DFT for convolution like operation.

In summary, the discussion involves using a cyclic convolution to calculate z in NLOG N time instead of N^2. The upper limit of the sum in the algorithm can be arbitrarily cut off at g without affecting the overall complexity. The solution involves taking the FFT of the re-defined x and y sequences, multiplying them point-wise, and then taking the inverse FFT of the result. However, there is a potential issue with the second sum depending on k, which prevents the double sum from factoring into the product of DFTs.
  • #1
menahemkrief
5
0
Let x,y be finite real valued sequences defined on 0...N-1 and let g be a non negative integer .

define
View attachment 2231
also on 0..N-1.

In addition, the DFT of y is known in closed form.
Is there a way to write z as some cyclic convolution, so that with the help of the convolution theorem z can be calculated in NLOG N istead of N^2?

Thank you
 

Attachments

  • Untitled.png
    Untitled.png
    1.7 KB · Views: 67
Mathematics news on Phys.org
  • #2
It would seem you can do that. First note that the unusual upper limit of your sum, arbitrarily cut off at $g$, poses no issue for this algorithm. Simply re-define the $x$ and $y$ sequences to include only the terms that show up in the convolution sum. I'm not sure that knowing the FFT of $y$ in advance will help you, since you really need to truncate the sequence, at least in general.

1. Take the FFT's of the re-defined $x$ and $y$. Complexity is $\mathcal{O}(n \, \log(n))$.
2. Multiply the two transformed sequences point-wise. Complexity is $\mathcal{O}(n)$.
3. Take the inverse FFT of the result. Complexity is $\mathcal{O}(n \, \log(n))$.
 
  • #3
Hi,

lets assume there is no g (g=inf).

notice the the sum end at the index n, not N.

I'm trying to see directly the convolution theorem but I get stuck:
View attachment 2233

The problem is that the second sum depends on k so the double sum doesn't factor to the product of DFTs.

what am I missing?

thank you
 

Attachments

  • Untitled.png
    Untitled.png
    5.5 KB · Views: 64

FAQ: DFT for convolution like operation.

What is DFT for convolution like operation?

DFT stands for Discrete Fourier Transform, which is a mathematical operation that converts a discrete signal from its original domain (often time or space) to a representation in the frequency domain. This operation is often used in signal processing and image analysis to analyze the frequency components of a signal or image.

How does DFT for convolution like operation work?

The DFT for convolution like operation involves multiplying the Fourier transform of two signals and then taking the inverse Fourier transform of the result. This process is essentially a convolution operation in the frequency domain, which allows for efficient computation of convolution in the time or space domain.

What are the benefits of using DFT for convolution like operation?

Using DFT for convolution like operation allows for more efficient computation of convolution compared to direct convolution in the time or space domain. It also allows for analysis of the frequency components of a signal, which can be useful in various applications such as filtering and noise reduction.

What types of signals or data can DFT for convolution like operation be applied to?

DFT for convolution like operation can be applied to any type of discrete signal or data, including audio signals, images, and time series data. It is commonly used in digital signal processing and image analysis applications.

Are there any limitations or drawbacks to using DFT for convolution like operation?

One limitation of using DFT for convolution like operation is that it assumes the signal is periodic, which may not always be the case in real-world data. Additionally, the computation can become more complex and less efficient for larger signals or data sets. Other techniques, such as Fast Fourier Transform (FFT), have been developed to address these limitations.

Similar threads

Back
Top