Speech dereverberation using linear prediction with estimation of early speech spectral variance.
Mahdi ParchamiWei-Ping ZhuBenoît ChampagnePublished in: ICASSP (2016)
Keyphrases
- linear prediction
- speech signal
- prediction error
- minimum variance
- image coding
- speech enhancement
- lossless compression
- vocal tract
- speech recognition
- spectral analysis
- subband
- automatic speech recognition
- cepstral coefficients
- speaker identification
- parameter estimation
- hidden markov models
- non stationary
- acoustic features
- bit rate
- model selection
- bayesian networks