Analog/Digital Signal Processing

Analog Signal Processing

Tutorial 2: Sampling and Reconstruction
Prepared by Prof. Dr. Thuong Le-Tien
March 2020

1. The analog signal x(t) =10 sin(4t) + 10sin(12t) + 5sin(20t), where t is in seconds,
is sampled at a rate of fs = 8Hz. Determine the signal xa(t) aliased with x (t) . Show that the
two signals have the same sample values, that is, show that x(nT)=xa(nT). Repeat the above
questions if the sampling rate is fs =14Hz.
2. The analog signal xt)=sin(6t)[1+2cos(4t)], where t is in milliseconds, is sampled at a rate
of 4kHz. The resulting samples are immediately reconstructed by an ideal reconstructor.
Determine the analog signal xa(t) at the output of the reconstructor.
 
3. The sequence x(n)  cos n ,    n   was obtained by sampling a CT signal
4 
xa (t )  cos(0t ),    t   at a sampling rate of 1000samples/sec. What are two
possible values of 0 that could have resulted in the sequence x(n)?
4. The Continuous Time (CT) signal xa(t)=sin(20t) + cos(40t) is sampled with a sampling
 n   2n 
period T to obtain the Discrete Time (DT) signal x(n)  sin   cos 
 5   5 
a. Determine a choice for T consistent with this information
b. Is your choice for T in Part (a) unique? If so, explain why. If not, specify another
choice of T consistent with the information given.
5. The Continuous Time (CT) signal xa(t) = cos(800t) is sampled with period T to obtain a
Discrete Time (DT) signal xd(n)=xa(nt)
a. Compute and sketch the magnitude of the CT Fourier Transform of xa(t) and the DT
Fourier Transform of xd(n) for T=1ms
b. Repeat part (a) for T=2ms
c. What is the maximum period Tmax such that no aliasing occurs in the sampling
process?
6. Consider the following sound wave, where t is in milliseconds:
x(t)=sin(10t) + sin(20t) +sin(60t)+ sin(90t)
This signal is prefiltered by an analog antialiasing prefilter H(f) and then sampled at an
audio rate of 40 kHz. The resulting samples are immediately reconstructed using an ideal
reconstructor. Determine the output ya(t) of the reconstructor in the following cases and
compare it with the audible part of x(t):
a. When there is no prefilter, that is, H(f) 1.
b. When H(f) is an ideal prefilter with cut off of 20 kHz.
c. When H(f) is a practical prefilter that has a flat passband up to 20 kHz and attenuates
at a rate of 48 dB/octave beyond 20 kHz. (You may ignore the effects of the phase
response of the filter.)
 Digital Signal Processing - Tutorial No. 3
Quantization Error, Coding and Noise Shaping
Prepared by Prof. Dr. Thuong Le-Tien
March 2020

1. Consider an A/D converter with full scale range of R=4 volts. Determine the quantized
value as well as the 3 bit representation of the following analog input values: x=1.9, 1.1,
0.7, -2,2, -1.3, -0.4, -1.7, -0.9 by writing a table of results. Repeat using an offset binary
converter and discussing about the unipolar binary?

2. Consider the signal x(n)=2sin(2fn), where f=0.04cycles/sample. This signal is to be

quantized using a 4-bit successive approximation bipolar ADC whose full-scale range is
R=8volts. For n=0,1,….9, compute the numerical value of x(n) and its quantized version
xQ(n) as well as the corresponding bit representation at the output of the converter. Do
this both for an offset binary converter and a two’s complement converter.

3. A digital audio mixing system uses 16 separate recording channels, each sampling at a
44.1kHz rate and quantizing each sample with 16 bits. The digitized samples are saved
on a hard disk for further processing.
a. How many megabytes of hard disk space are required to record a 3-minute song
for a 16-channel recording?
b. Each channel requires about 25 multiplier/accumulation (MAC) instructions to
perform the processing of each input sample (this corresponds to about 12 second-
order parametric EQ filters covering the audio band). In how many nanoseconds
should each MAC instruction be executed for:
(i) Each channel?
(ii) All 16 channels, assuming they are handled by a single processor?

