Spectral Subtraction Steered by Multi-Step Forward Linear Prediction For Single Channel Speech Dereverberation.
Keisuke KinoshitaTomohiro NakataniMasato MiyoshiPublished in: ICASSP (1) (2006)
Keyphrases
- speech enhancement
- linear prediction
- speech signal
- single channel
- image coding
- prediction error
- vocal tract
- speech recognition
- multi channel
- lossless compression
- sound source
- automatic speech recognition
- subband
- background noise
- spectral analysis
- autoregressive
- signal to noise ratio
- non stationary
- neural network
- noisy environments
- wavelet transform
- multiscale
- image compression
- filter bank
- brain activity
- high quality
- image segmentation