Speech Enhancement Using Fast Adaptive Kalman Filtering Algorithm Along With Weighting Filter

The speech signal which is degraded by noise is improved by the technique called speech enhancement. For this purpose we use many filters such as conventional, Fast Adaptive, weighted filter, etc. conventional Kalman filter require calculating the parameters of AR (auto-regressive) model, and inverse matrix operation, which is non-adaptive. In this paper we proposed adaptive Kalman filter along with perceptual weighting filter which is used to eliminate the matrix operations, to reduce the calculating time and complexity. Perceptual weighting filter is based on masking characteristics of human auditory system and constantly updating the first value of state vector Z (n) in adaptive filtering which automatically amend the estimation of environmental noise by observation data. The simulation results show that the performance is improved compared to Conventional method