EE6403 - Digital Time Signals and signal processing

B.E. / B.Tech. DEGREE EXAMINATION, APRIL/MAY 2017

PART A – (10 X 2 = 20 marks)
1.What are the energy and power of discrete signal?
The average power P of a discrete time signal x(n) is defined as,

N
1
Power  lim 
2
x(n) (for aperiodic signal)
N  2 N  1
n  N

1 N 1
Power   x(n) (for periodic signal)
2

N n 0

The energy E of a discrete time signal x(n) is defined as,


Energy  
2
x ( n)
n 

If E is finite then x(n) is called as an energy signal.

If E is infinite but P is finite then x(n) is called as power signal
2. State sampling theorem.
A band limited continuous time signal, with highest frequency ‘B’ Hertz, can be
uniquely recovered from its samples provided that the sampling rate Fs≥2B samples
per second.

3.Write the properties of region of convergence?

i)The ROC cannot contain any poles
ii)The ROC must be a connected region
iii) The ROC of an LTI stable system contains the unit circle
iv) The ROC is a ring in the Z plane centered ay the origin.

4.Find the convolution of the input signal {1, 2, 1} and its impulse response {1, 1, 1}
using Z transform.
x n   1,2,1 h n   1,1,1
 
2
X  z    x n  Z n  1  2z 1  z 2
n 0
2
H  z    h n  Z n  1  z 1  z 2
n 0

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 1

 EE6403 - Digital Time Signals and signal processing

Y  z   X  z  H  z   1  2z 1  z 2 1  z 1  z 2 
 1  2z 1  z 2  z 1  2z 2  z 3  z 2  2z 3  z 4
 1  3z 1  4z 2  3z 3  z 4
y n   1,3,4,3,1

5.Define twiddle factor. Write its magnitude and phase angle.
The N point DFT is given by,

N 1  j 2 kn
X (K )   x (n )e N , k=0,1,2……..,N-1
n 0
 j 2
Let wN e N which is called as twiddle factor.
N 1
X (K )   x (n )wN kn
n 0
2
Its magnitude is one and phase angle is N
6.Compute the number of multiplications and additions for 32 point DFT and FFT.
DFT FFT

No of additions needed -- 32x1023=32736 32 log 32=32x5=160

No of multiplications needed—322=1024 (32/2) log 32= 16x5=80

7.Write the advantages and disadvantages of digital filters.

8.Define prewarping effect.

When the desired magnitude response is piece-wise constant over frequency, the
effect of the non-linear compression at high frequencies can be compensated by
introducing a suitable pre scaling or pre warping the critical frequencies by using the
formula   2 tan  .
T 2

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 2

 EE6403 - Digital Time Signals and signal processing

9.What is pipelining and how do define its depth?

10.Write some commercial DSP processors.

A digital signal processor ('DSP) is a specialized microprocessor with its architecture

optimized for the operational needs of digital signal processing.

PART B – (5 X 13 = 65 marks)
11.(a)Determine the following systems are linear, stability and time invariance of the
system (i) y(n) = x(2n), (ii) y(n) = cos x(n), (iii) y(n) = x(n) + n x(n+1) (13)
Linearity:
i) y[n] = x[2n]
The output y(n) for an arbitrary input x1(n) is, y1 (n)  x1 (2n)
The output y(n) for an arbitrary input x2(n) is, y2 (n)  x2 (2n)
Thus the sum of the individual response is,
y1 (n)  y2 (n)  x1 (2n)  x2 (2n) 1
When the inputs are applied simultaneously then the response due to sum of inputs
is,
y1 (n)  y2 (n)  x1 (2n)  x2 (2n) 2
As an equation (1) is equal to the equation (2) the given system is linear.
ii) y[n] = Cosx[n]
The output y(n) for an arbitrary input x1(n) is, y1 (n)  Cos x1 (n)

