Design and Performance Analysis of FIR Filtering Operations on Speech Signal Using Windowing Techniques

Abstract

One of the important applications of digital signal processing is digital filter which is used in speech processing applications such as speech filtering, speech enhancement, noise reduction etc. In this project different windowing methods are used for designing the finite impulse response filter. A window function is a mathematical function that is zero esteemed outside of some chosen interval. The objective is to analyze the spectrum of input speech signal and to design and implement low pass and high pass FIR filters using different windowing methods. Hamming, Hanning, Blackman, Rectangular and Kaiser windows are used for designing FIR filters and responses of the FIR filters are observed. Finally performances of these windows are evaluated based on equivalent noise bandwidth and windowing power. Comparing the results of different windows, it is observed that the rectangular window has low equivalent noise bandwidth and high windowing power. This shows that performance of the rectangular window is better and it is suitable for designing low pass and high pass FIR filters.