Using continuous lexical embeddings to improve symbolic-prosody prediction in a text-to-speech front-end.
Asaf RendelRaul FernandezRon HooryBhuvana RamabhadranPublished in: ICASSP (2016)
Keyphrases
- distance measure
- text to speech
- speech synthesis
- vector space
- prosodic features
- programming tool
- prediction accuracy
- data sets
- improve the prediction accuracy
- word processing
- text to speech synthesis
- english text
- prediction model
- high level
- neural network
- visual information
- speech recognition
- multi modal
- hidden markov models
- artificial neural networks
- predictive clustering trees