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 3

 EE6403 - Digital Time Signals and signal processing

The output y(n) for an arbitrary input x2(n) is, y2 (n)  Cos x2 (n)
Thus the sum of the individual response is,
y1 (n)  y2 (n)  Cos x1 (n)  Cos x2 (n) 1
When the inputs are applied simultaneously then the response due to sum of inputs is,
y1 (n)  y2 (n)  Cos  x1 (n)  x2 (n) 2
As an equation (1) is not equal to the equation (2) the given system is non linear.
iii) y[n] = x[n]+ n x[n+1]
The output y(n) for an arbitrary input x1(n) is, y1 (n)  x1 (n)  nx1 (n  1)
The output y(n) for an arbitrary input x2(n) is, y2 (n)  x2 (n)  nx2 (n  1)
Thus the sum of the individual response is,
y1 (n)  y2 (n)  x1 (n)  nx1 (n  1)  x2 (n)  nx2 (n  1) 1
When the inputs are applied simultaneously then the response due to sum of inputs
is,
y1 (n)  y2 (n)  x1 (n)  x2 (n)  n  x1 (n  1)  x2 (n  1)  2
As an equation (1) is equal to the equation (2) the given system is linear.
Time-invariant vs. time-variant:
i) y[n] = x[2n]
if x1[n] = x[n − n0]
The response due to delayed input is, y1[n] = x1[2n] = x[2(n − n0)]
while the delayed response is, y[n − n0] = x[2n − n0]
So the system is time variant.
ii) y[n] = Cos x[n]
if x1[n] = x[n − n0]
The response due to delayed input is, y1[n] = Cos x[n − n0]
While the delayed response is, y[n − n0] = Cos x[n − n0]
So the system is time-invariant.
iii) y[n] = x[n]+ n x[n+1]
if x1[n] = x[n − n0]
The response due to delayed input is, y1[n] = x[(n − n0)]+ n x[(n − n0)+1]
while the delayed response is, y[n − n0] = x[(n − n0)]+ [(n − n0) x[(n − n0)+1]
So the system is time-variant.
Stability:
i) y[n] = x[2n]
The impulse response is, h(n)   (2n)

The condition for stability is,




n 
h(n)  ...  0  0  1  0  0  0  ...  1  

Hence the system is stable.

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 4

 EE6403 - Digital Time Signals and signal processing

ii) y[n] = Cosx[n]

The impulse response is, h(n)  Cos  (n) where  (n)  1 for n  0
The condition for stability is,



n 
h(n)  ...  1  1  Cos1  1  1  1  ...  

Hence the system is unstable.

iii) y[n] = x[n]+ n x[n+1]
The impulse response is, h(n)   (n)  n (n  1)

The condition for stability is,




n 
h(n)  ...  0  1  1  0  0  0  ...  2  

Hence the system is stable.

11.(b) (i) Explain the process of quantization and its error types.
(10)
Quantization is the process of conversion of discrete time continuous valued
signal into a discrete time discrete valued signal. The difference between actual
output and quantized output is called quantization error.
As the number of bits in the digital word is finite, A/D conversion results in a finite
resolution. The digital output can assume only a finite number of levels, and
therefore an analog number must be either rounded off or truncated to the nearest
digital level. So the error that exists due to the quantization process is called
quantization error.

Let xq(n) denote the sequence of quantized samples at the output of the quantizer.
Hence xq(n) = Q[x(n)]. Then the quantization error is a sequence eq(n) defined as the
difference between the quantized value and the actual sample value.

Thus eq(n) = xq(n)-x(n). Quantization is done two ways.

(i) Truncation (excess digits are discarded)

(ii) Rounding (rounded off to next integer level). The distance  between
two successive quantization levels is called the quantization step size (or) resolution.
 
