Professional Documents
Culture Documents
Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
room: SE5.119
mekyska@feec.vutbr.cz
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 1 / 98
Course Overview
Course Organization
Attendance
12 weeks of lectures + numerical exercises
absence is allowed, but not recommended:-)
Evaluation
12 pt: full attendance
28 pt: small exam (closed book) in the 13th week
60 pt: final exam (open book) in January
at least 50 pt required to pass the course
Literature
lecture notes
slides for numerical exercises print them before each lecture
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 2 / 98
Course Overview
Books
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 3 / 98
Course Overview
Course Outline
Fundamentals of Signals
Fundamentals of Systems
Fourier Series
Continuous-time Fourier Transform
Discrete-time Fourier Transform
Sampling Theorem
The z-Transform
Modulation
Filtering
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 4 / 98
Introduction
Examples of Signals
0.4
0.2
0
s(t )
-0.2
-0.4
0 0.5 1 1.5 2 2.5 3 3.5
t [s]
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 5 / 98
Introduction
What is a Signal?
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 6 / 98
Introduction
Signal Classification
Continuous-time signal
x (t), where t is a real-valued variable denoting time, i.e., t R
we use parenthesis () to denote a continuous-time signal
1
0
x (t )
1
0 1 2 3 4
t
Discrete-time signal
x [n], where n is an integer-valued variable denoting the discrete
samples of time, i.e., n Z
we use square brackets [] to denote a discrete-time signal
under the definition of a discrete-time signal, x [1.5] is not defined, for
example 0.5
x [n]
0
0.5
0 20 40 60 80
n
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 8 / 98
Introduction
0 T 2T 3T 4T t 0 T 2T 3T 4T t
digital signal
J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Introduction to Signals 9 / 98
Introduction
Exercise Block 1
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 10 / 98
Signals Signal classification
Signal classification
Basic classification: signals
with discrete time f (x ), y (t) functions
I continuous
I impulse
with discrete time f [n], h[m] series
1
1 2 3 4
1
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 11 / 98
Signals Category of sets
Set Classification
finite
e.g. {0, 2, 4, 6} or a set of all apples in a basket
infinite
I countable, e.g. N, Z, Q
I uncountable, e.g. R
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 12 / 98
Signals Signals definition
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 13 / 98
Signals Signals definition
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 14 / 98
Signals Signals definition
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 15 / 98
Signals Basic operations of signals
Operator notation
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 16 / 98
Signals Basic operations of signals
Time shift
Exercise 1.6. We have given discrete signal s[n] = {3, 1, 2}. Next, we
0
have its derived signal r [n] = T4 s[n] = s[n 4]. Draw signals s[n] and r [n].
3
4 3 2 1 1 2 3 4 5 6
4 3 2 1 1 2 3 4 5 6
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 17 / 98
Signals Basic operations of signals
Exercise 1.7. Draw signal with continuous time z(t) = D1/2 x (t) = x ( 2t ),
t + 1 for t (1, 0]
where x (t) = 12 t + 1 for t (0, 2]
0 for other t.
1 1
3 2 1 1 2 3 4 5 6 3 2 1 1 2 3 4 5 6
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 18 / 98
Signals Basic operations of signals
3
2
1
3 2 1 1 2 3
Exercise 1.10. (HW) Draw g[n] = h[2 n], where h[n] is from Ex. 1.8.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 19 / 98
Signals Basic operations of signals
Combination of operations
x (bt a) = (Db Ta x )(t) = (Ta/b Db x )(t) (derive as HW) (1.1)
Exercise 1.11. Draw signal y (t) = x (3t 6), where x (t) was defined in
Ex. 1.7.
x (t)
2 1 1 2 3 4 5 6 7 8 9 t
2 1 1 2 3 4 5 6 7 8 9
2 1 1 2 3 4 5 6 7 8 9
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 20 / 98
Signals Basic operations of signals
Amplification / attenuation
Exercise 1.12. (HW) We have x (t) = sin t. Draw signal y (t) = 2x (t).
