Calculating FFT for discrete values

Click For Summary

Homework Help Overview

The discussion revolves around calculating the Fast Fourier Transform (FFT) for a discrete set of values defined as f(n) = (1, 3, 2). Participants are exploring the implications of using FFT versus Discrete Fourier Transform (DFT) for this small dataset.

Discussion Character

  • Exploratory, Conceptual clarification, Mathematical reasoning

Approaches and Questions Raised

  • The original poster attempts to calculate the DFT and is uncertain about the necessity of using FFT for only three points. Some participants discuss the symmetry properties of the FFT and how they might simplify calculations. Others question the equivalence of certain exponential terms in the context of periodicity.

Discussion Status

Participants are actively engaging with the concepts of FFT and DFT, with some providing insights into the mathematical properties involved. There is a recognition of the symmetry in the function and its implications for calculations, but no explicit consensus has been reached regarding the best approach for this specific problem.

Contextual Notes

There is a focus on the mathematical properties of the exponential functions involved in the FFT calculations, particularly regarding periodicity and symmetry. The discussion reflects a learning environment where assumptions about the necessity of FFT for small datasets are being questioned.

electronic engineer
Messages
145
Reaction score
3

Homework Statement


We have a set of values: f(n)=f(0,1,2)=(1,3,2)

so f(0)=1, f(1)=3, f(2)=2, where n=0,1,2the number of values N=3

The question is to calculate the FFT of this signal, the Fourier spectrum the power spectrum and phase spectrum.

I'm not sure concerning FFT. And also about the Fourier Spectrum what should I use to conduct it? the DFT or FFT?

Homework Equations


[/B]
The calculation of DFT for these values.

The Attempt at a Solution



I personally calculated the Discrete Fourier Transform in this way:

$$ F[n]=\sum_{k=0}^{N-1}f(k)e^{-j \frac{2\pi}{N}nk} $$

so I had a set of values: F(0),F(1),F(2)
 
Physics news on Phys.org
For only 3 points, it seems unnecessary to try to do the FFT. However, the point of the FFT is to take advantage of symmetry.
F(1) = f(0) +f(1) e^{-j 2pi/N * 1 }+f(2) e^{-j 2pi/N * 2 }
F(2) = f(0) +f(1) e^{-j 2pi/N * 2 }+f(2) e^{-j 2pi/N * 4 }= f(0) +f(1) e^{-j 2pi/N * 2 }+f(2) e^{-j 2pi/N * 1 }
So whay you really have is (for real-valued inputs):
## F(0) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\1\\1} , ##
## F(1) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\ e^{-j 2\pi/N }\\ e^{ j 2\pi/N }}##
## F(2) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\ e^{j 2\pi/N }\\ e^{- j 2\pi/N }} = F(1) ^* ##
As I said, for 3 values, this is more a demonstration of the concept that F(N-k) = F(k)*, which should save you about 1 calculation by just replacing F(1) with its conjugate for F(2).
 
RUber said:
For only 3 points, it seems unnecessary to try to do the FFT. However, the point of the FFT is to take advantage of symmetry.
F(1) = f(0) +f(1) e^{-j 2pi/N * 1 }+f(2) e^{-j 2pi/N * 2 }
F(2) = f(0) +f(1) e^{-j 2pi/N * 2 }+f(2) e^{-j 2pi/N * 4 }= f(0) +f(1) e^{-j 2pi/N * 2 }+f(2) e^{-j 2pi/N * 1 }
So whay you really have is (for real-valued inputs):
## F(0) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\1\\1} , ##
## F(1) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\ e^{-j 2\pi/N }\\ e^{ j 2\pi/N }}##
## F(2) = \pmatrix{ f(0), f(1), f(2) } \pmatrix{ 1\\ e^{j 2\pi/N }\\ e^{- j 2\pi/N }} = F(1) ^* ##
As I said, for 3 values, this is more a demonstration of the concept that F(N-k) = F(k)*, which should save you about 1 calculation by just replacing F(1) with its conjugate for F(2).

That is a very good and summarized answer. Thank you RUber :)
 
But I still don't understand how e^(-j*2*pi/N *2)= e^(j*2*pi/N). Are these two angles identical??
 
This is the symmetry of the function. ##e^{ix}## is 2pi periodic, so ##e^{ix}=e^{ix+i 2\pi}##
##e^{-j2\pi/N 2+j 2\pi}= e^{-j4\pi/3 +j6\pi/3}=e^{j*2*pi/N}.##
 
RUber said:
This is the symmetry of the function. ##e^{ix}## is 2pi periodic, so ##e^{ix}=e^{ix+i 2\pi}##
##e^{-j2\pi/N 2+j 2\pi}= e^{-j4\pi/3 +j6\pi/3}=e^{j*2*pi/N}.##

OK, thank you again.
 

Similar threads

  • · Replies 12 ·
Replies
12
Views
2K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 7 ·
Replies
7
Views
2K
Replies
26
Views
6K
Replies
4
Views
2K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 1 ·
Replies
1
Views
3K
Replies
6
Views
2K