0% found this document useful (0 votes)
55 views78 pages4&5unit Signal
Signal and system
Uploaded by
SATENDRA KUMAR KUSHWAHACopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
55 views78 pages4&5unit Signal
Signal and system
Uploaded by
SATENDRA KUMAR KUSHWAHACopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 78
Q.1. What is convolution ? What is the relationship between correlation
ad convolution ? (R.G.P.V., June 2009)
Ans. Convolution — Convolution is a mathematical operation which is
used to express the input output relationship in linear time invariant system.
Mathematically, the convolution of two time functions x,(t) and x(t) may be
defined by the following expressions —
t
x(t) @xg(t) = Jui@x2tt —1)dt
The symbol @ is used to represent the convolution of two time functions.
Relationship between Correlation and Convolution — The correlation
function is used to indicate the degree of similarity between two signals. Higher
the value of correlation function is more the degree of similarity.
There are two types of correlation function —
(i) Auto-correlation function (ii) Cross-correlation function.
vojer auto-correlation function is denoted by R(t) and it is defined as
R(q)= fx(t).x(t—tde if x(0) is real valued
| com Now, if we compare the convolution equation with the definition of auto
lation stated above, we concluded that the process of autocorrelation is
ve ”
“Similar to that of convolution.
22. Describe the response of LTI systems to arbitrary inputs.
SI Or ;
nun that the response y(n) of the LTI system is a convolution sum of the
ignal x(n) and the unit impulse response h(n). (R.G.P.LV., Dec. 2014) sample We den, reeatlaeaats
Seuence at ne Pons of Me
~ Se
she cei Kena, Specig” PyeaTt ‘analyse of Discrete Time Signals 145
sk) = ° :
the toca, ens 6 Samp eo qi (iv) reduces 10
an a tion of ett ig yg) 8 ten it
scaled outpan k= X(®), the timp the ting c 7 we - Sonn -k) (wit)
the rege else. tg yi xt y
Ponse ge the jn ok . we
f the «Duis LTT system is completely characterized
Finally, i¢ 9, #0 W) = the gy MPU tk the relaxes .
of wel ifthe inp se rs ten Me ver namely its FSP se 10 the unit sample sequence
the ari? © the 4H ti® 1 chat zation of the output of a time variant,
» that is, ALY signal 0 itt nite number ‘of unit sample response functions,
~ XO) tha ines je delay:
tn x(n) = x x is ex or on (vi) that gives the response y(n) ofthe LTT system
om the response of ir (k)3(n_, eae ut signal s(n) and the unit sample (impulse) response
"DUIS, that is, the system to , y yield te cues 9) ‘We shall now explain the
oP the response Y(1)» ‘mathematically and
(viii)
(a) =
TIx(n] = tf 3
ia _ |
= Lawrie ny
kee
2H,
Clearly
systema ion (yy e that the
We 204s Kacy (UO from the form a product sequence ‘Tre ouput 00 Laie
system to a ot pe Mposition ope POSiti ‘of the product sequen ynce hiay = Bi
both xa) iy ea as sum tion open nfh folding h(k) about k = 0 (the Oe org. roa in
It ‘Press; is thet
00 seq ion fc INC). The folded sequence is 0 :
system bon oor etiving 8 ora) This x th espana ing comping the onvuson Ne apy and Mk involves Be
GY) applies oan time {DE EAation yy AE 0 the esi i funn wing four steps — qo duin¥®
I the Property, ped the I impulses a1) |. Folding ~ Fold 1p about k= 010 0528 7
Considerably er, Si Cimesvarag ys inary propery fu |i) Sting = shift he #) by 80 he ight (et ifm 8 POSES
Seauence &(n) is denon ia, ) system, ion in equate [peative), to obtain b(ng ~ ¥-
th * wo a ot he in equation (i (ii) Multiplication - M017 xy! yoy Hoobs PONE
1en by the & h(n) 2 of (iv) sin a
a ues time-invarian) » 780) is system to the frend Vag (= ae = ave vaso product sequence
sequei Property, iv) Summation — SY
nce B(a — jg Mh re 0 oan valve of the out pa ingle time
Pons no(K) to obtain the val st a single
hen - b= 71% SOF the 5 0D tis teense of SIRE espns of
a~ Wy Ystem to the delayed This procedure results in om in vai the response of
felayed stant, say n= No. In gener we en ely, se 2 OH
system over time instants = © ile time shifts ~ °° <" a
Te repeated, forall
yi) 146 Signals and Systems (Sem, Ex .
