Unsupervised Speech Enhancement with Speech Recognition Embedding and Disentanglement Losses.
Viet Anh TrinhSebastian BraunPublished in: ICASSP (2022)
Keyphrases
- speech recognition
- speech enhancement
- noisy environments
- speech signal
- noisy speech
- speech synthesis
- vocal tract
- spectral subtraction
- hidden markov models
- automatic speech recognition
- language model
- pattern recognition
- background noise
- speaker identification
- linear prediction
- noise reduction
- additive noise
- spectral analysis
- signal to noise ratio
- speaker recognition
- single channel
- speech recognition systems
- multi channel
- sound source
- probabilistic model