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 21 / 98
Signals Basic operations of signals
Exercises
Exercise 1.13. (HW) Rewrite using operator notation (and proof):
f (4t + 5) =
f [4 n] =
cos(2t ) =
3
5 sin[2n 4 ] =
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 22 / 98
Signals Basic operations of signals
In the case of product, we work analogically. E.g. see picture from Matlab
in eLearning.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 23 / 98
Signals Basic operations of signals
Modulation
Influencing the parameter of one signal using value of another one,
e.g. simple amplitude modulation
s(t)
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 24 / 98
Signals Basic operations of signals
Convolution
We denote (discrete linear) convolution of signals x [n] and h[n] as
and define as
X
y [m] = x [n] h[m n], m Z. (particular samples)
n=
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 25 / 98
Signals Basic operations of signals
Convolution
Exercise 1.16. Calculate convolution z[n] = s[n] p[n], where:
s[0] = 3, s[1] = 1, s[2] = 2; p[1] = 1, p[2] = 1.
Result: z[1] = 3, z[2] = 2, z[3] = 1, z[4] = 2.
You can try to calculate z[n] = p[n] s[n]. Do you observe any change in result? How is this
property called?
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 26 / 98
Signals Basic operations of signals
Statement 1.18. Note the first non-zero sample of signal x [n] as MS and
the last one MF , similarly in the case of h[n] using NS and NF . Then, their
convolution y [m] has non-zero values in range
MS + NS m MF + NF .
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 27 / 98
Signals Basic operations of signals
Convolution
Exercise 1.19. (HW) Calculate convolution of signals x [n] and h[n],
defined as: x [0] = 1, x [1] = 1, x [2] = 0, x [3] = 1; h[3] = 2, h[4] = 5.
Exercise 1.20. (HW) Calculate convolution of signals x [n] and h[n],
defined as: x [50] = 1, x [51] = 8; h[0] = 2, h[1] = 2.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 28 / 98
Signals Basic operations of signals
Convolution applications
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 29 / 98
Systems
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 1 30 / 98
Systems
Exercise Block 2
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 31 / 98
Complex numbers
Complex numbers
1
imaginary unit j = 1; j2 = 1; j = j
complex number in the Cartesian form: x = + j
real part Re(x ) = , imaginary Im(x ) =
complex conjugate: x = j
Complex numbers
p
modulus of complex number: |x | = 2 + 2
phase/argument of complex number: arg(x ) = arctan for 6= 0
Simplification: since the results is within the range
2 , 2 , we dont consider sign of
0 for x > 0
phase of real number: arg x = for x < 0
undefined for x = 0.
Exercise 3.2. Calculate and draw modulus and phase of numbers from
exercise 3.1.
What are the relations between modulus and phase of complex conjugate numbers?
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 33 / 98
Complex numbers
Complex numbers
Reminder:
Complex numbers
Euler formula: ejy = cos y + j sin y for y R
exponential function in C: ex +jy = ex (cos y + j sin y ) for x , y R
1
Exercise 3.3. Draw in complex plane numbers d1 = ej 4 and d2 = e 2 +j 4 .
What are the modules of these numbers?
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 35 / 98
Complex numbers
Complex numbers
Statement 3.4. ejy = 1 for arbitrary y R.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 36 / 98
Complex numbers
Complex numbers
ejy + ejy
cos y = (3.1)
2
ejy ejy
sin y = (3.2)
2j
x = r ej = |x | ej arg(x ) (3.3)
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 37 / 98
Complex numbers
Complex numbers
Exercise 3.5. Find formula of complex number transformation into the
polar form.
Exercise 3.6. Express number x = 1 + 3j in polar form.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 38 / 98
Harmonic oscillation
A. . .
. . .
2
I T = ...
I f = 2 . . .
. . .
cos t cos(t + 3 )
t t
cos(2t + 3 ) 1
4 cos(2t + 3 )
t t
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 40 / 98
Harmonic oscillation
Harmonic oscillation
Exercise 4.2. (HW) Draw signal h(t) = 2 cos(t + 4 ).
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 41 / 98
Harmonic oscillation
we call c and c . . . . . . . . . . . . . . . . .
|c| =
|c | =
arg c =
arg c =
c=
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 42 / 98
Harmonic oscillation
Im
Re
http://www.utko.feec.vutbr.cz/~rajmic/applets/FazorovyDiagram.html
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 43 / 98
Harmonic oscillation
A cos(t + ) (4.6)
= c ejt + c ejt (4.7)
= a cos t + b sin t (4.8)
when
a = A cos c = 21 (a jb)
b = A sin A = 2 |c| = a2 + b 2
A j b
c= 2e = arctan a .