0.3. What are the various ahem
Ans, There ae following properties gp cm
(i) Commutative Property — Fry betes,
is that its a commutative operation, thay jt PBC py,
xto}* Nal = ha)» gy
This can be proved ina strightfony,
x(a} * Ala) = Yxk)hin—ky =
=
Therefore, It is seen that with this substi
X{o} and h(n} are inerchanged. According le,
LT system with input x{n) and unit samy Tesponge te
cuxput ofan LT system with input h(a} and un samp is x
(i) Associative Property ~ Second usaf Prope
is that it is associative, that is, Petty og
Xlo} * Qh) * bala) = (xin «by tny » yn
This propery is proven by a staightforwang
summations involved. Interpretation of the associative
figs. 4.1(@) and (b). Systems shown in these block diagrams «
withthe indicated unit sample responses. Thig Pictorial rp
panicularly convenient way to represent LT] Systems in bec
‘also reemphasizes the fact thatthe impulse response ofan
‘characterizes its behaviour,
In fig. 4.1(@),
Yio} = win] * hyfn) = (x{n} bla) * bn)
In fig 4.100,
ylo) = x{n} * hfn} = x{nj * (bila) * b,fop
1 ne asociaive propery, the series interconnection ft
in fig. 4.1(a) is equivalent to the
ln] = x{a] + byl] * hain}
eAambiguous. That is, according to equay i, it does not
what order we convolve thee ‘igual
wm
‘Analysis of Discrete Time Signals 147
=> sim
nial = Bylo] bya]
(b)
biol = hylal« y(n) sia}
stl
1101
©
via)
ayo =f Sie) => oe
fl
@
Property of Convolution and the Implication of this
iatve
sil ¢ Property for the Series Inerconnection of LTI Systems
Comm
: a
sctributive Property ~ Third property of convolution is
Dials that convolution dsaibutes over addition so tht
{a} * Qylo) + hgfn}) = x(n] * hyln) + xfa) * hylan) ...(w)
1
(i)
also be proved ina straightforward manner. Distributive property
iso
erpretation. Consider two LTI systems in parallel, as shown in
te men hy {a} and hy{n] have same inputs and their outputs
A
sila)
Oe via} sme in
ale]
(@ )
‘Interpretation of the Distributive Property of Convolution for a
Parallel Interconnection of LTI Systems
Yale] = x(n} * hytn}
Yaln] = x{n] * hf} 148 Signals and Systems (Il-Sem,, Ex.
the system of fig. 4.2(a) has output —
yl] = x{n} * h
8 ‘Analysis of Discrete Time Signals 149
shown ote of the weighed pulses i
em bave Pre qutput 9(t) coresponding to X(t) is the
as n
Wi
111+ atgy
i ‘What remains is to take the limit of
Corresponding to the righthand side g¢ a Af gs ee a a ie
system of fig, 4.2(b) has ~ Nation ee a Oe ee med ee be ple
yl = xin] * (hytny 4 balay le a ih he gotwal ee unit impulse as & — 0, the response
Corresponding to the left-hand Side of gay... ea? L nds 10 8 uch an impulse in the limit. Thus, if h(t)
ofthe dributv propery ofconvluton, apy Thee Beno respons fo the unit impulse &(-1) located at time t,
can be replaced by a single LT system, Un want He ie mg aim
sum ofthe individual unit sample responses Te spon ) peo
Patty ca} (ii)
O4. Mate Younnderstandy drama sin, SII
mn? s) = 50 ye
Or Rept eee hand side becomes an integral because of & —> 0,
Define discrete convolution and state i pipe, pit ic in fig. 43. Hence, o
’ wee (iv)
Ans. Convolution isa mathematical operation eq YJ (= Lona 0 the one for
Convolution is important in DSP (digital sana ea oma, x ono uation Go) i exacly anaogous
twosequnes inthe time domain sequvalenio mag ier] t ad) cam De denoted as
the frequency domain, Convolution finds Ping thea, Any iP
iS application in roca
sore
erpesefe
especially analysing the output of a sy
stem. Consider the
and x,(n) and the convolution of these
2 [ xpat- Dat
: x=
two signals can be.
jnuuitively think of x(0) as a “sum” of weighted shifted
can int
With this
pee) otis o€ AP Natt on the impulse &(t-t) is x(t}dt
. s eS,
y jes, where the {iv simply denotes the superposition ofthe ron :
7 Eetntew ie and, by linearity, the weight on the response hy
inputs, and,
sow a ees also x(a).
