On the uncertainty principle for positive definite densities (Q2365278)

Let \(p\) be a positive probability density and \(f\) be its characteristic function. The adjoint density of \(p\) is denoted by \(\widetilde p\) and is defined by \(\widetilde p(x)= (\int^{+ \infty}_{-\infty} f(x) dx)^{-1}f(x)\), \(x\in \mathbb{R}\). Denote by \(\sigma^2\) and \(\widetilde \sigma^2 \) the variances of \(p\) and \(\widetilde p\), respectively, and by \(\Lambda\) the greatest lower bound for all products of variances \(\sigma^2\widetilde\sigma^2\) of adjoint densities. It is known that \(0.454< \Lambda< 0.8907\). The author improves the bounds as \(0.527<\Lambda \leq 0.8609 \dots\).