On the Importance of Super-Gaussian Speech Priors for Pre-Trained Speech Enhancement.
Robert RehrTimo GerkmannPublished in: CoRR (2017)
Keyphrases
- speech enhancement
- pre trained
- noisy environments
- speech signal
- noise reduction
- signal to noise ratio
- vocal tract
- single channel
- linear prediction
- speech recognition
- training data
- noisy speech
- multi channel
- wiener filter
- training examples
- control signals
- smoothing algorithm
- additive noise
- maximum likelihood
- prior information
- background noise
- speech synthesis
- pattern recognition
- edge detection
- learning algorithm