Properties of Discrete Convolution ~ Refer he aso 3
5. Whatis convolution integral? Explain the properties
time convolution,
Ans, An arbitrary continuous-time signal could be expressed ate
form of a linear combination of shifted pulses —
ied impulse xi)
+
= lis ~kA)A
x(t) im, ZrkaBac kA).
Lets define f(t) asthe response ofan LTI system tothe inpt 8
‘Then, from equation () and the superposition property of linear sts.
sto)hyind
x)
y= im Dxckayiyg (va
ie i
Where x(t) and y(t) are the input and output of a linear system. Inf
il s
'nput x(t) and its approximation %(t) have been depicted while infes+ (b) 1 ‘Analysis of Discrete Time Signals 151
toa unit impulse. Sy
. enti givens ‘mbelically the comvotution of two signals x(t)
a UT yan) = x00) * AD)
ite te sae symbol * is used ty
gisereretime and continuous-time Myon
ote content will generally pe
fe Mistinguish the (Wo cases,
ct COMIN has he
(vii)
shady 0d ities discussed for discrete-time
| 9 PT Continuous-time convolution
as > murative, associative, ang 2 + ti8
’ oui ‘That is, Fig. 44
; pf) xD * HY = h(t) * x(t)
rr 7 Gi) 10 * THO * HOI = bay) 5 - (Commutativity)
. f ) x(t) * Ly( + byl) = faq = D ___ (associatv
5 io Gi OT MOs ty eae
iven LTI system, determine (distributivit
> gs Foragen Jormula for convolutinn na
“Sp.
1
fi T
© ‘ (Rory, .
tw 0 Let us assume an LTI system which is *V, Dec. 2017)
" spenbeinput to the system isan impulse, then the gyore su) Felaxed at t = 0,
juoyTas is known as impulse response of the « "stem output is represented
> BO =T BO) "stem and is represented as
ny arbitrary signal x(t) is expressed as i)
° | 20 = [8-0 .
Fig. 4.3 Graphical Interpretation of the Response of a nh OCT i “©
Continous-time Linear System RO]
Equation (iv) shows the general form of the response of a linear systa bal 1f X(1)8(t~ adr
inconinuos tine. finaditon to being linear the system s also tme inva |e near system, on ~Aitl)
then hy) = hgt-t) Then unit impulse response ht) can be defined as ~ c
; A(t) =ho(t) YO=]" xQ TBA
fie, 8() - b(O) In this case, equation (iv) becomes as — We sytem response bocaue me ‘Okie wAiv)
Wt = fis more “if teause of ~~ impulse is giv pee 54) ish(), then the system
; ae . IL D=T [8
ine known as the convolution integral or the superpulit} Fon equation (iy), we peels 9
im continuous-time counterpart of the convolution sum i!
"© arepesenation of a continuous-time LTT system in te ¥0= f° [email protected] EX - Branch)
‘systems (m-Sem.
152 Sonat and yea by tsec ise (Othe py
“The output because of input delay
for a LTI system,
a nawo=
ation (v), We Bet
es y= f_ xcpnct= Oat
ntegral or simply convolution,
causal ifand only if its impulse roy,
na-d
This is known as convolution it
(0.7. Prove that an LTI system is
‘Ans. Ifthe response of
subsequent values of the exci
‘A causal system as one WI
and past inputs but does not depends on
time invariant system having an output at
equation in the form as —
tation, then it is known as causal filter,
hose output at time n depends only on py,
future inputs. Let We consider gj,
ying) = J n(K)x(n9 —K) al
faa
Suppose we subdivide the sum into terms such as first set ino
present and past values of the input, i.e., x(n) for
involving future values of the input, ic., x(n) for n > no. Thus we obtain
above expression in the form as —
2 =1
yltg) = Sb¢K)x(ng Kk) + SYR —K) 8
0 kee
= [h(O)x(ng) + H(L)x(ng-1) + h(2)x(ng-2) + ...]