The quantization error eq(n) in rounding is limited to the range of 2
to 2
. In other
words, the instantaneous quantization error cannot exceed half of the quantization
step. If x min & x max represent the minimum and maximum values of x(n) & the
number of quantization levels, then

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 5

 EE6403 - Digital Time Signals and signal processing

x max  x min
 
L1

The quantization error decreases and the accuracy quantizer increases as the number
of quantization level increases. The quantization error eq(t) = xa(t) – xq(t)

Let  denotes the time that xa(t) stays within the quantization level. The mean-
square error power Pq is,

 
1 1
Pq 
2  e q (t ) dt    e2q (t ) dt
2

 0


eq ( t )  t
2


  2
1 2
 Pq  
t dt 
  2  12
0

If the quantizer has ‘b’ bits of accuracy and the quantizer covers the entire range 2A,
2A 4 A2 A2
the quantization step is   2b
. Hence,  Pq  22 b 12  22 b 3

 3t 
11.(b) (ii) Compute the Nyquist sampling frequency of the signal x (t )  4 Sinc   (3)
  
3t
Sin 
 3t   4Sin(3t)
x (t )  4 Sinc    4 3t 
    3t

The maximum frequency component is,
2 f m  3  f m  3/ 2
Nyquist rate is, Fn  2 f m  3/  Hz
12.(a)(i) State and prove convolution and Parseval’s theorem of Z transform. (6)
Convolution theorem
If x1 (n) & x2 (n) are two sequences
x1 (n) 
z
 X 1 ( z ) with ROC : R1 &
x2 (n) 
z
 X 2 ( z ) with ROC : R2
x (n)  x2 (n)   X1 ( z) X 2 ( z) with ROC : R1  R2
Then 1 z

Proof:

 
  
Z .T  x1 (n)  x2 (n)      x (k ) x (n  k ) z
1 2
n
  x (k )  x (n  k ) z
1 2
nk

n  k  k  n 


  x (k )z
k 
1
k
X 2 ( z)  X1 ( z) X 2 ( z)

Hence proved.

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 6

 EE6403 - Digital Time Signals and signal processing

Parseval’s theorem:
If x1 (n) & x2 (n) are two complex valued sequences
x1 (n) 
z
 X 1 ( z ) with ROC : R1 &
x2 (n) 
z
 X 2 ( z ) with ROC : R2
N 1
1
Then Z .T  x1 (n) x2 (n)  
N
 X (k )Y
K 0
*
(k )

Proof:

12.(a). (ii) Find the Z transform of the system x(n )  Cos(n )u(n ) (7)
Solution:
 e j n  e j n
Cos (n )u (n)   for n  0
 2

 

  Cos (n )u (n)   0.5  e j n z  n   e  j n z  n 
 n 0 n 0 
   e j n   e j  n 
  Cos (n )u (n)   0.5        
 n 0  z  n 0  z  

z  z  2z  e  e  
j  j
 z
 0.5     2 
 z  z  e  e   1 
j  j 
 z  e z  e  2 j  j

z 2  z cos 
  Cos (n )u (n)  
z 2  2 z cos   1
12. (b).Find the inverse Z transform of X(z) = (z+1) / (z+0.2) (z-1), z  1
Using residue method. (13)
Solution:

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 7

 EE6403 - Digital Time Signals and signal processing

1
2 j 
n 1
x ( n)  X ( z )z dz
C

  residues of X ( z ) z n 1at poles of X ( z ) z n 1 within C

( z  1) z n 1
  residues of at poles of same within C
( z  0.2)( z  1)
The closed contour C, begin in the ROC z>1 encloses the poles at z=-0.2, z=1 and
for n=0 the pole is at z=0. Therefore for n=0
( z  1)
x(0)   residues of at poles z  0, z  1 and z  0.2
z ( z  0.2)( z  1)
( z  1) ( z  1) ( z  1)
 ( z)  ( z  0.2)  ( z  1)
z ( z  0.2)( z  1) z 0 ( z  0.2)( z  1) z  0.2
( z  0.2)( z  1) z 1

