Improving Prosody for Unseen Texts in Speech Synthesis by Utilizing Linguistic Information and Noisy Data.
Zhu LiYuqing ZhangMengxi NieMing YanMengnan HeRuixiong ZhangCaixia GongPublished in: CoRR (2021)
Keyphrases
- speech synthesis
- noisy data
- linguistic information
- legal texts
- speech recognition
- text to speech
- prosodic features
- linguistic features
- structural information
- semantic information
- high dimensional
- part of speech
- input data
- missing data
- training data
- high dimensionality
- database
- noisy environments
- multiword
- language model
- pattern recognition
- computer vision