+ [h(-I)x(ngtl) + (2) x(ng#2) +
Now we observe that the terms in the first sum involve input sii
component x(n), x(Ny-1), ... which are the present and past values of tie
input signal. The terms in the second sum involve input signal compon:
x(nig+1), x(ng¢2), ... and it is clear that the impulse response of the systs
must satisfy the condition as —
h(n) = 0 forn <0
‘Because that h(n) is the response of the relaxed linear time invariant syste
toa unit impulse applied at n = 0 and it follows that response h(n) = 0 ft
1 < 0 is both are necessary and a sufficient condition for causality. There
a linear time invariant system is causal if and only if its impulse response is °
for negative values of n. Since for a causal system h(n) = 0 for n-< 0 then lit
limits on the summation of convolution formula may be modified from tis
restriction. This we can write the two equivalent forms of the convolul®
(i)
negative (RGBV, Dee
is values of M- Dee
at digital filter at a specific instant is independ fe
ti
time n =o impart by the convoy
nS ng and second wf hat the EAU
Aratyia of Discrete Time Signals 159
oe
yn) = Shkyxtn— ky
ra)
= Yxtyin— iy
ie
sal linear
he input to 9 cau ime invariant system sequence
’ (for <0) the limits onthe convoitn ermla we hates e
the two equivalent forms of the convolution formula obtain
Pe
wi)
x(n) =
y(n) = Yo b¢k)xin— ky
io (iy
= YL xtkyh(n- ic) |
k=O ‘visa
We observe from the equations (vii) and (viii) that in this case the lower
sits on the summation fortwo altemative forms ae ideal andthe upper
imc is growing with time. Since the response y(n) = 0 for n Ny
h(n) = 0, n <0 and n 2 Np.
In the case of frequency-domain,
Y(@) = H(@) X(o)
Y(k) = Y(@)y—mkn K=O 12
= X(@) HO) 4 -yuin k= % 12,
Y(k) = X(k) H(k)
1 +Np-1
ad a —_
54. Sigal and Sytos (Sem, EX - Branch) ea,
calculated as ~ qin) becomes Anais oy
Since, y(n) can be calcu p Ny Orson Tne
y(n) = IDFT [X(K) HOI (mM) = San Spr 155 {
The Noi sla onvoitin of X0 WHN HO) HM, mo alt "Wed Wy
" ‘ith hin) f
to the linear convolution of x(n) wit ae 4p ll") 04 Neng
Tete omc eg Neh 119 ENE rt
[Np respectively will pro on less than N, for implement, >| wpe 6h * th Yi
When) a8 0) ee, Ny 1 aN 1 228 ae ay Pye abe SHO agg A a
convolution using : their Ie toN, a
end of x and hn respectively 1 nee ae {Ye pp tana the ftovn 4 ya
o ore sie XK) ad KA) ae (i) Linear convolution (in wi aly
Paes eee epectively and they canbe defied, (i) Sectioned convoy," Creer try
Nel F , —
mWIN k= 0,1, (i) Linear convolution Rory
x= Yume , 2 ay AM ~ Retr te 1, hee
et atl i Wi rela contin - Rte gg 8 a7,
(ii) Sectioned convolution of ai
The
pn the impulse response B(n) of a digi)
pa —j2mk/N
x= LHP, k=012 0.
ing linear convolution for Fk ra
2
f length N and Nn
assume that x3(n) be another sequence oF ‘en its N, ; her. ¢;
a X;(K) which 3 product of X(k) and Xe. : employed an be vine Ping
X3(k) = XK) X20, k= 0, 1,2. N= 1 woes dls. Because he ameet
By taking inverse DFT of Xs(k) ie.
