Résumé:
FIR Filters are used, nowadays, in different applications because of their phase linearity as well as their intrinsic stability. This research introduces an innovative window-based near optimal FIR Filters design, especially for high order FIR Filters (complex mathematical calculations). And showcasing its effectiveness in enhancing speech signal quality by minimizing noise. In this research we try to rely on an innovative window called the Fractional order window whose characteristics depend on a parameter α. The fractional window function can adjust the spectral characteristics of FIR Filters, i.e., Ripple ratio, Main Lobe Width, and Side Lobes Ripple ratio by using two parameters: order ‘N’ and coefficient ‘α’.
Order of window N! Yes, A characteristic of this window is that its main lobe and the amplitude of its side lobes also depend on the number of samples N, the Ripple ratio can be reduced to the lowest levels, something which does not exist for conventional windows. To performance appraisal of Fractional window in speech signals de-noising, FIR filter has been designed using Fractional window and compared with filter designed using Kaiser window. It has been shown that proposed window gives far better performance than Kaiser window function. The enhanced performance and computational efficiency of FIR filters with these innovate window make them particularly suitable for real-time applications. This research provides valuable insights into the optimization of FIR filters, laying the groundwork for future advancements in adjustable and hybrid filtering methods that promise to further refine speech signal processing and improve audio technologies.
