Improving Character Error Rate Is Not Equal to Having Clean Speech: Speech Enhancement for ASR Systems with Black-box Acoustic Models.
Ryosuke SawataYosuke KashiwagiShusuke TakahashiPublished in: CoRR (2021)
Keyphrases
- error rate
- black box
- word error rate
- speech recognition
- automatic speech recognition
- speech signal
- noisy speech
- black boxes
- speech enhancement
- noisy environments
- acoustic models
- test set
- speech recognition systems
- hidden markov models
- broadcast news
- noise reduction
- language model
- speech recognizer
- pattern recognition
- signal to noise ratio
- linear prediction
- background noise
- vocal tract
- feature extraction
- acoustic features
- image processing
- multiscale
- speaker independent
- machine learning