xy(m) = DFT [X3(6)]
is ws
2 LF xa09 PH = EY xo ao"
k=O
id these problems. The ite fete mane
ret a0 8. The overap aa domain, i
pe to perform fast (sectioned) convolution by eee methods are
k=O r
and the final convolut
no ._ Jcoput is computed by combining the tion
Et/ Ne} - yest Parial convol
! s Sa meemes | [Es (e aan ae cach block. In oveepadd me, ved ot es ae es
ko n=0 1-0 It lke overlap-save method, instead we overlap (M1) amps (vine
isthe number of samples in b(n) sequence) and are wed lr for adding
two values,
Inoverlap-save method, we overlap the input sections and discard those
lsat the output after circular convolution, that ae not valid and sve the
sl samples in each section,
MERE IR iankm-n-DIN
oo Fae Sef Ye
N n=O iO k=0 a
Now, we consider the term with in the brackets. 1t has the
for a=1
N-I
k N
y a” for azl Numerical Problews
ko I-a ; ten
where, a = oF cm--ON, when (m—n— is @ multiple of ; yee Compute the convolution y{n] = x[n) * h(n] of the following
vige aN value of a #0. ‘signals —
otherwise a = 1 for any Wa pem-—n+ PN ln) = auf, hfn] = ula), (RLG.PV, Dee. 2011)
Nay { -
Therefore, >: 0. otherwise J
y 156 Sor0s ana systems (-S2™ EX - Branch)
oo —
/ sop We know tlw sweet ,
= Yxkynin-) in making use of infinite YA of scray
ve as : wei kat te 165 ony Tin Signy 157
oo ‘ Cr
The product of x18) Ma pas alvays 000 ZF0 SMP oi i " tape ecm gti
messi . Oa i
yore Yxwna-0 y= Fangs | xt)
kee ene comet ouput of is
k © shown in fg. 45 ba
yor De
ken
yt 2 mimes!
k=O
sik] =a]
emia
The Outpat¥ (ny
Fig. 4.6 The Ouput yn)
prob2. Compute convolution yn) ofthe signals
sor|
ja" , -ISnss
0, elsewhere
ee,
ime sns4
0, elsewhere.
(RGPY, June 2010)
Sol Plot the sequence values ofthe two signal s shown ini, 47.
Fig. 45 : oH) a}
cing insane FNS
0 °
Jit = >
kere 20
azo.
324
Tae
Hence, y(n) takes a constant value for
or2sase
Again fora <0 the product of 0) = —k) has
Therefore, for n <0, the output yin) is
y(n)= yx) hin-k) = xy
non-zero sapies 18°"
Fig. 47 SOT
153 Sissons 70 (
The convolution
when 0
when
when
when
ex-- Branch)
fo site dratonsoenees 6 One a,
yo = Yxwr-
ran
=0 -
yor Dxconc®)
- 3) HO) + X= 2) BQ) + x
aol
Be MO
+ x(0) (0) + =
+o?
4!
y= Dxeond-0
i
ee x 3) A) +X 2) BG) + 26
4+ (0) BCL) + xC1) by
soieateai+i+a
n=2 -
Yona’)
ee
a crceees + M(=3) OS) 4 8(-2) bes
Fx) (3) + x0) B42) +
oad
y(2) =
eovat+atel+ara
cateateleare
n=3 -
yQ)= YDrtona-b)
“em (3) WOH) + ¥ 2) 6
Sxl) 4d) + X(0) BG) + XE) D
4 x(2) DCL) + x3) MO) +
eatelearare®
ned
yay= Yxdond-W)
—
eT ae 3) nC) + XD) AH) # ED BE) **
Fad hid) + x) HO) + XG) BD +2) 0)
a POs 0s 04 lsararat
elea+re+a+ae
WS of Discrete Time Signals 159
go.
ye DRenS-W)
eo
yO"
X-3) b) + X(-2) 7) + 2-1) (6)
+
¥(0) WS) + x(1) h(a) +x(2) b(3) + x3) h(2)
et O4040404 $ x4) BC) + (5) bo
Beso Tete nor
za
yo = DAWEH)
-
ewe + XCC3) BO) + X(-2) b(B) + x 1) AO) + 0) HG)
ay MS) + x2) WA) +308) 49) +24 82) a
2 HOF0F0F0F0 HPF Meares ec.
cot+a+at+oae 7
a=7
Dxwnd-»
=f
vee (3) HHO) + (2) (9) # XC)
8)
A) bo) + (2) b1S) + x) bd) + x04) WG) + eee
oO. +04040+0+a%sateaimoeatsa’
o=8 5
y= Lacon -b)
i
enue + (3) BUTI) + x(-2) (10) + x(-1) (9) + (0) (8)
“+ x(1) I(T) + x(2) (6) + x(3) b( 5) + x(4) h(4) + x(5) B13) +
= 4 040404040+0404 04a
=at+o
azo
2O= Yxckyho-k)
me
= +x C3) (12) + x 2) B (ID) + x I) B10) +O) HO)
“+ x(1) hh) +x 2) (1) + x (3) (6) + x4) (5) + x15) BA)
=4040404040+0+040+0 = 0 systems (Sem EX - Branch)
160. Signals and
} when n="! -
y= Yxdonet-)
rail
Sy x3) WO) + ADHD 4 x1,
se art oe ",
when 9=—? .
