alleviate the muffled sound effect. Experimental results demonstrate that the DT method significantly enhances the discriminative power between acoustic classes (generative classes) in the objective evaluation and effectively alleviates the muffled sound effect in the subjective evaluation.en_US
dc.language.isoen_USen_US
dc.subjectVoice conversionen_US
dc.subjectdiscriminative trainingen_US
dc.subjectGMMen_US
dc.titleAlleviating the Over-Smoothing Problem in GMM-Based Voice Conversion with Discriminative Trainingen_US
dc.typeProceedings Paperen_US
dc.identifier.journal14TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2013), VOLS 1-5en_US
dc.citation.spage3061en_US
dc.citation.epage3065en_US
dc.contributor.department電機工程學系zh_TW
dc.contributor.departmentDepartment of Electrical and Computer Engineeringen_US
dc.identifier.wosnumberWOS:000395050001156en_US
顯示於類別:會議論文