IMPLEMENTATION OF SPEECH DENOISING USING SPECTRAL SUBTRACTION AND KALMAN FILTERING

In this paper, I havedescribed about the speech denoising using spectral subtraction and kalman filtering. Speech denoising aims to improve the speech quality and intelligibilityby using various techniques and algorithms. Speech communication is the process to have noise unrestricted speech signal with good clarity and qualityalong with high performance. In real world, it is very difficult to get clean signal all time for speech communication system. Speech processing is required in communication systems to apply noise reduction techniques so as to get desired speech signal from the corrupted speech signal. The proposed method, spectral subtraction and kalman filtering algorithms uses to suppress the additive background noise. It is easy for implementation and widely used in simplicity. Denoising is used to estimate average noise spectrum for noisy speech signal spectrum.The performance of spectral subtraction and kalman filtering algorithm for speech is evaluated by using Mean square error, Peak Signal to Noise Ratio, Normalized Root mean square error and Distortion. In this method Simulink model and hardware implementation in raspberry pi has been used.