Model-based speech enhancement using SNR dependent MMSE estimation.
Thomas EschPeter VaryPublished in: ICASSP (2011)
Keyphrases
- speech enhancement
- signal to noise ratio
- minimum mean square error
- noise reduction
- noisy environments
- single channel
- speech signal
- linear prediction
- edge detection
- vocal tract
- wiener filter
- smoothing algorithm
- edge preserving
- multi channel
- additive noise
- background noise
- automatic speech recognition
- speech recognition
- multiscale