Improving Prosody with Linguistic and Bert Derived Features in Multi-Speaker Based Mandarin Chinese Neural TTS.
Yujia XiaoLei HeHuaiping MingFrank K. SoongPublished in: ICASSP (2020)
Keyphrases
- text to speech
- neural network
- prosodic features
- feature extraction
- structural information
- feature set
- artificial neural networks
- feature vectors
- co occurrence
- linguistic information
- linguistic knowledge
- linguistic features
- associative memory
- data sets
- genetic algorithm
- speaker verification
- extracting features
- structural features
- pattern recognition
- high level
- image features
- feature space