查看文章

Thenormal trendin Linear Prediction (LP) techniques is fixed frame windowing. In this paper, however, dynamic window concept is introduced whereframe size is kept variable in order to achieve efficient outcome in terms of computational cost, mean square error and prediction gain. The three famous LP techniques namely Normal Equations, Levinson Durbin Algorithm (LDA) and Leroux Gueguen Algorithm (LGA) are briefly discusse dusing variable frame windowing and the above mentioned parameters are analyzed. Simulation results for the above three algorithms suggest that LDA and LGA have shown better performance than Normal Equation method based on reduced prediction error, low computational time and high prediction gain.