4. If the quantized value xQ is obtained by truncation of x instead of rounding, show that the
truncation error e = xQ - x will be in the interval [–Q, 0]. Assume a uniform probability
density p(e) over this interval, that is

a. Find the mean of error

b. Find the variance of error

5. In a audio codec, it is desired to maintain quality of 16-bit resolution at 44.1kHz

sampling rates using a 12-bit oversampled noise shaping quantizer. For quantizer order
p=1,2,3, determine the corresponding oversampling ratio L and oversampling rate Lfs in
Hz.
 Digital Signal Processing
Tutorial No.4b for LTI Systems
Prepared by Prof. Dr. Thuong Le-Tien
Sept. 2021

1) Verify whether the system are stable or not?

a. h[n]  b nu[n]
b. h(t )  t sin t u (t )
c. h(t )  e  j 2t u (t )
d . y[n]  4n 1 u[n  2]
2) Check the Linearity and Time Invariance of the systems :
a. y (t )  x(t  1)  2tx(t  2)
b. y[n]  x[n]cos  n
c. y ( n)  e x ( n )
3) Check the systems are memory or memoryless systems:
a. y (t )  x(t )  x(t  1)
b. y (t )  x 3 (t )
c. y[n]  x[n  1]  x[1  n]
d . y[n]  e jnu[n]
4) Is the following systems causal?
1
a. y[n]  x[n] 
x[n  1]
b. y[n]  2 x[n]  3 x[n  3]
5) Whether the system is linear?
y[ n]  x[ n]  nx[n  1]
6) Find the impulse response h[n] for each of the causal LTI discrete-
time systems satisfying the following difference equations and
indicate whether each system is a FIR or an IIR system.
a) y[n]  x[n]  2 x[ n  2]  x[n  3] ;
b) y[ n]  2 y[n  1]  x[ n]  x[ n  1]
1
c) y[ n]  y[ n  2]  2 x[ n]  x[ n  2]
2
7) Convolution : LTI system
x[n]  (2,3, 4,1, 2)
Find y(n) when
h[n]  (1, 2,1, 2)
 DIGITAL SIGNAL PROCESSING
TUTORIAL No4: Discrete Time Systems
Dated: June 2021
Lectured by Prof. Dr. Thuong le-Tien

Question 1
A discrete time signal x(n) is given in the figure

Sketch and label carefully each of the following signals:

a. x(n-2)
b. x(4-n)
c. x(2n)
d. x(n)u(2-n)
e. x(n-1)(n-3)
Question 2
Consider an LTI system with the frequency response
( ) 1+e + 4e
H e =e , − π < ≤
1 + 0.5e
Determine the output y(n) for all n if the input for all n is: x(n)=cos(0.5n)

Question 3
A linear time-invariant system is described by the input-output relation (I/O relation)
y(n) = x(n) + 2x(n-1) + x(n-2)
a. Determine h(n), the impulse response of the system
b. Is this a stable system?
c. Determine H(ej), the frequency response of the system. Use trigonometric identities to
obtain a simple expression for H(ej).

Question 4
The input and output of a stable LTI system

a. Find the response of the system to the signal x1(n)

b. Find the impulse response h(n) of the system
 Digital Signal Processing
Tutorial 5: Discrete Time Systems, Transients of Convolution
Prepared by Prof. Dr. Thuong Le-Tien
Date: October 2021