yee Lxahe?— k)
fa
et NES) HDD + XC2) NO) = ga
ty
phen yoo= Lx n(-3-k)
Te NO) + =e?
smese sequence values are drawn in AB: 48.
teen %
j wy sts Be sy
Pegees2
28 aa
2 ett he s
a lle
Beh le le [eye
2 PAEEEER
sage le lela fe
abl EEL
B72 15 [6 |= |e
+f]? s
fe
fl
=
% eels
ees
%
Fig. 48
a iaat tl
4. The impulse response of a linear timerinvarian’
a yD cof the system tothe inp
a) = (12,1, B}- Determine the response
x)=U, 2,3,1). (RG.PY, Det
Sol Convolution sum can be expressed aS —
yin)= x(k) h(n k)
Kone
Anatysis of
the given sequence, as Oscrote
prot ™ shown in fp 49 Tie Signa 10)
aon Fig. 49
Ny=4 12 = 4, thus Ny + a
tien = 0, N*M Ie as gy ey
yO = Lxtk) nk) y
+ x(0) HO) + x(1) hr)
21x2+2x14040=4 ) + x12) bt-2) + 103) bi-3) «
When n= 1, ia)
y= x(k) h(l-k)
Pa
= san + X(0) HCL) + (1) 10)
+ x(2) C1) + x(3) he-2) +...
ale 1l+2x24+3x14+0=8
When n= 2,
yQ2)= x(k) h(2-k)
. + x(0) hQ2)
+ x(1) h(1) + x(2) h)
+ x(3) HEI) +
= 1x (1) 42x 143x241 x1=8
When n= 3,
¥@) = Yx(k) hG-k)
k=-c0
4+ x(0) (3) + x(1) h(2) + x2) H(A) + x03) bO) +
042% Cl) +3xleix2=3 —
162. Signals and Systems (l-Sem
San
EX - Branch)
When n= 4, we kne
yd) = yond)
=F xq) na) +0)
coro dxCDtixl
‘When n= 5,
yo DE n6-K)
5 gq) m5) + CH) + x2) BB) + 20) HQ)
aor 0+0+ 1x C= .
1h) # x) BQ) + x) by,
=
poe l
‘Analysis of Discrete Time Signals
ow that, ime Stats 169
Lxcohin- iy
y(n)
porn =O
(0) = Sxtkyh(-k) = -5
ken
yl) = YxGeyhd- ky) =-8
ke
When n=l, pean YO" YxGn2-K=7
- k=
yed= Yxdoncl-# «
ay yet) +a) M2) +X) +30) H4 ganz Y= ZadonO-W=-s
x1+0+0+0= e
sequence values ae plod in 88 pened y= DxWOHA-b)=-7
; 14.883 kw
Prob. Convolve graphically feneS, Y= YxWQh-b)=-4
xn) = oT et
kn = Cl 2 3 4} forne- he yl Yxh-1-k)=-3
Pat
oe ah forn=-2, ye2) = Yxten-2-W
w= 34 ‘The given sequences canbe solved by using graphical method as shown
in fig. 412
For n=0,
‘The given sequences can be drawn in fig. 4.11
na)
Fig. 4.11
nek) 0) = a0) A)
3
Product Sequence,
ye be dp Peed 28 ‘Analysis of Discrate Time Signals 165
114) tk) we)
4 A
Product Sequence, =
nets *
12) =1008 2-4)
6) x0) i
e
Product Sequen
y)eedededs
|
Forn=3,
9) = 41k) 20-4) rel
Blot ky eoSnittat-b)
tw ef by one and Systems (Sem, EX - Branch)
‘Analysis of Discrete
* Time Signals 167
5, Compute the convoluto
11h oh, Hh and inverse wanton PE fia by
im
le) =e wl)
hy) = we.