10 5
 5    0
3 3
i.e, x(0)  0
for n  1
( z  1) z n 1
x(n)   residues of at poles z  0.2 and z  1
( z  0.2)( z  1)
( z  1) z n 1 ( z  1) z n 1
 residue of at z  0.2  residue of at z  1
( z  0.2)( z  1) ( z  0.2)( z  1)
( z  1) ( z  1)
 ( z  0.2)  ( z  1)
( z  0.2)( z  1) z 0.2 ( z  0.2)( z  1) z 1

2 5
  (0.2) n 1 
3 3
2 5
Therefore, x(n)   (0.2) n1 u(n  1)  u (n  1)
3 3
13.(a)Determine the 8 point DFT of the sequence x(n) = {1, 1, 1, 1, 1, 0, 0,0} (13)
N 1 j 2 kn 7 j 2 kn
( ) ( )
X k    x n  e N   x n  e 4 where N  8 and k  0,1, 2, 3, 4, 5, 6, 7

n 0 n 0
j k
 j 3 k j 5 k j 7 k
 j k   
X k   x (0)  x (1)e 2  x (2)e  x (3)e 2
 x (4)e  j 2 k  x (5)e 2
 x (6)e  j 3 k  x (7)e 2

X (0)  x (0)  x (1)  x (2)  x (3)  x (4)  x (5)  x (6)  x (7)  5

j
 j 3
 j 
X (1)  x (0)  x (1)e 2  x (2)e  x (3)e 2
 x (4)e  j 2
j 5 j 7
 
 x (5)e 2
 x (6)e  j 3  x (7)e 2
  j 2.414
 j  j 2
X (2)  x (0)  x (1)e  x (2)e  x (3)e  j 3  x (4)e  j 4
 x (5)e  j 5  x (6)e  j 6  x (7)e  j 7 1
j 3
 j 9
 j 3 
X (3)  x (0)  x (1)e 2  x (2)e  x (3)e 2
 x (4)e  j 6
j 15 j 21
 
 j 9
 x (5)e 2
 x (6)e  x (7)e 2
  j 0.414

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 8

 EE6403 - Digital Time Signals and signal processing

 j 12
 x (2)e  j 4
 j 2 
X (4)  x (0)  x (1)e  x (3)e 2
 x (4)e  j 8
 x (5)e  j10  x (6)e  j12  x (7)e  j14  1
j 5
  j15
X (5)  x (0)  x (1)e 2  x (2)e  j 5  x (3)e 2
 x (4)e  j10
j 25 j 35
 
 x (5)e 2
 x (6)e  j15  x (7)e 2
 j 0.414
j 6
 j 18
 j 6 
X (6)  x (0)  x (1)e 2  x (2)e  x (3)e 2
 x (4)e  j12
j 30 j 42
 
 x (5)e 2
 x (6)e  j18  x (7)e 2
1
j 7
 j 21
 j 7 
X (7)  x (0)  x (1)e 2  x (2)e  x (3)e 2
 x (4)e  j14
j 35 j 49
 
 x (5)e 2
 x (6)e  j 21  x (7)e 2
 j 2.414
X k   5,  j 2.414,1,  j 0.414,1, j 0.414,1, j 2.414

13. (b)Compute 8 point DFT of the given sequence using DIT algorithm
n n 7 
x (n )   
0 otherwise  (13)

The twiddle factors associated with the flow graph are,

2 1
j
w8  1; w  e
0
8
1 8
 0.707  j 0.707

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 9

 EE6403 - Digital Time Signals and signal processing

2 1
j
w41  w82   j; w83  e 8
  0.707  j 0.707
X ( K )  28, 4  j9.6565, 4  j 4, 4  j1.656, 4, 4  j1.656, 4  j 4, 4  j9.6565
14.(a)Design a 15 tap linear phase filter using frequency sampling method to the
 1 0  k  3
 2 k   
H  0.4 k 4 
following discrete frequency response  15  
 0 k  5,6,7 
 

