Acoustic BPE for Speech Generation with Discrete Tokens.
Feiyu ShenYiwei GuoChenpeng DuXie ChenKai YuPublished in: ICASSP (2024)
Keyphrases
- continuous variables
- bayesian networks
- speech recognition
- speech sounds
- speech recognition systems
- acoustic features
- speech signal
- acoustic signal
- emotional speech
- generation process
- speech recognizers
- audio visual
- sensor networks
- speaker independent
- prosodic features
- endpoint detection
- noisy environments
- underwater acoustic
- neural network
- line segments
- multimedia