Filtering Techniques to reduce Noise from Corrupted Audio Signal

The aim of the project is to process the audio signal contain the corrupted sound source to reduce the noise utilizing various filter and processing techniques in MATLAB. In this task we are using the FFT for filtering, median filter, lowpass filter, bandpass filter, Equiripple Lowpass filter and Butterworth lowpass filter to reduce the noise from the audio signal. Background noise can substantially damage the quality of intended signal therefore, communications and signal processing devices are likely to function in challenging circumstances. Noise reduction techniques are created and applied to noisy signals with the goal of increasing signal quality and reducing background noise. Low-pass filtering is a technique often applied in preparing voice recordings for acoustic analysis. The presence of external noise in a speech signal has been found to inflate values of jitter, shimmer, and correlation dimension, among other metrics, affecting acoustic measurements of perturbation and nonlinear dynamic voice features. Due to unnaturally high noise levels, environmental noise poses a risk of false-positive diagnoses of vocal fold disorders. Because the components of vocal analysis are located at low frequencies, lowering the high frequency component of environmental noise can help improve vocal analysis outcomes. The spectrum subtraction approach is a well-known noise reduction technique. The noisy audio signal is first converted from the time domain to the frequency domain using the quick Fourier transform in this method. The noise spectrum is then computed in the audio pauses and subtracted from the noisy audio signal's frequency spectrum before using the inverse FFT to transform the noisy audio signal from the frequency domain to the time domain.

Theory