14. (b)Using bilinear transformation, design a high pass filter, monotonic in

passband with cutoff frequency of 1000 Hz and down 10 dB at 350Hz. The sampling
frequency is 5000Hz. (13)
Solution:
The high pass filter specifications are,
 p  3dB;  p  2   1000  2000 rad / sec
 s  10dB; s  2    350  700 rad / sec
1 1
T   2 104 sec
f 5000

2   pT
 2  2000  2 104 
 pH  tan    4
 tan    7265 rad / sec
T  2 2 10  2 
 T   700  2 10 
4
2 2
 SH  tan  s   4
 tan    2235 rad / sec
T  2  2 10  2 
The characteristics are monotonic in both pass band and stop band. Therefore, the filter is
Butterworth filter.
Design of prototype LPF:
 pL  1
 pH
 SL   3.25
 SH

Determination of order:

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 10

 EE6403 - Digital Time Signals and signal processing

 100.1 s  1  100.1(10)  1
 log 0.1 p  log
 10  1  100.1(3)  1  0.932
N  
 log   SL   log 3.25
   
 PL  
Choose N=1
Determination of cutoff frequency:
C   pH  7265rad / sec

C
The transfer function of highpass filter is obtained using the transformation S  i.e.,
S
7265
S
S
1 S
H (S )  
1 S S
7265 S  7265
S

Using bilinear transformation

H ( z )  H ( S ) S  2  1 z 1 
T  1 z 1 

 1  z 1 
10000  
S  1  z 1 
 
S  7265 S  2  1 z 1   1  z 1 
210 4
 
1 
 1 z  10000  1 
 7265
 1 z 
0.5792(1  z 1 )
H ( z) 
1  0.1584 z 1
15.(a)Discuss the features and architecture of TMS 320C50 processor. (13)

Architecture of TMS 320C50 processor:

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 11

 EE6403 - Digital Time Signals and signal processing

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 12

 EE6403 - Digital Time Signals and signal processing

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 13

 EE6403 - Digital Time Signals and signal processing

15. (b)Explain the addressing modes and registers of DSP processors. (13)
Refer Q.No. 15. (b) May 2015 Answer Key for addressing modes

PART C – (1 X 15 = 15 marks)
16.(a)The analog signal has a bandwidth of 4KHz. If we use N point DFT with N =
2m (m is an integer) to compute the spectrum of the signal with resolution less than
or equal to 25 Hz. Determine the minimum sampling rate, minimum number of
required samples and minimum length of the analog signal. What is the step size
required to quantize this signal. (15)

The bandwidth of the analog signal (B)=4kHz

Minimum sampling rate= fs =2B= 8000 Hz

Minimum number of samples required:

Resolution =25 Hz =fs/N= 2B/N

N=8000/25 =320

or

1 1 8000
 25 L L  320 samples
LT 25T 25

Minimum length of the analog signal:

1
LT   320  0.04sec
8000

The step size required to quantize the analog signal:

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 14

 EE6403 - Digital Time Signals and signal processing

16.(b)Convert the single pole low pass filter with system function
0.5(1  Z 1 )
H (z )  into band pass filter with upper and lower cutoff
1  0.302Z 1
frequencies u and l respectively. The lowpass filter has 3dB bandwidth and
p   /6 and u  3 /6 , l   /4 and draw its realization in direct form II.
[15]

a1 
2 K
& a2 
 K  1
 K  1  K  1
   l   p
K  Cot  u  tan
 2  2
 u  l    u  l  
  Cos   /Cos  
 2   2 
p   /6 and u  3 /6 , l   /4
We have,
We find,   0.414, K  1.4,a1  0.483& a2  0.167

Prepared by Prof.S.Nagammai, HOD/EIE, KLNCE Page 15