I am attempting to prove Lobachevsky's integral formula, which states that for an integrable even function $f$ with a period of π, the following integral equality must hold,$$\int_{0}^{\infty} \frac{\sin(x)}{x}f(x)dx = \int_{0}^{\pi/2}f(x)dx$$ I am aware of a classic proof,but I want to try proving this integral formula from the perspective of Fourier series expansion,Due to the function being even, we can let $$f(x)=\frac{1}{2}a_0+\sum_{n\geq 1}a_n\cos(2nx)$$, and then switch the order of summation and integration, thus obtaining $$\begin{aligned}\int_{0}^{\infty} \frac{\sin(x)}{x}f(x)dx &= \sum_{n\geq 1}a_n\int_{0}^{\infty} \frac{\sin(x)}{x}\cos(2nx)dx+\frac{\pi}{4}a_0 \\ &= a_0\frac{\pi}{4} =\frac{\pi}{4}\frac{4}{\pi}\int_{0}^{\pi/2}f(x)dx \\ &= \int_{0}^{\pi/2}f(x)dx\end{aligned}$$.
- However, there is a problem here. The Fourier series of a continuous function may not necessarily converge pointwise to the function itself. But Carleson's theorem tells us that the Fourier series of continuous functions converge almost everywhere. Therefore, if we consider the Lebesgue version of the integral formula, then this proof should be feasible.
- My main question is, if we consider the Riemann integral, is the proof method using Fourier series expansion feasible?