Question 1
The Impulse response h(n) of a filter is non zero over the index range of n be [3,6]. The input
signal x(n) to this filter is non zero over the index range of n be [10,20]. Consider the direct and
LTI forms of convolution

a. Determine the overall index range n for the output y(n). For each n, determine the
corresponding summation range over m, for both the direct and LTI forms.
b. Assume h(n) = 1 and x(n) = 1 over their respective index ranges. Calculate and sketch
the output y(n). Identify (with an explanation) the input on/off transient and steady state
parts of y(n).
Question 2:
Given x(n)={-1.0.2.1.-3.2.5.3.-2,2} and h(n)={-1,-2, 0,3,2}, calculate y(n)=x(n)*h(n) by
a. Using the direct form
b. Convolution table
c. Overlap add method with block 4 samples
d. Flip and slide form
Question 3
An LTI filter has impulse response h(n) = an u(n), where /a/<1. Using the convolution
summation formula in question 1, derive closed-form expressions for the output signal y(n)
when the input is:
a. A unit step, x(n) = u(n)
b. An alternating step, x(n) = (-1)n u(n)
c. In each case, determine the steady state and transient response of the filter.
Question 4
Consider the filter with I/O equation: y(n) = x(n) – 2x(n-3)
a. Determine the impulse response sequence h(n) for all n(> or =)0
b. Draw the block diagram realization of the filter
c. The input sequence x(n) = [1,1,2,2,4,-3,-2,0,2,-1]. Using convolution, compute the
corresponding y(n) using the Flip and slide method.
d. Repeat question (c) with the overlap add blocks of 4samples/block.
e. Classify the filter be FIR or IIR responses?
Question 5:
Consider the filter with I/O equation y(n) = 0.8y(n-1) + x(n)
a. Determine the impulse response h(n) in the forms: Recursive form; vector form with
specific filter coefficients; maths equation.
b. Draw the block diagram realization of the filter
c. Classify the filter be FIR or IIR responses?
 Digital Signal Processing
Tutorial No.6 - Z-Transform
Prepared by Prof. Dr. Thuong Le-Tien
Date: October 2021

Question 1: Compute the z-transform of the following sequences and determine

the corresponding region of convergence:
a. x(n)=2(0.9)n cos(n/2)u(n) c. x(n)=(0.9j)nu(n)+(-0.9j)nu(n)
b. x(n)=(0.5)nu(n)+(-0.5)nu(n) d. x(n)=(-0.5)n [u(n)-u(n-10)]
Question 2: Using partial fractions or power series expansions, determine all
possible inverse z-transforms of the following z-transforms, sketch their ROCs,
and discuss their stability and causality properties:
3(1  0.3 z 1 ) 6  2 z 1  z 2
a. X ( z )  e. X ( z )  ,
1  0.81z 2 (1  z 1 )(1  0.25 z 2 )
6  3 z 1  2 z 2 1
b. X ( z )  f. X ( z )  4 
1  0.25 z 2 1  4 z 2
6  z 5 4  0.6 z 1  0.2 z 2
c. X ( z )  g. X ( z ) 
1  0.64 z  2 (1  0.5 z 1 )(1  0.4 z 1 )
10  z 2
d. X ( z ) 
1  0.25 z 2
Question 3:
Consider the RLC circuit shown in Figure. Find the different equation relating the output
current y(t) and the input voltage x(t)
R L

y (t ) C

Question 4.
Consider the RL circuit shown in figure
a. Find the different equation relating the output voltage y(t) across R and the input
voltage x(t).
b. Find the input response h(t) of the circuit.
c. Find the step response s(t) of the circuit.
L

x (t ) y (t )
x (t )
R
 Digital Signal Processing
Tutorial No. 7: LTI Systems and Z Tranform
Prepared by Prof. Dr. Thuong Le-Tien
November 2021

Question 1:
Three systems A,B, and C have the inputs and outputs indicated in the figures. Determine
whether each system could be LTI. If your answer is yes, specify whether there could be
more than one LTI system with the given input-output pair. Explain your answer

Question 2:
Using partial fractions or power series expansions, determine all possible inverse z-
transforms of the following z-transforms, sketch their ROCs, and discuss their stability
and causality properties:
4  0.6 z 1  0.2 z 2 10  z 2
a. X ( z )  ; b. X ( z ) 
(1  0.5 z 1 )(1  0.4 z 1 ) 1  0.25 z 2

