In This paper Introduces a new technique for speech enhancement to enhance speech degraded by noise corrupted signal. When speech signal with an additive Gaussian white noise is the only information available processing. So many techniques are to improve the efficiency for enhancement. Kalman filter is one of the technique for speech enhancement. But in kalman filter needs to calculate the parameters of Auto-Regressive model, For this a lot of matrix operations required. In which speech signal is usually modeled as autoregressive (AR) process and represented in the state-space domain. In this paper presents a alternative solution for estimate the speech signal. In proposed technique to eliminates the matrix operations and calculating time by only constantly updating the first value of state vector X(n). The experiments results show that the Improved algorithm for adaptive Kalman filtering is effective for speech enhancement
展开▼