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 2 45 / 98
Harmonic oscillation
Exercise Block 3
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 46 / 98
Periodic signals
Periodic signals
One possible signal classification:
aperiodic
periodic
I we can find T > 0, so that x (t + T ) = x (t) for all t R
I we can find N N, so that x [n + N] = x [n] for all n Z
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 47 / 98
Periodic signals
1 RT
average value: Ia = T f (t) dt
s 0
1 RT
effective value: Ief = T f 2 (t) dt
0
3 2 1 1 2 3 4
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 48 / 98
Fourier series
Integration recommendation
Integration of selected functions:
cos kt dt = sink
kt
R t
e dt = et
R R
1 dt = t
t n+1 ejkt
R n R cos kt
R jkt
t dt = n+1 sin kt dt = k e dt = jk
Linearity: R R
f (t)dt = f (t)dt
R R R
[f (t) + g(t)]dt = f (t)dt + g(t)dt
Division of integration interval:
Z b Z p Z b
f (t)dt = f (t)dt + f (t)dt, where a p b
a a p
Spectral analysis:
periodic signal f (t), t R coeff. of Four. series {ck }, k Z
Spectral synthesis:
periodic signal f(t), t R coeff. of Four. series {ck }, k Z
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 50 / 98
Fourier series
complex
X X
ck e jkt = c0 + ck e jkt + ck e jkt
f (t) = (6.2)
k= k=1
goniometric
a0 X
f (t) = + [ak cos(kt) + bk sin(kt)] , (6.3)
2
k=1
bk
p Ak
where: ck = 12 (ak jbk ), Ak = 2 |ck | = ak2 + bk2 , ck = 2
ejk , k = arctan ak
.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 51 / 98
Fourier series
b0 = 0 always
c0 R, ak , bk R
f (t) is even bk = 0 for k N ck R for k Z
f (t) is odd ak = 0 for k N {0} ck purely imaginary, k Z
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 52 / 98
Fourier series
T t
T T X 1 2kt
Resulting FS: f (t) = sin .
2 k=1 k T
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 53 / 98
Fourier series
1 1 1
K=0 K=1 K=2
s0(t)=1/2
0 0 0
0 2 0 2 0 2
1 1 1
K=3 K=10 K=20
0 0 0
0 2 0 2 0 2
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 54 / 98
Fourier series
Magnitude spectrum of per. signal {|ck |}kZ says how intensive are particular
harmonic components, while phase spectrum {arg ck }kZ describe their phases.
It holds that ck = |ck | ej argck .
5 4 3 2 1 1 2 3 4
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 55 / 98
Fourier series
t
- 2
2
3 5
2 2
2
Solution: signal is even, so bk = 0. It can be proofed that a0 = , a1 = 12 , for
k
(1) 2 1 2
k > 1 odd are ak = 0 and for k even it holds that ak = 2
k 1 . Resulting FS is
1 1 2 2 2
f (t) = + cos t + cos 2t cos 4t + cos 6t . . .
2 3 15 35
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 56 / 98
Fourier series
1.2 1 1.2
K=1
1
1
K=2
K=0
0.8
0.8
0.5
0.6
0.6
0.4
0.4
0.2
0
0.2
0
0
0.2
1 1 0.5
K=4 K=6
0.8 0.8 0.4
0 0 0
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 57 / 98
Fourier series
RT
ck (y ) = 1
T 0 y (t)ejkt dt =
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 58 / 98
Fourier series
2
10 8 6 4 2 0 2 4 6
2
2
10 8 6 4 2 0 2 4 6
2
2
10 8 6 4 2 0 2 4 6
1
1
10 8 6 4 2 0 2 4 6
http://www.utko.feec.vutbr.cz/~rajmic/applets/KrizKorelace.html.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 60 / 98
Correlation function and its relation to FS Cross-correlation function
Examples
Exercise 7.1. Calculate cross-correlation
function of 2-periodic signals
1 for t [0, )
x (t) and y (t), where x (t) = , y (t) = sin t.
1 for t [, 2)
Result: Rx ,y ( ) = 2 cos
(using sin( + ) = sin cos + cos sin ).
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 3 61 / 98
Correlation function and its relation to FS Cross-correlation function
Exercise Block 4
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 62 / 98
Autocorrelation function
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 63 / 98
Autocorrelation function
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 64 / 98
Autocorrelation function
Properties
Rf ( ) is T -periodic.
Rf ( ) is even function.
