The contribution of changes in F0 and spectral tilt to increased intelligibility of speech produced in noise.
Youyi LuMartin CookePublished in: Speech Commun. (2009)
Keyphrases
- speech recognition
- signal to noise ratio
- noisy environments
- speech enhancement
- noise reduction
- spectral features
- noise level
- linear predictive coding
- noisy data
- missing data
- hidden markov models
- noise model
- speech signal
- background noise
- random noise
- linear prediction
- cepstral coefficients
- gaussian noise
- spectral analysis
- automatic speech recognition
- neural network
- image processing
- low signal to noise ratio
- pattern recognition
- multiscale