On spectral and time domain energy reallocation for speech-in-noise intelligibility enhancement.
Tudor-Catalin ZorilaYannis StylianouPublished in: INTERSPEECH (2014)
Keyphrases
- signal to noise ratio
- speech recognition
- spectral features
- speech enhancement
- noisy environments
- frequency domain
- noise reduction
- random noise
- spectral analysis
- additive noise
- linear predictive coding
- image processing
- speech signal
- energy consumption
- spectral domain
- noisy data
- image noise
- image enhancement
- power spectrum
- linear prediction
- noise level
- background noise
- infrared
- missing data
- pattern recognition
- short time fourier transform
- speech quality
- hidden markov models
- image contrast
- vocal tract
- speech synthesis
- access patterns
- median filter
- energy minimization