Some considerations about 2D window functions used for bispectrum estimation
T. Oroian, Florin Şerban, Constantin-Iulian Vizitiu · 2008
It is well known that higher-order statistics is a new useful tool in signal processing area. In some specific applications of signal processing, they bring additional information about analyzed signal compared with the case of standard second-order statistics use. An important problem in this new domain is to estimate the cumulants and polyspectra when a finite sequence of time samples is available. In order to obtain a better estimate for bispectrum of the signal, different types of 2D window functions are used. Also, these windows are investigated in terms of their resolution and leakage effect.