One-sided Littlewood-Paley inequality in \(\mathbb R^n\) for \(0<p\leq 2\) (Q549432)

The following theorem is proved. Suppose that \(\{f_m\}\) is a sequence of functions in \(L^1(\mathbb{R}^n)\) whose Fourier transforms \(\hat{f}_m\) are supported in pairwise disjoint parallelepipeds \(\Delta_m\) having sides parallel to the coordinate axes. Then for \(0<p<2\) one has \[ \bigg\|\sum_m f_m\bigg\|_{L^p(\mathbb{R}^n)}\leq C_{p,n} \|\{f_m\}\|_{L^p(\mathbb{R}^n,\ell^2)}, \] where \(C_{p,n}\) does not depend on the functions or on the parallelepipeds. The proof proceeds by a reduction to a corresponding boundedness result of the mapping from the multiparameter real Hardy space \(H^p(\mathbb{R}\times\cdots\times\mathbb{R})\) to the vector-valued \(L^p\)-space. The theorem is reduced to a special case in which there are only finitely many parallelepipeds that are all obtained from dyadic ones by an 8-fold dilation, that is, there exist \((k,j)=(k_1,\dots, k_n, j_1,\dots, j_n)\in\mathbb{Z}^{2n}\) such that \[ \Delta_m=[j_12^{k_1},(j_1+8)2^{k_1}]\times\cdots\times [j_n2^{k_n},(j_n+8)2^{k_n}] \] and that \(\hat{f}_m\) is supported in \((3/4) \Delta_m\). In this case, it is shown to be enough to prove \(H^p(\mathbb{R}^n,\ell^2)\) to \(L^p(\mathbb{R}^n)\) boundedness of the operator \[ S(h) (x_1,x_2,\dots, x_n) = \sum_{(k,j)} e^{2\pi i j_1 2^{k_1}x_1}\cdots e^{2\pi i j_n 2^{k_n}x_n}\, (\Phi_k\ast h_{k,j})(x_1,\dots, x_n), \] where the sum extends over the vertex multi-indices \((k,j)\) of the \(\Delta_m\), \(h=\{h_{k,j}\}_{(k,j)\in\mathbb{Z}^{2n}}\in L^2(\mathbb{R}^n,\ell^2)\), and \(\Phi=\phi_{k_1}(x_1)\cdots \phi_{k_n}(x_n)\) with \(\varphi_k (t) =2^k\varphi(2^k t)\). The bound again depends only on \(p\) and \(n\). The proof of this boundedness reduces to a bound of \textit{A. Carbery} and \textit{A. Seeger} [Trans. Am. Math. Soc. 334, No. 2, 719--747 (1992; Zbl 0770.42010)] that states that an \(L^2\)-bounded operator that satisfies an appropriate \(L^p\)-bound for multiparameter atoms extends to an operator bounded from \(H^p\) to \(L^p\).