Sons
Forn=-2
nab)
yea) =x) 2 =k)
1, convolution Property states that (RG.PY., Dee, 2011)
x(t) * HO) © XG), Ho)
Product Sequence,
yeay==1 qhuss
1
(2+ jo)?
sioilaly HAni)| = HGw) = —
4+ jo
Fix) = XGo) =
u 1
2+ jo)? 4+ jo) ih
By partial fraction expansion method, :
X(jo).HG@) =
AL Ay AS
La XG HGo)=YGay= OE A
+0)" 2+ 30)" as jo? 7
‘On solving above equation, we get
1 1
Ays ge A= 0. Ag=
Fig. 4.12 Graphical Method
‘The convolution is given by 2
1 1
Yin) = A B-$-BeT STA ‘Ts, ee
v ws 69) = 44450)" 204 jo?
Verify - ‘On taking inverse Fourier transform, we get
Aan be
= + ate Luo
yO=\4e + ul)
Hence, x(t)*h(t) = ea tie Ans.
rob6, Compute (i) near and (i) crear periodic convolution of
"following sequences —
x(n) = {1,1,2,2} and x(n) ={1,2,3,4) (R.G.PV., June 2009)
given as follows, the linear
Aw) Sol () Using matrix representation
“=volution of the two sequences can be
a) = (3-8-7 T4) Hence, x0) 4.14. ‘ x2
[papinnijealimipealoaaany —
»= Tuto n) xz (m—),(Mod N), 1m = 0.
a0
For m=
x0) = Za »
n,(Mod 4) meme ) folded. The folded seyuen
wc ag see te
0) = 1S
For i, 5
a) = Daim. Mod:
=
From the product sequence, the sum is xy(1) = 17
For m= 2,
4) =
Hoided Sequence
Product Sequence a,()=2 -
‘Analysis of Discrete Time Signals 171
Mod4) is the sequence x3(n) folded x3 (0) is obtained by
(ayes 2 juct sequence, point by point and takit
aye 0 pos whe pre @=ni4 " camels
3
x3(D = ¥1(0)x3(1~n,Mod4)
. 0
ay3)=4 a ne product sequence, the sum is
J Rowateé from x3(1) = 16
3
x3(2)= %(n).x2(2-n, Mody
porp=> cal
the product sequence the sum i
From x32) = 14
10)e4
3
23, 3) = Y)x1(0).x2(3- n, Mod)
or P 0
luct sequence the sum is
From the ono) = 16
the circular convolution of the two sequences xy(n) and x3) is
aye
aye
tits a Time Hence OES (p) = (14s 16, 14,16) ra
Fig. 4.13 Circular Convolution t
Prob.7. Compute the circular convolution of the Oo segueac, suet ae
xy ln} = (1, 1,22, x2 1a) = U1, 2,3, g)
(AGP
Sol. Refer the sol. of Prob 6. ow HO? ayes ae
Prob.s. Perform the linear and circular convolution of a.) 2)"
Sequences x(n) and x(n) whem xj © (Jo2.5 4) and (3.1.42)
(RGPY, jw ay0re! 04
Sol Students solve yourself, same as Prob 6 aa, 4
Prob. For the two following sequences ~ zy
PLAY, x=(P2IO
FFind out the sequence x, (p) which is equal to circular con’ aes ayet ‘ 2
above two sequences. (RGPY, Jus
Sol. Circular convolution of the two sequences is shown in fg +
Net
X30) Yxy(a)xg(p-m,ModN) P= 0. —S aye? 1
“0 Folded Sequnce of x(n) Product Sequence
Fap=0 =} 140) = 2444642
*Zn@nCaMods) 140)=14
ete : Aratpa 1 ecrete Tre Sepals 173
a : Nj= Sand M, = 4
Wee N@NOM, Toten yey
0g 22008 Wo the sberre
nperaing 22008 sermences ie (NJ and (My 1) revon
i ebsed othe 208 L319) 8 Mn epecrey. Th
; i win) = (1.2.9.2, 1,0, 0,9
bin) = 1-2.-3,1,2.0,6,05 }
> } 9.9)
convolution of these padde
’ The convolution °d sexmences s cheamed anny matt
Product Sequence pethiod %
AAD ETB SS $40), 1o00F 2025p -,,
2 7 O12 3
0 0 J 1
000
2 2
; ‘ fl 100 f
0 2 0
° 0 021 F
Product Sequenc v 202
. ;
Folded Sequence Rotated by x2) =24+24+4+6214 Ie t2 a
Two Units in Line 3 1,0.0.0}6 (-
SAMPLING OF CT SIGNALS AND ALIASING
0.10. Discuss the need of sampling. (RGPV, Dec. 2014)
Ans, Basically, sampling refers to uniform sampling which convers an
#1 analog signal into a sequence of samples that are uniformly spaced a ume