Rf (0) |Rf ( )| for all normalized autocorr. function
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 65 / 98
Autocorrelation function
Exercises
Exercise 8.1. Calculate effective value of signal
s(t) = 2 cos(t) + cos(2t + 3 ).
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 66 / 98
Autocorrelation function
Exercises
Exercise 8.2. (HW) Calculate and draw magnitude and phase spectrum
of signal s(t) from Exercise 8.1.
|ck |
3 2 1 1 2 3 k
arg ck
3 2 1 1 2 3 k
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 67 / 98
Function sinc
Function sinc
Definition:
sin(t) for t 6= 0
t
sinc(t) = (9.1)
1 for t = 0.
Properties:
sinc(t) is even
Ra jt Ra
e dt = cos(t) dt = 2a sinc(a)
a a
R
sinc(t) dt =
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 68 / 98
Function sinc
Function sinc
Exercise 9.1. Draw functions sinc(t) and sinc(t). Mark important
values on both axis.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 69 / 98
Back to Fourier series
T t
T2 2
2
T
2
T
k
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 70 / 98
Back to Fourier series
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 71 / 98
Back to Fourier series
2
1
6 5 4 3 2 1 1 2 3 4 5 6 7 8 t
Solution:
P 4
FS: s(t) = 3
sinc( ) + 29 sinc( ) ejk5 ejkt , where = 2
9
.
k=
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 72 / 98
Fourier transform
in time domain:
in frequency domain:
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 73 / 98
Fourier transform Definition and properties
Definition
+
[ .
Notation: g(t) = G()
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 74 / 98
Fourier transform Definition and properties
Properties of FT
Note RS () = Re{S()} and XS () = Im{S()}, therefore
S() = RS () + j XS (). Then for real signals it holds that:
1
Z
RS () = s(t) cos t dt (11.4)
Z
XS () = s(t) sin t dt (11.5)
2
RS () = RS () (11.6)
XS () = XS () (11.7)
S() = [S()] (11.8)
2 cos
2
Solution: F () = 1 2
for 6= 1, F (1) = 2
f (t), F ()
2
t,
-9 -7 -5 -3 - 2
2
3 5 7 9
(two pictures in one, vertical axis are equal, horizontal different: time and frequency resp.)
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 76 / 98
Fourier transform Exercises
Properties of FT continuation
For definition of the other properties we will use notation s(t)
d = S(). It
holds that:
\
s(t ) = S() ej or T\ d j
s(t) = s(t) e
[ = 1 1
s(bt) |b| S( b ) or D\
b s(t) = |b|
D 1 S()
b
\
s(bt ) = 1
|b| S( b ) ej or Db\
T s(t) = 1
|b|
D 1 S()
b
ej
for , a R, a 6= 0.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 77 / 98
Fourier transform Exercises
8 t 8
f , 4F (4) t
Re f \
4 4 +1
t
Im f \
4 +1 t
f 4 +1
1 t, 1 t,
2 3 2 2 4 3
4 4
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 78 / 98
Fourier transform Exercises
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 4 79 / 98
Fourier transform Exercises
Exercise Block 5
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 80 / 98
Fourier transform Fourier transform reminders
Reminders of FT
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 81 / 98
Sampling of signal with continuous time
|S()|
s s
|S()|
s s
0.4
0.2
0.2
0.4
0.6
0.8
1
0 2 4 6 8 10 12 14 16
samples
sampling frequency 22,05 kHz
One of consequences: If the signal contains both frequencies (13 and 9,05 kHz) before sampling,
then, during the reconstruction, we will get only 9,05 kHz with incorrect amplitude and phase.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 83 / 98
Sampling of signal with continuous time
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 84 / 98
Analysis of discrete systems
0,1 z 1 0,5
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 85 / 98
Analysis of discrete systems
Exercise 14.3. (HW) Calculate first five samples of the systems output
(Ex. 14.1) for input signal x [2] = 5, x [3] = 0, x [4] = 2.
Additional questions:
What will be the output if we send to the input sine signal with specific frequency? (you can use Matlab)
Can you experimentally determine if the system from Ex. 14.1 is high-pass filter? How would you do it?
