Time-domain deterministic plus noise model based hybrid source modeling for statistical parametric speech synthesis
详细信息 查看全文
- 作者：N.P. Narendra ; narendrasince1987@gmail.com" class="auth_mail" title="E-mail the corresponding author ; narendranp666@gmail.com" class="auth_mail" title="E-mail the corresponding author ; K. Sreenivasa Rao ksrao@iitkgp.ac.in" class="auth_mail" title="E-mail the corresponding author
- 关键词：Statistical parametric speech synthesis ; Deterministic plus noise model ; Hybrid source modeling ; Pitch-synchronous residual frame
- 刊名：Speech Communication
- 出版年：2016
- 出版时间：March 2016
- 年：2016
- 卷：77
- 期：Complete
- 页码：65-83
- 全文大小：1426 K
文摘
- •
The source signal is modeled as a combination of deterministic and noise components.
- •
The estimated deterministic components are stored in the form of a decision tree.
- •
The noise components are parameterized in terms of spectrum and amplitude envelopes.
- •
During synthesis, deterministic component is selected from the leaf of a decision tree.
- •
The noise component is generated from the natural instance of the noise signal.