Speech dereverberation with multi-channel linear prediction and sparse priors for the desired signal.
Ante JukicToon van WaterschootTimo GerkmannSimon DocloPublished in: HSCMA (2014)
Keyphrases
- linear prediction
- multi channel
- single channel
- speech signal
- image coding
- compressive sensing
- prediction error
- lossless compression
- signal reconstruction
- non stationary
- vocal tract
- autoregressive
- speech recognition
- signal processing
- noisy environments
- additive noise
- sound source
- prior information
- subband
- high frequency
- probabilistic model
- speaker identification
- hidden markov models
- spectral analysis
- prior knowledge
- wavelet transform
- high quality