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 86 / 98
Analysis of discrete systems Z transformation
Exercise 14.4. Find the transfer function of system from Exercise 14.1
and calculate its zeros and poles.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 87 / 98
Analysis of discrete systems FIR filtering of random process
2
2
0
1
0 0
5
1
2
2
4 10
3
0 100 200 300 400 500 600 700 800 900 1000
Magnitude (dB)
0 100 200 300 400 500 600 700 800 900 1000
n n
80
60 20
magnitude
magnitude
60
40
25 40
20
20
30
0 0
0 100 200 300 400 500 600 700 800 900 1000 0 5 10 15 20 0 100 200 300 400 500 600 700 800 900 1000
f Frequency (kHz) f
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 88 / 98
Discrete Fourier transform
DFT
Definition of DFT for signal s[n] with length N:
N1
2
s[n]ejk N n ,
X
S[k] = RN [k] kZ (15.1)
n=0
Exercise 15.1. (HW) Calculate DFT of signal s[n], where
s[0] = s[1] = 1, s[2] = s[3] = 0.
Solution: S[0] = 2, S[1] = 1 j, S[2] = 0, S[3] = 1 + j
2
n o
Pozn.: How do the series of numbers ejk N n look like? They are
complex discrete harmonic signals; e.g. for N = 8:
2
k = 0 : ej0 8
n
{1, 1, 1, 1, 1, 1, 1, 1}
j1 2 3 5 3 7
k=1: e 8
n
{1, ej 4 , ej 2 , ej 4 , ej , ej 4 , ej 2 , ej 4 }
..
.
2 7 3 5 3
k = 7 : ej7 8
n
{1, ej 4 , ej 2 , ej 4 , ej , ej 4 , ej 2 , ej 4 } .
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 89 / 98
Some useful properties
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 5 90 / 98
Some useful properties
Exercise Block 6
Pavel Rajmic
translation and further edit: Ji Mekyska
Department of Telecommunications
Faculty of Electrical Engineering and Communication
Brno University of Technology
-
rajmic@feec.vutbr.cz
mekyska@feec.vutbr.cz
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 91 / 98
DFT reminders
DFT reminders
Definition:
N1
2
s[n]ejk N n ,
X
S[k] = RN [k] k Z. (*15.1)
n=0
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 92 / 98
Circular convolution
Modulo operation
z = modx y
m
z is remainder after integer division of number y by x
m
L Z : Lx + z = y
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 93 / 98
Circular convolution
Modulo operation
Exercise 18.2. Draw signal r [n] = x [mod4 n] for |n| 4, where x [n] is
n 0 1 2 3
defined by table: .
x [n] 1 5 1 2
Solution:
x [n]
4
3
2
1
4 3 2 1 1 1 2 3 4 n
r [n]
4
3
2
1
4 3 2 1 1 1 2 3 4 n
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 94 / 98
Circular convolution
Circular convolution
(Discrete) circular convolution of signals x [n] and h[n], both equally
defined in time samples 0, . . . , N 1, is denoted as
and defined as
N1
X
z[m] = RN [m] x [n] h[modN (m n)], m Z. (particular samples)
n=0
(18.2)
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 95 / 98
Circular convolution
N1
X
z[m] = RN [m] x [n] h[modN (m n)], m Z.
n=0
h[2] h[1] h[0] h[7] h[6] h[5] h[4] h[3] h[2] h[1] h[0] h[7] h[6] h[5] h[4]
m = 0:
h[3] h[2] h[1] h[0] h[7] h[6] h[5] h[4] h[3] h[2] h[1] h[0] h[7] h[6] h[5]
m = 1:
..
.
h[1] h[0] h[7] h[6] h[5] h[4] h[3] h[2] h[1] h[0] h[7] h[6] h[5] h[4] h[3]
m = 7:
See applet http://www.utko.feec.vutbr.cz/~rajmic/applets/KruhKonvoluce.html
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 96 / 98
Circular convolution
Circular convolution
Exercise 18.3. (HW) Calculate circular convolution of discrete signals
x [n] and h[n], defined as:
x [0] = 2, x [1] = 1, x [2] = 0, x [3] = 2;
h[0] = 1, h[1] = 2, h[2] = 3, h[3] = 4.
Use can use table as in the case of linear convolution, but you must make h periodical.
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 97 / 98
Circular convolution
Fast convolution
So called fast convolution is in fact calculation of circular convolution
(18.2) using FFT. Calculation: ifft(fft(x).*fft(h)).
P. Rajmic, J. Mekyska (DT, FEEC, BUT) CASS Signals and Systems Analysis Exercise Block 6 98 / 98