A density criterion for frames of complex exponentials (Q1191397): Difference between revisions

A sequence \((e_ n)\) of vectors of a Hilbert space \(H\) is a frame if there exist positive constants \(C_ 1\) and \(C_ 2\) such that, for all \(f\) in \(H\), \[ C_ 1\| f\|^ 2\leq\sum|\langle f,e_ n\rangle|^ 2\leq C_ 2\| f\|^ 2. \] Let \(\Lambda=(\lambda_ n)_{n\in Z}\) be a sequence of distinct real numbers. The upper bound \(R(\Lambda)\) of all numbers \(R\) such that the sequence of functions \((e^{i\lambda_ n t})\) is a frame of \(L^ 2([-R,R])\) is called the frame radius of the sequence. Let \(U(\Lambda)\) be the set of all the subsequences of \(\Lambda\) with a uniform density \(d\). Then the frame density of \(\Lambda\) is defined by \(D^ f(\Lambda)=\sup_{\Theta\in U(\Lambda)} d(\Theta)\). In this paper, the following result is proved: If \(U(\Lambda)=\emptyset\), or if the numbers \(A_ n\) of elements of \(\Lambda\cap[n,n+1]\) are not bounded (\(n\) taking all integer values), then there exists no interval over which \((e^{i\lambda_ n t})\) is a frame. Otherwise, the frame radius of \(\Lambda\) is equal to \(\pi D^ f(\Lambda)\).