On the rate of pointwise divergence of Fourier and wavelet series in \(L^p\) (Q818461)

After \textit{P. du Bois-Reymond}'s construction of a continuous function whose Fourier series diverges at one point [Gött. Nachr. 1873.21, 571--582 (1873; JFM 05.0145.01)], Haar introduced an orthonormal basis of \(L^2\) in which the expansion of continuous functions converges uniformly on compact sets. In further developments of the theory of Fourier series, \textit{J.-P. Kahane} and \textit{Y. Katznelson} [Studia Math.\ 26, 305--306 (1966; Zbl 0143.28901)] proved that, given any \(F_{\delta}\) set \(A \subset T := R/Z\) of Lebesgue measure zero, there exists a continuous function whose Fourier series diverges everywhere on \(A\); meanwhile the Haar basis becomes the prototype of wavelet bases, and wavelet expansions of continuous functions also converge uniformly on compact sets, cf. [\textit{G. G. Walter}, J.\ Approximation Theory 80, No. 1, 108--118 (1995; Zbl 0821.42019)]. Here we see one of the main differences between Fourier and wavelet bases. Another difference is that wavelets yield unconditional bases of \(L^p\), \(1 < p < \infty\); this is false for the Fourier basis if \(p \neq 2\). In this paper a comparison is made between wavelet and Fourier bases on the matter of pointwise divergence for functions in \(L^p\). Since the works of \textit{L. Carleson} [Acta Math. 116, 135--137 (1966; Zbl 0144.06402)] and \textit{R. A. Hunt} [``On the convergence of Fourier series'' in: Orthogonal Expansions and their Continuous Analogues, Proceeding of the Conference, Southern Illinois University Press 1967, 235--255 (1968; Zbl 0159.35701)], we know that the Fourier series of a function in \(L^p(T)\), \(1\leq p \leq \infty\), converges almost everywhere. The goal of this paper is to study the rate of divergence at other points. It is shown that the set of points where the divergence is ``fast'' must be ``small'' in the sense of Hausdorff dimension. On the other hand, the wavelet expansion of a function in \(L^p\) also converges almost everywhere if \(1\leq p \leq \infty\), cf. [\textit{S. E. Kelly, M. A. Kon} and \textit{L. A. Raphael}, J.\ Funct.\ Anal.\ 126, 102--138 (1994; Zbl 0809.42019)]. The rate of divergence turns out to have a limitation equivalent to the one found for Fourier series. In both cases, the author shows that his bounds are sharp by constructing a function whose Fourier or wavelet series diverges pointwise at a given rate on any set satisfying the dimension condition.