Improved Window Function in the Application of MFCC Feature Parameter Extraction

In view of spectrum leakage and the contradictory problem of spectrum accuracy of main lobe and reducing spectrum leakage, MFCC algorithm based on improved window function is proposed. Improved window function is based on the mathematical analysis of Kaiser window, and under the condition of finite sampling points minuses weighted impact function where is at the frequencies that side lobe peaks of correspond to. The amplitude of improved window compared with Kaiser window is smaller, and main lobe width is the same, solving the conflicting problem of main lobe width and side lobe amplitude and reducing spectrum leakage. The experimental results show that speech recognition rate of MFCC feature parameter extraction algorithm based on improved window function is better than Kaiser window and Hamming window.