Scribd
Upload
45 views

DSP Tutorial-1

1. The document contains tutorial questions on discrete-time signals and systems. It includes questions on convolution, impulse response, periodicity, linearity, and operations on signals like shifting and upsampling. 2. Answers to the questions are provided. For question 1, it is shown that x(n-n1) * h(n-n2) = y(n-n1-n2) using the definition of convolution. 3. In question 5, the effect of shifting a signal by 2 samples and upsampling a signal by 2 are illustrated through examples.

Uploaded by

ARUOS Soura
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
45 views

DSP Tutorial-1

1. The document contains tutorial questions on discrete-time signals and systems. It includes questions on convolution, impulse response, periodicity, linearity, and operations on signals like shifting and upsampling. 2. Answers to the questions are provided. For question 1, it is shown that x(n-n1) * h(n-n2) = y(n-n1-n2) using the definition of convolution. 3. In question 5, the effect of shifting a signal by 2 samples and upsampling a signal by 2 are illustrated through examples.

Uploaded by

ARUOS Soura
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 14

DSP Tutorial-1

Tutorial Questions

1. Let 𝒚(𝒏) = 𝒙(𝒏)* 𝒉(𝒏)

(where ‘*’ represent convolution).

Then show that


𝒙(𝒏−𝒏𝟏)∗𝒉(𝒏−𝒏𝟐) = 𝒚(𝒏−𝒏𝟏−𝒏𝟐).
Tutorial Questions

2. The impulse response of a discrete LTI system is


given by 𝒉(𝒏) = (𝟎.𝟓)𝒏 𝒖(𝒏) of the input is

𝒙(𝒏) = 2𝜹(𝒏)+𝜹(𝒏−𝟑).

Find the output at 𝒏 = 𝟏 and 𝒏 = 𝟒.


Tutorial Questions
n
3. Consider the signal x[n]  cos(   ) .
4

Comment about the periodicity of the signal.


Tutorial Questions

4. Consider a discrete-time system:


y[n]  k  e x[ k ]
n

Is this a linear or a non-linear system?


Tutorial Questions
5. Consider a discrete sinusoidal signal denoted by
𝑥 𝑛 , shown in Fig.
Find out the following signals and sketch the
intermediate stages if required.
(a) 𝑦1 = 𝑥 𝑛 − 2
(b) 𝑦2 = 𝑥 𝑛 2

FIG.
Answers

1. To Prove: x[n-n1] * h[n-n2] = y[n-n1-n2]


From convolution definition,
∞
y[n] = x[n] * h[n] = 𝑘=−∞ x[k] h[n-k] …..(i)
∞
x[n-n1] * h[n-n2] = 𝑘=−∞ x[k-n1] h[n-k-n2] ……..(ii)
let k-n1 = p then k = p+n1, from eq(ii)
∞
x[n-n1] * h[n-n2] = p=−∞ x[p] h[n-n1-n2-p] = y[n-n1-n2], (from eq(i) by
replacing n by n-n1-n2)
Therefore, x[n-n1] * h[n-n2] = y[n-n1-n2].
2. y(n) = x(n) * h(n) = (2𝛿[𝑛]+𝛿[𝑛−3]) * ((0.5)𝑛 𝑢[𝑛])
y(n)=2(0.5)𝑛 (𝑛) + (0.5)𝑛−3 𝑢[𝑛−3] [utilizing sifting property]
y(1) = 2*(0.5)1 𝑢[1] + 0 = 1
y(4) = 2*(0.5)4 𝑢[4] + (0.5)4−3 𝑢[1] = (0.5)3 + (0.5) = ⅝

** Sifting property in convolution:


3. In third case, we have,
 n   n N  4 
x[n   ]  cos     cos  
 4   4 4 

Again making use of the periodicity equation, we get (for k= 1),


N  4
 2k
4
N 4

Hence, the signal is periodic with period N=4.


n n

 e 1 and y 2 [ n]  
x2 [ k ]
4. We have, y1[n]  x [k ]
e
k   k  

y [n]  y [n]  
n x 1[k] n x 2[k]
So,
1 2
Π
k  
e Π
k  
e
n n
Σ x 1[k] Σ x 2[k]

e k  
e k  

n
To prove additivity, y3[n]   e x [ k ] x [ k ] 1 2

k  
n n
n
 x1 [ k ]  x2 [ k ]
 y3[n]  e  y3[n]  e
x1 [ k ] x2 [ k ]
.e k  k 

k  

y [n ]  y [n ]  y [n ]
3 1 2

Therefore, the system is non-linear.


5 (a) 𝑦1 (𝑛) = 𝑥 𝑛 − 2
Here, the samples of 𝑥 𝑛 will be shifted by 2.

𝒚𝟏 (𝒏) 𝒙 𝒏−𝟐 𝒚𝟏 (𝒏) 𝒙 𝒏−𝟐


𝑦1 (−3) 𝑥 −5 𝑦1 (4) 𝑥 2
𝑦1 (−2) 𝑥 −4 𝑦1 (5) 𝑥 3
𝑦1 (−1) 𝑥 −3 𝑦1 (6) 𝑥 4
𝑦1 (0) 𝑥 −2 𝑦1 (7) 𝑥 5
𝑦1 (1) 𝑥 −1 𝑦1 (8) 𝑥 6
𝑦1 (2) 𝑥 0 𝑦1 (9) 𝑥 7
𝑦1 (3) 𝑥 1
1/16/2020 11
5 (a) 𝑦1 = 𝑥 𝑛 − 2
Here, the samples of 𝑥 𝑛 will be shifted by 2.

1/16/2020 12
5 (b) 𝑦2 (𝑛) = 𝑥(𝑛 2)
Here, the samples of 𝑥 𝑛 is up-sampled by factor 2.

𝒚𝟐 (𝒏) 𝒙 𝒏 𝟐 𝒚𝟐 (𝒏) 𝒙 𝒏 𝟐
𝑦2 (−3) 𝑥 −3 2 𝑦2 (4) 𝑥 2
𝑦2 (−2) 𝑥 −1 𝑦2 (5) 𝑥 5 2
𝑦2 (−1) 𝑥 −1 2 𝑦2 (6) 𝑥 3
𝑦2 (0) 𝑥 0 𝑦2 (7) 𝑥 7 2
𝑦2 (1) 𝑥 1 2 𝑦2 (8) 𝑥 4
𝑦2 (2) 𝑥 1
𝑦2 (3) 𝑥 3 2
1/16/2020 13
5 (b) 𝑦2 (𝑛) = 𝑥(𝑛 2)
The odd values of ‘n’ provide fractional values of ‘n/2’. For these cases,
𝑥 𝑛 2 does not have any specified values. We can put zeros instead.

1/16/2020 14

You might also like