Character-Level Linguistic Features Extraction for Text-to-Speech System
碩士 === 國立臺北科技大學 === 電子工程系研究所 === 104 === Good context dependent is a key part of the speech synthesis, the traditional context dependent depend on NLP (Natural Language Processing, NLP) parser text analysis. It is hence difficult to design one especially for speech synthesis. To alleviate these draw...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Online Access: | http://ndltd.ncl.edu.tw/handle/nh7kh2 |
| Summary: | 碩士 === 國立臺北科技大學 === 電子工程系研究所 === 104 === Good context dependent is a key part of the speech synthesis, the traditional context dependent depend on NLP (Natural Language Processing, NLP) parser text analysis. It is hence difficult to design one especially for speech synthesis. To alleviate these drawbacks, establishing an end-to-end speech synthesis system, we propose to use character-level word2vector and recurrent neural networks (RNNs) to directly convert input character sequences into latent linguistic feature vectors for context dependent. In the end, we use a mixed English-Chinese speech synthesis system to test this idea. Experimental results show that proposed approach provides comparable performance with conventional NLP parser-based methods.
|
|---|