Suppression of Two Noise Sources in the Same Direction by Instantaneous Spectral Subtraction.
Kenji OzawaKanji WatanabeShuichi SakamotoPublished in: GCCE (2019)
Keyphrases
- spectral subtraction
- speech recognition
- speech signal
- signal to noise ratio
- noise model
- hybrid method
- power spectrum
- automatic speech recognition
- speech enhancement
- hidden markov models
- background noise
- noisy environments
- additive noise
- noise reduction
- noisy speech
- multi band
- non stationary
- hearing aids
- noise level
- edge detection
- linear prediction
- sound source
- vocal tract
- frequency band
- multiresolution