Sidelobe Reduction for Window Functions by Multiplication of Cosine-Sum Windows
Tomoya Yamaoka, Tadashi Oshima · IEEE Access · 2025
Various studies have considered the reduction in sidelobes when using window functions, and further sidelobe reduction using existing design results is an important perspective. In this study, we propose reducing side lobes by multiplying cosine-sum windows with existing window functions. This process shifts and synthesizes the frequency profile of the window function and may effectively reduce sidelobes. This is advantageous in that unlike the convolution between window functions, the data length is short. The proposed method was first applied to Gaussian windows, and it was found that they exhibit better performance than that with Gaussian windows alone and approximate confined Gaussian windows. We also showed that the peak and integrated sidelobe ratios of the proposed method are superior to those of cosine-sum windows designed with a single parameter. Moreover, we demonstrated that the proposed method can control the sidelobe decay of a Gaussian window. The second application of the proposed method was to the Nuttall, Blackman-Harris, and Blackman-Nuttall windows. We clarified the derived properties, showing that their lower sidelobes could be reduced further.