Generalisation of Parseval's Theorem via Convolution Theorem

In summary, the conversation discusses Parseval's theorem, which states that the sum of the squared Fourier coefficients of a 2\pi-periodic, integrable function is equal to the integral of its squared function. It then poses the question of finding the Fourier coefficients of the q-th power of a function, and whether repeated application of the convolution theorem is the usual method for this. The provided solution suggests that the manipulation is correct as long as everything converges, and that the convolution should be on \mathbb{Z} rather than \mathbb{R}.
  • #1
FeDeX_LaTeX
Gold Member
437
13

Homework Statement


[/B]
Suppose we have a [itex]2\pi[/itex]-periodic, integrable function [itex]f: \mathbb{R} \rightarrow \mathbb{C}[/itex] whose Fourier coefficients are known. Parseval's theorem tells us that:

[tex]\sum_{n = -\infty}^{\infty}|\widehat{f(n)}|^2 = \frac{1}{2\pi}\int_{-\pi}^{\pi}|f(x)|^{2}dx,[/tex]

where [itex]\widehat{f(n)}[/itex] are the Fourier coefficients of [itex]f[/itex].

Suppose we instead want to replace [itex]f(x)[/itex] with [itex]f(x)^{q},[/itex] say: then it would suffice to determine the Fourier coefficients of the [itex]q[/itex]-th power of [itex]f[/itex]. Is repeated application of the convolution theorem the usual way of finding powers of the Fourier coefficients of functions, where the Fourier coefficients of the original function are already known?

Homework Equations


[/B]
[itex]f \ast g[/itex] denotes the convolution of [itex]f[/itex] and [itex]g[/itex], given by [itex](f \ast g)(t) := \int_{-\infty}^{\infty} f(\tau)g(t - \tau)d\tau,[/itex] and [itex]\widehat{f \ast g} = \hat{f} \cdot \hat{g}[/itex] is the convolution theorem for the Fourier transforms of [itex]f[/itex] and [itex]g[/itex].

The Attempt at a Solution


Suppose that we are interested in [itex]\int_{-\pi}^{\pi}|f(x)|^{4} dx[/itex]. I would like to know if it is valid to say the following:

[tex]\frac{1}{2\pi}\int_{-\pi}^{\pi}|f(x)|^{4}dx = \frac{1}{2\pi}\int_{-\pi}^{\pi}|(f(x))^{2}|^{2}dx = \sum_{n = -\infty}^{\infty} |\widehat{f(n)^{2}}|^{2} = \sum_{n = -\infty}^{\infty} | (\hat{f} \ast \hat{f})(n)|^{2}.[/tex]

The reason I am interested in this is because I'm working on bounding a class of [itex]L^{p}[/itex]-norms using the asymptotics of Fourier coefficients, and hoping to modify this slightly to integrate functions over a [itex]d[/itex]-cube [itex][0,2\pi)^d[/itex]. This seems to be a complicated procedure however, since additional conditions need to be imposed on the functions to guarantee the convergence of the integral in [itex]\mathbb{R}^d[/itex].
 
Last edited:
  • #3
Thanks for the automated response. I posted the same question on MathSE, and was told that the manipulation is correct as long as everything converges, although one should note that the convolution is on [itex]\mathbb{Z}[/itex] rather than on [itex]\mathbb{R}[/itex]. The only question I have left is whether that means I should be using the discrete convolution, or if I should still be using the continuous convolution restricted to [itex]\mathbb{Z}[/itex] (if such a thing exists).

EDIT: I've just realized that the discrete convolution is precisely the continuous convolution restricted to the integers, so the question for this post has been answered. However, if anyone has any comments about the manipulation in post #1, I would greatly appreciate hearing them.
 
Last edited:

FAQ: Generalisation of Parseval's Theorem via Convolution Theorem

What is Parseval's Theorem?

Parseval's Theorem is a mathematical theorem that relates the energy of a signal in the time domain to its frequency domain representation. It states that the sum of the squared magnitudes of a signal in the time domain is equal to the sum of the squared magnitudes of its Fourier transform in the frequency domain.

What is the Convolution Theorem?

The Convolution Theorem is a mathematical theorem that states that the convolution of two signals in the time domain is equivalent to the multiplication of their Fourier transforms in the frequency domain. In other words, it provides a shortcut method for calculating the convolution of two signals.

How does the Convolution Theorem relate to Parseval's Theorem?

The Convolution Theorem can be used to generalize Parseval's Theorem. By applying the Convolution Theorem to the squared magnitudes of a signal and its Fourier transform, we can show that Parseval's Theorem holds for a wider range of signals, including complex-valued signals and signals with finite energy.

Why is the generalization of Parseval's Theorem important?

The generalization of Parseval's Theorem allows us to apply this important theorem to a wider range of signals, making it a more useful tool in various fields such as signal processing, communication systems, and image processing. It also provides a deeper understanding of the relationship between a signal and its frequency domain representation.

What are the applications of the Convolution Theorem and Parseval's Theorem?

The Convolution Theorem and Parseval's Theorem have numerous applications in areas such as signal processing, communication systems, image processing, and quantum mechanics. They are used to analyze and manipulate signals, design filters, and solve differential equations. They are also used in the study of Fourier series and transforms, and in the development of efficient algorithms for signal processing.

Similar threads

Back
Top