Skip to content

DFS

Helpful Resource (DTFT, DFS, DFT)

https://ethz.ch/content/dam/ethz/special-interest/mavt/dynamic-systems-n-control/idsc-dam/Lectures/Signals-and-Systems/Lectures/Fall2018/SigSys_Lect5.pdf


The Basics

DFT/DFS Representation of a Periodic Signal

The Discrete Fourier Series perfectly represents a periodic sequence \(x[n]\) as a series (sum) of sinusoids (or Complex Sinusoids).
- Below is the complex forms of the Synthesis (Discrete Frequency to Discrete Time), and Analysis (Discrete Time to Discrete Frequency)
\(\(\begin{array}{}\LARGE x[n]=\frac{1}{N}\sum\limits_{k=0}^{N-1}X[k]e^{j2\pi \frac{k}{N}n}\\\text{Synthesis Equation, IDFS}\\\LARGE X[k]=\sum\limits_{n=0}^{N-1}x[n]e^{-jk \frac{2\pi}{N}n}\\\text{Analysis Equation, DFS}\end{array}\)\)
- Since \(x[n]\) is periodic, so is the DFS, with period N.
- The lowest frequency sinusoid of the DFS is \(\large \frac{2\pi}{N}\), aka the fundamental frequency
- We need N complex exponentials to represent a DT periodic signal with period N.
- Aka Discrete-Time Fourier Series (DTFS)

Example Coefficient Calculation (Analyzing Frequency Content of a Sequence)

Given the periodic sequence, \(x[n]=\{...,2,0,4,-5,2,0,4,-5,2,0,4,-5,...\}\), and that the first -5 is n=0 find the DFS coefficients \(X[k]\).
(Alternative prompt for DFT: Given a sequence defined below, find the DFT coefficients for the sequence)
\(\(x[n]=\Bigl\{ \begin{array}{l}{-5,2,0,4},\; \text{ for } 0\le n <N \\0,\; otherwise\end{array}\)\)
1. The pattern repeats every 4 n, so N=4.
2. DFS Coefficients Definition: \(\large X[k]=\sum\limits_{n=0}^{N-1}x[n]e^{-jk \frac{2\pi}{N}n}\)
3. Apply N=4: \(\large X[k]=\sum\limits_{n=0}^{3}x[n]e^{-jk \frac{2\pi}{4}n}\)

  1. Evaluate Sum: \(\large =x[0]e^{-jk \frac{2\pi}{4}0} + x[1]e^{-jk \frac{2\pi}{4}1} + x[2]e^{-jk \frac{2\pi}{4}2} + x[3]e^{-jk \frac{2\pi}{4}3}\)

  2. Simplify: \(=\large -5(1) + 2e^{-jk \frac{\pi}{2}} + 0e^{-jk\pi} + 4e^{-jk \frac{3\pi}{2}}\)

    1. \(\large X[k]=-5 + 2e^{-jk \frac{\pi}{2}} + 4e^{-jk \frac{3\pi}{2}}\)
  3. Find \(\{X[k]\}\):

    1. \(\large X[0]=-5 + 2e^{-j0 \frac{\pi}{2}} + 4e^{-j0 \frac{3\pi}{2}} = -5+2+4=1\)
    2. \(\large X[1]=-5 + 2e^{-j\frac{\pi}{2}} + 4e^{-j\frac{3\pi}{2}} = -5+2(-j)+4(+j)=-5+2j\)
    3. \(\large X[2]=-5 + 2e^{-j2\frac{\pi}{2}} + 4e^{-j2\frac{3\pi}{2}} = -5+2(-1)+4(-1)=-11\)
    4. \(\large X[3]=-5 + 2e^{-j3\frac{\pi}{2}} + 4e^{-j3\frac{3\pi}{2}} = -5+2(+j)+4(-j)=-5-2j\)
  4. Answer: \(X[k]=\{1,-5+2j,-11,-5-2j\}\)

But what do the coefficients mean? (Synthesizing a signal from it's frequency content)
  • Let's use the previous example's coefficients: \(X[k]=\{1,-5+2j,-11,-5-2j\}\)
  • The Synthesis equation: \(\(\LARGE x[n]=\frac{1}{N}\sum\limits_{k=0}^{N-1}X[k]e^{j2\pi \frac{k}{N}n}\)\)
  • If we manually build the sinusoid form of those coefficients using Eulers Formula, we'll find that all of the imaginary parts of the our equation will either cancel out or be zero-valued for integer values of n, leaving us with just the real portions at our sampling points:
    \(\(\large x[n]=1-5cos(2\pi \frac{n}{4})-11cos(2\pi \frac{n}{2})-5cos(2\pi \frac{3n}{4})-2sin(2\pi \frac{n}{4})+2sin(2\pi \frac{3n}{4})\)\)
  • Here is the graphical representation of the added real series (lined) and the original sequence (circles). Keep in mind that we could have also graphed the complex magnitude of the DFT, or graphed the imaginary part on it's own.
    1200
    clear all
    x=[-5 2 0 4];
    m=0:3;
    n=0:0.1:3;
    y=1-5*cos(2*pi*n/4)+-11*cos(2*pi*n/2)+-5*cos(2*pi*3*n/4)-2*sin(2*pi*n/4)+2*sin(2*pi*3*n/4);
    plot(m,x,'o',n,y/4,'x');
    title("Real Fourier Series vs Real Periodic Sequence")
    xlabel("Discrete Time, n");
    ylabel("Amplitude")
    ylim([-5.5,4.5]);
    xlim([-0.5,3.5]);
    

What we think, we become.
Buddha