Real-Time Speech Enhancement Using Spectral Subtraction with Minimum Statistics and Spectral Floor.
Georgios IoannidesVasilios RallisPublished in: CoRR (2023)
Keyphrases
- spectral subtraction
- speech enhancement
- speech signal
- speech recognition
- noisy environments
- hybrid method
- signal to noise ratio
- noise reduction
- linear prediction
- single channel
- power spectrum
- noisy speech
- automatic speech recognition
- vocal tract
- noise model
- background noise
- spectral analysis
- hearing aids
- non stationary
- language model
- hidden markov models
- edge preserving
- wiener filter
- neural network
- multi channel
- image quality
- additive noise
- speech synthesis
- image coding
- information retrieval