Controllable speech synthesis by learning discrete phoneme-level prosodic representations.
Nikolaos EllinasMyrsini ChristidouAlexandra VioniJune Sig SungAimilios ChalamandarisPirros TsiakoulisParis MastorocostasPublished in: CoRR (2022)
Keyphrases
- speech synthesis
- speech recognition
- prosodic features
- text to speech
- learning process
- learning algorithm
- learning systems
- online learning
- knowledge level
- higher level
- learning problems
- word processing
- active learning
- reinforcement learning
- neural network
- supervised learning
- learning activities
- low level
- automatic speech recognition
- multiple representations