Frequency Domain Multi-channel Acoustic Modeling for Distant Speech Recognition.
Minhua WuKen'ichi KumataniShiva SundaramNikko StromBjörn HoffmeisterPublished in: CoRR (2019)
Keyphrases
- frequency domain
- speech recognition
- multi channel
- single channel
- speech recognition systems
- spatial domain
- speaker independent
- speech recognizers
- hidden markov models
- language model
- speech recognizer
- automatic speech recognition
- cross correlation
- speech signal
- feature extraction
- speaker identification
- power spectrum
- noisy environments
- speech synthesis
- acoustic models
- pattern recognition
- acoustic features
- denoising
- noisy speech
- wavelet domain
- independent component analysis
- gaussian mixture model
- subband
- non stationary