OAI/Analysis/LipschitzHilbert/Fourier/PublishedBandReductionInput.lean

GitHub source file

1 declarations, 0 contain sorry, 0 declared as axioms

The precise published band-to-full input

Di Plinio, Guo, Thiele and Zorin-Kranich, Square functions for bi-Lipschitz maps and directional operators, J. Funct. Anal. 275 (2018), 2015--2058, DOI 10.1016/j.jfa.2018.07.005, author version arXiv:1706.07111, Corollary 1.4, p₀=2 (printed p. 2), following Corollary 1.3.

We use only its conclusion for the fixed real, even, nonnegative and nonzero narrow cutoff. Its frequency support is ±[1,101/100]. This is the inverse Fourier transform of the paper's Ψ, up to the harmless scalar normalization in the reproducing formula. We require the hypothesis for all positive widths, which is stronger than the paper's requirement 0<t<t₀.

The proposition below records the rescaled specialization at R<=1/10000: set (x,y)=R(X,Y). The resulting field is u(RX,RY), whose vertical Lipschitz constant is at most R, hence at most 1/100. The width changes from w to w/R and the same two-dimensional L2 scaling occurs on both sides. Our sampling uses a minus sign instead of the paper's plus sign; the symmetric principal values differ by an overall minus sign.

  1. def OAI.LipschitzHilbert.PacketRepresentation.BandToFullInput line 35