Expt.

No :1 Date:30-07-2024

Implementation of DFT and IDFT

Aim
1. To find frequency domain representation of a discrete-time signal using DFT.
2. To find the time domain sequence from the spectrum of the signal using IDFT

Theory
Frequency content in a continuous-time signal can be obtained using continuous-
time Fouriertransform(CTFT).
Let, x(t) be a continuous-time signal and its CTFT is given by

j t
X(j ) x(t)e dt (1)

where and t are continuous variables.

Let, x(n) be a discrete-time signal and its discrete-time Fourier transform(DTFT) is
given by

j j n
X (e x(t)e (2)

where is continuous variable and time n is discrete .

Discrete Fourier Transform (DFT)

The discrete Fourier transform(DFT) of a length-N time-domain seqeuence x(n) is
given by
N 1
X [k] 0 k N 1 (3)
n0

where
N1
j j 2 nk / N
X [k] X (e ) | 2 nk / N x[n]e 0 k N 1 (4)
n 0

Comparing Eq.(2) and Eq.(4) it is observed that the N-point DFT sequence X(k) is
precisely the set of frequency samples of the Fourier transform X(e ) of the length-N
sequence x(n) at N equally spaced frequencies k = 2 /N, 0 K N 1. Hence Eq.(3) is the
frequency domain representation of the sequence x(n).

1
 Algorithm
1. Fix the N-point DFT of the sequence
2. Fix the time index n and frequency index k to run from 0 to N-1.
3. For each value of n allow k to run from 0 to N-1 and sum the product given by
Eq.(3).
4. Repeat step 3 for all values of n.

Amplitude Spectrum
The discrete Fourier transform(DFT) of a length-N time-domain seqeuence x(n) is
given by

N 1
j 2 nk / N
X [k ] x[n] e 0 k N 1 (5)
n0

Since each calculated DFT value is a complex number, it is not convenient to plot
it versus its frequency index. Hence, after evaluating Eq.(3), the magnitude of each DFT
value (referred as the amplitude spectrum ) can be determined and plotted versus its
frequency index. The amplitude spectrum is defined as
1
Ak (6)
N N
The amplitude spectrum can be modified to a one sided amplitude spectrum by
doubling theamplituded in Eq.(6), keeping the original DC term at k = 0. Thus

1
X [0] , k 0
N
Ak (7)
2
N

2
 X [k ] , k 0,1, 2, ..... N / 2

Matlab Programs

1. Given a sequence, write a matlab program to find its DFT and IDFT
(a) without using matlab function.
(b) with matlab function
2. A continuous-time signal is given by
x(t) = 100sin(100 ) + 20sin(300 ) + 3sin(500 ) (8)

(c) write a matlab program to obtain the frequency components of the

signal with minimum number of samples.
(d) Write a matlab program to plot the amplitude spectrum using fft showing the
frequen- cies present in the signal.

Matlab Syntax and Description

Y = fft(x) computes the discrete Fourier transform (DFT) of X using a fast Fourier
transform (FFT) algorithm.

3
MATLAB CODE:

1.

4
OUTPUT :

5
6
7
8
 OUTPUT GRAPH:

Results and Inference:

 The code begins by plotting the original input sequence, providing a reference
for the time-domain signal. This is followed by computing the Discrete Fourier
Transform (DFT) using both a direct method and the Fast Fourier Transform
(FFT) function, ensuring the accuracy of the DFT computation.

 The direct method calculates the DFT by summing the product of the input
sequence and a complex exponential factor for each frequency index
 k. This approach, while computationally less efficient than the FFT algorithm,
offers a detailed insight into the DFT computation process.

 The FFT function computes the DFT using an optimized algorithm. The result
should closely match the DFT obtained via the direct method, confirming the
transformation's correctness.

 To verify the consistency between the two methods, the code calculates and
displays the error
 e between the direct method DFT and the FFT result. Ideally, this error should
be close to zero, indicating that both methods yield consistent results.

 The DFT results are then converted from Cartesian coordinates (real and
imaginary parts) to polar coordinates, specifically magnitude
r and angle. This conversion helps in understanding the amplitude and phase
9
 The magnitude and phase of the DFT are plotted, offering a visual
representation of the frequency spectrum of the input sequence. These plots are
crucial for identifying the frequencies present and understanding the
corresponding phase shifts.

 The code proceeds to compute the inverse DFT (IDFT) to reconstruct the
original sequence from the DFT coefficients. The reconstructed sequence x(11)
 is displayed and should ideally match the original input sequencex, verifying the
accuracy of the DFT and IDFT processes.

 Lastly, the error

 𝑒1 between the original sequence and the reconstructed sequence is computed
and displayed. A near-zero error confirms that the IDFT has accurately
restored the original sequence, validating the overall process.

Name: ___AKSHIT GHUNAWAT ____ _ ____ __ Reg. No.: ____CB.EN.U4EEE22059 ___ _

ASSESSMENT

Submission

Signature

Student Task Max. Marks Graded Marks

Conduction 5

Results & Conclusion 5

Total 10

1
0
1
1