PSD estimation in Beamspace for Estimating Direct-to-Reverberant Ratio from A Reverberant Speech Signal.
Yusuke HiokaKenta NiwaPublished in: CoRR (2015)
Keyphrases
- speech signal
- speech recognition
- speaker identification
- automatic speech recognition
- spectral analysis
- estimation process
- noisy environments
- background noise
- hidden markov models
- non stationary
- linear prediction
- vocal tract
- automatic speech recognition systems
- speech enhancement
- speaker recognition
- sound source
- fundamental frequency
- speech quality
- pattern recognition
- language model
- cepstral coefficients
- feature extraction