A Single-Channel Non-Intrusive C50 Estimator Correlated With Speech Recognition Performance.
Pablo Peso ParadaDushyant SharmaJose LainezDaniel BarredaToon van WaterschootPatrick A. NaylorPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2016)
Keyphrases
- speech recognition
- single channel
- multi channel
- speech enhancement
- language model
- speech signal
- independent component analysis
- automatic speech recognition
- hidden markov models
- frequency domain
- prior information
- speech synthesis
- maximum likelihood
- speech recognition systems
- wavelet decomposition
- wiener filter
- pattern recognition
- signal processing
- noisy environments
- sound source
- brain activity
- machine learning
- noise reduction
- heart rate
- multiresolution
- image sequences