Question 3
The system L in the figure is known to be linear. Show are three output signals y1(n), y2(n),
y3(n) in response to three input signals x1(n), x2(n) and x3(n), respectively

a. Is the system L Time-Invariant?

b. Find the possible impulse response of the system?
DSP Lectured by Prof. Dr. Thuong Le-Tien

DIGITAL SIGNAL PROCESSING

Tutorial No.8 - System Properties and z-transform
November 2021
Question 1:
Assume that the zero-state response of an LTI system to input
x(n)=2-n u(n) is y(n) = (1/3)n u(n). Use the system properties to find
the system’s response (h(n) to a unit pulse input.

Question 2:
Compute the convolution x(n)*h(n) for the x(n) and h(n) below:
a. x(n)={-1,2,1}, h(n)={1,0,-1,2}
b. x(n)=(-4)-n u(n), h(n)={1,2,-3}
c. x(n)=(-1)-n u(n) and h(n)=e-n u(n)
d. x(n)=u(-n), h(n)=n(u(n)-u(n-3))

Question 3:
Find the z-transform (if it exists) and the corresponding region of
convergence for each of the following signals:
a. {1,0,-1,0,1,-1,3}
b. 2n u(n)-3n u(-n-1)
c. (0.8)n u(n) + 0.9n u(-n-1)
d. (1/2)n (u(n)-u(n-26)
e. (1/2)n cos(n/3+/4) u(n)

Question 4:
Find the inverse z-transform of
z ( z  4) z2 1
 
a. b. c.
z  6z  6
2
z2  9 z 10 z 2  2z  2

Question 5:
The 2-sided z-transform of x(n) is given by
z 1
X ( z) 
 
1  3z 1 1  5 z 1 
a. Determine all possible ROCs for X(z)
b. For each ROC in (a), find x(n)
c. Discuss the stability and causality of each case.

November 2021 1
 Tutorial No.9 for Digital Signal Processing
Discrete Fourier Transform and Fast Fourier Transform

Prepared by Prof. Dr. Thuong Le-Tien

November 2021

Q1. A 256ms portion of an analog signal is sampled at a rate of 16kHz and the
resulting L samples are saved for further processing. What is L? The 256 point
DFT of these samples is computed.
a. What is the frequency spacing in Hz of the computed DFT values?
b. What is the total number of required multiplications if the computations
are done directly using the definition of DFT?
c. What is the total number of required multiplications if the L samples are
first wrapped modulo 256 and then 256-point DFT is computed?
d. What is the total number of required multiplications if a 256-point FFT is
computed of the wrapped signal?

Q2. By definition, the first and second of Fibonacci numbers are 0 and 1 (e.g.
h(0)=0 and h(1)=1), and each subsequent number is the sum of the previous two.
a. Write the Fibonacci sequence h(n) for n=0, 1, …,9 described by the above
definition?
b. Calculate 4-point FFT of the sequence of ten numbers in (a) using the
definition in matrix form.
c. Recomputed by first reducing x modulo 4 (wrapped signal) and then
computing the 4-DFT of the result.
d. Finally, compute the 4-point IDFT of the result and verify that you recover
the mod-4 wrapped version of x.

Q3. Compute the 8-point FFT of the length-8 signal x, in which these samples are
the first 8 samples of x(n)=4cos(n/2)+cos(n), discuss whether the 8 computed
FFT values accurately represent the expected spectrum of x(n). What FFT indices
correspond to the two frequencies of the cosinusoids?

