$L^p$ version of Square Function

110 Views Asked by At

I have taken an introductory course into Harmonic analysis and in the course, we went over square functions

$s_\phi f(x) =\left(\int_0^\infty |f *\phi_t (x)|^2 \frac{dt}{t}\right)^{1/2}$

where $\phi \in \mathcal{S}$ is Schwartz with average zero: $\int \phi = 0$.

Essentially by Plancharel's theorem, we get $\| s_\phi(f)\|_{L^2}^2 \lesssim \| f\|_{L^2}^2$ and then one can show this also holds with regards to $L^p$ norm.

I am curious if this result can be generalized in the following way: instead of defining the square function with regards to an $L^2$ convolution integral, can we do it against an $L^p$ convolution integral instead? In other words, what happens if we look at

$$\left(\int_0^\infty |f * \phi_t(x)|^p \, \frac{dt}{t}\right)^{1/p}$$

for $p>1$? Can we establish a similar estimate on the "p-square" function?