Fewer-Token Neural Speech Codec with Time-Invariant Codes.
Yong RenTao WangJiangyan YiLe XuJianhua TaoChu Yuan ZhangJunzuo ZhouPublished in: ICASSP (2024)
Keyphrases
- network architecture
- speech recognition
- neural network
- neural model
- error correction
- speech signal
- speech synthesis
- video codec
- bio inspired
- bitstream
- automatic speech recognition
- video coding
- error correcting codes
- audio visual
- coding method
- recognition engine
- noisy environments
- text to speech
- endpoint detection
- motion estimation
- biologically plausible
- open loop
- spoken language
- logical operations
- motion compensated