Integrating Discrete Word-Level Style Variations into Non-Autoregressive Acoustic Models for Speech Synthesis.
Zhaoci LiuNing-Qian WuYajie ZhangZhenhua LingPublished in: INTERSPEECH (2022)
Keyphrases
- autoregressive
- speech synthesis
- word level
- speech recognition
- non stationary
- language independent
- document images
- text to speech
- automatic speech recognition
- machine translation
- n gram
- hidden markov models
- language model
- random fields
- word recognition
- document analysis
- speech signal
- word segmentation
- noisy environments
- character recognition
- sar images
- pattern recognition
- sentence level
- image processing
- neural network
- conditional random fields
- pairwise
- information retrieval