0
Product S¢
Folded Sequence Rotated by aye ete p16. | Sunplng process is necessary to select he sampling rae utahy 0 that he
Three Unit in Line 2313) = sequence of samples can uniquely define the original analog signal
Prob.10. Convolve following sequences using matrix method -
x(n)
hin)
Sol. Given,
x(n)
hin)
Fig, 4.14 Circular Convolution Q.11. State and prove sampling theorem. (RGPY., Dec. 2014)
Or
State sampling theorem. (RG.V, June 2016)
‘Ans, The sampling theorem provides the basis for transmitting analog signals
y using digital techniques. The sampling theorem may be stated in two parts —
(A band-limited signal having no frequency components higher
than, Hz is fully described by its sample values at equal intervals less than ot
= 20 29
cl t
= (2 3 1 4
1 (R.G.RV, Dec. 20
‘ual to x sec apart, This statement is in frequency domain, pe
jing 10 ey
apatite sa pe knee of its sampien”
e009 ond. THIS gatement i time doy
1 eo yas uniform SAMPLE theory,
Beyond the frequency ,
transform. In the true may
noe its Fourier transform
orm diminishes after ayo
rected, and the Fourie ys
taken a bandslniegy
mis
ran durations:
atu of Fourier
r pand-timiteds SH
seul Signal ge ofthe rans
i it.can be nes!
froency 00h Tis pe of signals are
.
om stated adOVE CAN DE rove,
8 es considera band-limited sip)
nts beyond fn Hs fs Flo) is 2
‘onal is multiplied by a periodic in,”
‘When this signal is mu ima,
soe (Oa 2 aves sequence of pulses located at unin.
function S(t). ‘sirength of resulting impulses is equal to the valy.
ines fT 6000 ants, Function f) and its Foss ansfom i,
con aT 5 (a). Fig. 4.15 b) shows the periodic impulse func,
is shown in fg 415 0. TE 5 (a). Te product off and 8
a at fig. 15 (€. This product of 0 and Bt) shows the sampl
fi, located at uniform interval of T
Let us represent this sampled version o!
£00) = ft) BO)
specu (Fourier tansform) off) can be ahleved by using equ
convolution theorem. Accordingly
ves ;
s of sampling ther’
f f(t) as £,(0. Then —
{00 z {Fa @0,8.,(0)}
Wy _ 1
Substituting oe = jim the above equation, it is obtained that
1
Loe {Fo 28,,(@)} fi
Teer, the spectum of (0) canbe achieved by convving Fi
and, (©). By flipping the function §,(t) around the vertical ax
z
x0.
ny
x)
ly X(0-ko,)
: @--h
os x cm in fig. 4.17.
ing is :
1 1 cy consiingof aun of sified
Pe i junction of
Se
cigs shown in fig. 4.18
oe i
is 3 pea by
7 8 xo)
7 | | 1 seo
1
| |
7.
17
a
ko
Fig. 4,
tn any practical pulse amplitude modulation
system, itis impor
becomes small io maintain a constant time-average
Power inthe module
O16. Explain impulse train sampling, (RGPV, Dee
Or
x,(0)
Explain impulse tain sampling with the help of neat sketches,
(RGPY, Dec, 1
cmp les esmpg eter
wat samplng pride ansarees frequency of p(t), ©, = 2
“Sesame frequen. nie tine done
A = x0) pct
where p(t) Loa-n7
The 54(0)i an impulse train wi A
YT tha is — o
%0 = Leo 30-g7, meals _
180 Signals and Systems (M-Sem.. EX - Branch)
In fig. 4.18 (€). a= (% — Ow
; between the shifted f X(o0),
is mo overlap pernese is overlap ne 418.
1, < 204, eT itiples of the ing freq
at integer
eee ‘exactly from X,(0) by means of a I
cao oft frequency BTeater IN @yy_ and less thal?
rotiowed bY & mapping o -
Isc train to a sequence. Fi
ig steps are a3 follows