Q4. The 8-point DFT X of an 8-point sequence x is given by

X=[0,4,-4j,4,0,4,4j,4].
a. Using the FFT algorithm, compute the inverse IFFT: x=IFFT(X).
b. Using the given FFT X, express x as a sum of real-valued (co)sinusoidal
signals.
Q5.
a. Compute the DFT of the sequence x(n) = {1,0,2,0}
b. Without explicitly computing the DFT sum, find the DFT of the sequence
y(n) = {0,2,0,1}, using your answer to part (a)
c. Find the inverse DFT of the sequence X(m) = {1, e-j3/4, 0, ej3/4}
d. Without explicitly computing the inverse DFT sum, find the inverse DFT
of the sequence Y(m) = {1, e-j/4, 0, ej/4}, using your answer to part(c).
 Digital Signal Processing - by Prof. Dr. Thuong Le-Tien

DIGITAL SIGNAL PROCESSING

TUTORIAL No. 10: Realization and Frequency Response
Lectured by Prof. Thuong Le-Tien
Date: December 2021

Question 1:
n
11
When the input to a causal LTI system is x(n)     u (n)  2 n u n  1 and the z-
4
3 2 3
transform of the output is
1  z 1
Y ( z) 
1  z 
1

1  0.5 z 1 1  2 z 1 
a. Find the z-transform of x(n)
b. What is the region of convergence of Y(z)?
c. Find the impulse response of the system
d. Is the system stable?

Question 2:
Consider an LTI system that is stable and for which H(z), the z-transform of the impulse
response, is given by
2  4 z 1  2 z 2
H ( z) 
1  3z 1  4 z 2

Suppose x(n), the input to the system, is a unit step sequence.

a. Find the output y(n) by evaluating the discrete convolution of x(n) and y(n)
b. Find the output y(n) by computing the inverse z-transform of Y(z)
c. Sketch the Pole/zero pattern of H(z) and the magnitude of the frequency response
d. Realize the above filter with the direct form and Canonical form.

Question 3.
a. Determine the unit step response of the causal system for which the z-transform of
1 z3
the impulse response is H ( z ) 
1 z4
b. Realize the above filter with the direct form and Canonical form

1
 DIGITAL SIGNAL PROCESSING

TUTORIAL No. 11: Realization and Frequency Response

Lectured by Prof. Thuong Le-Tien
Date: December, 2021

Question 1:
3  7 z 1  5 z 2
The impulse response is given by H ( z )  . Suppose x[n], the input
1  2.5 z 1  z 2
to the system,
a. Find all available impulse responses in the time domain h[n] with their ROCs
and stabilities.
b. Discuss about the causality and stability properties of the results in (a)
c. Sketch the zeros and pole pattern in the Z-plane of the above system then
sketch the frequency response of the system.
d. Realize the block diagram of the system in Direct Form I and Direct form II
(Canonical Form)

Question 2:
Find the overall impulse response of following system?

Question 3:
A digital reverberation processor has frequency response:
 0.5  e  j 8
H ( ) 
1  0.5e  j 8
where  is the digital frequency in [radians/sample].
a. Determine the causal impulse response h(n), for all n>=0 and sketch it versus
n.
b. Find zeros and poles of the system.
c. Sketch frequency response of the system
d. Realized the block diagram for the system.
DSP Prepared by Prof. Dr. Thuong Le-Tien

DIGITAL SIGNAL PROCESSING

Tutorial No. 12: Signal Flow graphs
December 2021
Question 1:
Consider the signal flow graph shown in the figure is an implementation of a Digital Filter,

a. Find the transfer function of the system relating the z-transform.

b. Determine the difference equation relating to the input signal x(n) to y(n)
c. Realize the block diagram of the system that has the smallest possible number of delay
elements.

Question 2:
The 2-sided z-transform of an impulse response of a system h(n) is given by
z 1
H ( z) 
(1  3z 1 )(1  5 z 1 )
a. Find and sketch the Pole/zero pattern of the response then determine all possible ROCs
and discuss the stability and causality of each case for H(z)
b. Find all possible solution for h(n)
c. Find the I/O equation of the system
d. Realize the direct form I of the system.

Question 3:
z 3
A causal LTI digital filter has the transfer function H ( z ) 
4  z 2
a. Find and sketch the pole/zero pattern of the filter.
b. Find the impulse response h(n) of the filter
c. Is the filter stable?
d. Sketch the magnitude response │H(j)│of the filter
e. Realize the block diagram with the canonical form.

December 2021 1