An Adjustable Window-Based FIR Filter and Its Application in Audio Signal De-Noising

This paper presents a new window function which is adjustable as well as it consists of linear phase response indication. Using a simple parameter, r, the shape of the window can be regulated. Besides, using the proposed window, a low-pass FIR filter has been designed, and its behavior has been compared with Gaussian, Kaiser, Tukey window and another window proposed by Rakshit et al. The outcome implies that the proposed window-based filter is showing 26.31 dB higher side-lobe roll-off ratio than the Gaussian filter. In addition, the proposed window-based filter has 26.81 dB better side-lobe roll-off ratio over Kaiser filter and 10.87 dB improved side-lobe roll-off ratio than the Tukey filter. Furthermore, all the mentioned filters have been applied on a noisy voice signal and it has been observed which filter has been de-noising the noisy voice data in a better way where the result illustrates that the proposed window-based filter is showing 4.4375 better signal to noise ratio than gaussian filter as well as 1.66, 1.3741 and 3.4447 higher signal to noise ratio over Tukey, Kaiser filter and another proposed window based filter respectively.