Title: A demonstration of a single channel blind noise reduction algorithm with live recordings
Abstract: Currently, most noise reduction algorithms are based on an a priori information such as signal-to-noise ratio (SNR) or noise parameters estimation. They are mostly performed in the spectral domain to reduce the background noise at each frequency bin. However noise reduction in the spectral domain may introduce musical noise and artefacts which are in some cases perceptually more annoying than the background noise itself. In this “show and tell”, we present a demonstration of a noise reduction algorithm based on dynamic range compression (DRC) using a timevarying and frequency-band dependant gain function deduced from the low-pass filtering of the temporal envelopes. The algorithm is considered as blind since it requires neither SNR nor noise parameters estimation. A graphical user interface (GUI) built under Matlab shows interactively the noise reduction in the temporal (waveform) and spectral (spectrogram) domains using live speech recordings mixed to pre-recorded noise signals.
Publication Year: 2014
Publication Date: 2014-01-01
Language: en
Type: article
Access and Citation
Cited By Count: 2
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot