Speech enhancement using Bayesian estimators of the perceptually-motivated short-time spectral amplitude (STSA) with Chi speech priors.
Marek B. TrawickiMichael T. JohnsonPublished in: Speech Commun. (2014)
Keyphrases
- speech enhancement
- linear prediction
- speech signal
- noisy environments
- noise reduction
- spectral subtraction
- signal to noise ratio
- single channel
- vocal tract
- spectral analysis
- noisy speech
- image coding
- bayesian networks
- lossless compression
- edge preserving
- prior information
- speech recognition
- subband
- multi channel
- smoothing algorithm
- information retrieval
- spectral features
- speaker identification
- wiener filter
- additive noise
- automatic speech recognition
- bayesian framework
- prediction error
- background noise
- maximum likelihood
- edge detection
- autoregressive
- prior knowledge
- feature vectors