Codec-ASR: Training Performant Automatic Speech Recognition Systems with Discrete Speech Representations.
Kunal DhawanNithin Rao KoluguriAnte JukicRyan LangmanJagadeesh BalamBoris GinsburgPublished in: CoRR (2024)
Keyphrases
- automatic speech recognition systems
- speech signal
- automatic speech recognition
- speech recognition
- noisy speech
- hidden markov models
- noisy environments
- vocal tract
- speaker identification
- speech enhancement
- training set
- non stationary
- speaker recognition
- background noise
- additive noise
- linear prediction
- broadcast news
- video coding
- motion estimation
- multiscale
- hearing impaired
- feature extraction
- image processing
- adaptive filtering
- bitstream
- noisy images
- acoustic features
- speech synthesis
- pattern recognition