In the beginning of the proof of Kovrijkine’s uncertainty principle , we consider the function \(g(x)=f(ax)\) which scales \(a\) to 1.
Indeed, dropping this step does not change the result.
Case \(p\neq \infty\) Link to heading
We repeat the proof step by step without the first step:
1. Apply complex lemma on intervals to get an estimate where some supremum is involved Link to heading
Let \(I\) be an interval of length \(a\). We may apply (0x6698f0dc) on \(\phi=\frac{a^{1/p}g}{\lVert f\rVert_{L^p(I)}}\) since \(f\) is analytic on the whole real line and due to (0x669a1456) a point \(x_0\in I\) exists and such that \(f(x_0)\ge\frac{\lVert f\rVert_{L^p(I)}}{a^{1/p}}\). Thus we obtain
\begin{equation*} \lVert f\rVert_{L^p(I)} \le \biggl(\frac{24a}{|S\cap I|}\biggr)^{2 \frac{\log M}{\log 2} + \frac{1}{p}} \lVert f\rVert_{L^p(S\cap I)}, \end{equation*}where \(M=\max_{|z-z_0|\le 4} a^{1/p}\frac{|f(z)|}{\lVert f\rVert_{L^p(I)}}\). Note that the scaling factor disappears in the inequality.
2. Estimate of the supremum on good intervals Link to heading
Using Kovrijkine’s Lemma we estimate \(M\) on each good interval \(I\) with
\begin{align} M &= \sup_{z\in D_{4a}(x_0)}\frac{a^{1/p}|f(z)|}{\lVert f\rVert_{L^p(I)}} \\ &= \sup_{z\in D_{5a}(x)}\frac{a^{1/p}|f(z)|}{\lVert f\rVert_{L^p(I)}} \\ &\le 2^{1/p}\exp(5Cab), \end{align}where \(x\) is the expansion point mentioned in the lemma. Note, that the \(a\) appears in exponential, because \(R=5a\).
3. Combine results of step 2 and 3 to obtain the result. Link to heading
Combining both, we obtain
\begin{align} \lVert f\rVert_{L^p(S)} &\ge \sum_{\text{\(I\) is good}} \lVert f\rVert_{L^p(S\cap I)} \\ &\ge \sum_{\text{\(I\) is good}} \biggl(\frac{|S\cap I|}{24a}\biggr)^{C(ab + 1/p)} \lVert f\rVert_{L^p(I)}, \\ &\ge \sum_{\text{\(I\) is good}} \biggl(\frac{\gamma}{24}\biggr)^{C(ab + 1/p)} \lVert f\rVert_{L^p(I)}, \\ &\ge 2^{-1/p}\biggl(\frac{\gamma}{24}\biggr)^{C(ab + 1/p)} \lVert f\rVert_{L^p(\mathbb{R})}, \end{align}where we also used the thickness property of \(S\) and the fact, that the family of good intervals is sufficiently large. Note, that \(C>0\) is some numerical constant.