0% found this document useful (0 votes)
83 views15 pages

DTSP Extc

1. Limit cycle oscillations arise in systems due to nonlinearities caused by quantization. Quantization of the product of two numbers in digital filter design can cause overflow errors, leading to limit cycles. 2. An additive noise model can be used to model quantization in linear systems. The quantization error can be modeled as an additive noise sequence added to the actual system output. 3. The difference between the actual and quantized values of a signal is called the quantization error or quantization noise. The maximum value of quantization error is +/- S/2, where S is the quantization step size. Reducing the step size reduces the quantization error.

Uploaded by

Purva K
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
83 views15 pages

DTSP Extc

1. Limit cycle oscillations arise in systems due to nonlinearities caused by quantization. Quantization of the product of two numbers in digital filter design can cause overflow errors, leading to limit cycles. 2. An additive noise model can be used to model quantization in linear systems. The quantization error can be modeled as an additive noise sequence added to the actual system output. 3. The difference between the actual and quantized values of a signal is called the quantization error or quantization noise. The maximum value of quantization error is +/- S/2, where S is the quantization step size. Reducing the step size reduces the quantization error.

Uploaded by

Purva K
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

FogeNo.

1
SEM-5 Dalc 25 21

Seat Na: 560016


Sub:Discrete Iime 8ignalProcessIng- P-PkaneRax

1MC9s
1option B 0 1tj 3

2option D Four

90pio.n 8 2 ,0,0,1
4option damplinq theimpulë re ponse of an
equivalent analog filter
S0ptión D ven and oddsamples

G loption B
GLLoption aif of the Quantizationstep

3aptio n D If HC)= CK thenH )


Cs-Pe (-e
3Option B _Q: 05 So Hz

9option h n) a nti symmetrt & Modd

1oOptian C multipliers.

Page No1 /5
SEM 5 PagaNo.2
Data 25 21

Seat NO: S6 0016


ub Dfscrete rine siqnal Processing9
PPkavleMa

i Condltionsor inear phase fTR flHers &e1 umple


far type of lin eor phase FIR flter.
Ans Linearphase tilttrinq impo THan t in alot of
üpplicatinos The fIR_tr h.ave linear phuse if i
Impulse teponie- is symmetric &qdtiymmetric lbis
basicaluq means that the impulse re9.pans&eSatfgti
the fou. condition b n ) s t h (M -1-n)
four ypeof FIR fiIe rs:_
Tyee 1 :5qmmetri 5.equen ce of odd length
Iupe : Symmetric cequehLe of even lehqth
Ty pe3 antssmmetne equenLe of_odd length
pe 4 ondiymme tri sequenee af e venlength

a [k 2 hM=E| k 12
Ty hk12hMEk kL,M
2

K L2,-
2

Type dfki 2 MtI-k ,2, 3


2
M
2

LnParsevaltheorem:
N-
then D:0
n) * (h) =1wher 1,tn) &

K CE)X*(k) X2(n) are N polnt:Sequence-

Stat ment_of Parse vals fh


Paqe na:215
SEM SS Page No. 3
Date 5 I 2U
Seat_No 56 0016
Sub Distxete ime signal pro.cessing P Pk a r l e k a

2 Ail bET DF

h) Yk)Par Seyal1
N
then 2o)y*h) E X(K)YYk)_th
D0 N

XCE X(K) x (K)epeen.h _eneryy tonttn-oF


Siq nal at qtven freq uency
Far paint equen ce N: 4
Let xCn-1,2,10 0
By def2
By def TCxn)] x CE)= WRZO
her w i Huiddu_facto matnix
1
X)= 2
-

Lc

+ 2 t|+0

1-2 t 0 2 'ztk))
1 - 2 t| -0
4,-2, o, 2j3-
Li +2-l-o N
L 2j
-

Enerau signal, nzo


2

=_6
Similar y X(E)=11 Kco)
N K0

+ x C |x (a)+ |xC3)3
F ro m

Pageno 3/15
SEM S
o2SI2

Seat No 560016
ub:Discvete time siqnal prote5ing P.Pkariekar

o+-11 +o o'42

1 6 43 6 4
from
n) N K0
x,K)
Partevals theore i pr Ved tor the sequence.

91 0): {11,1,1,1,10,02
2- XCK)
12co) A CO) GCo) XC6)

1 () AC XC1

BCo) G2)
1-1

6) Bt X(3

1 z0 CCo) HoA

1 X 5) 5)

1 Yc3) Do) wRHo)


sX C6)

DCO 7
N3)L- xC

PageNo: Z 5
SEM 5 Page No. S
Date 25 11 2/

Se at_ INO: 5600 1 6


Sub:Discrete time siqhal Processing
PP.Karle
a .

2B1 No
X Co) 3 f_C32 = 6

V2

XC2_ 1 t1 ( ) = 1 j

+0 0 2 t o 29 3
x C42 9 t (-1) 3 0
xC5)
2
V

o 70 0:2 93__
5milarly-
X CG)_ 1 +J

Paqt55
Sem 5 Page:G
Date:25/it/21
Seat No: 5.6.0.0116
Sub Discete Lime siqnal p ocessin9
P.P. K a r 1 e k a r

3 A}
i1 Limit cycle os.cillotions
Limitcy.cle isCloSedra jectoryjophase spa ce havinD
heproper ty
O As name SuagestSthere qre- two leel oscil1ations fbat
exst even dh en the input po vided tu the sus tem is
Zero

These oScllations are basicalq_disturbances thaf may-


exist in qn_otheru iie stable£ilte
These Ascillations are Yecursi/e in_nat ure as the q are
back to tht system
fecl

2exoilpimit cyeleasci. atise due to non lineo ri tie_


that are eued byquantiz2ation.
S Multiplicat" of o b bit no s qive nse6 2b bit
tesult The
tult then quantined
Thi 9uant2at_
givs eto 1eA ilp limit Cycle otcila1
Firstordzrl1R_diff eq _
y X ht a u
|Modifiedeq"yCn) = zn)t quahtlay(nd)) I
i Product quantizationtror
0 Wbenever the muliplîcatton operat js don e tn_ deaigninq
af digital fi|ter then the xdundtnq_f the teul of
the mulhplication fr_dene Bocaune oF this 4he errors are
4ener t hese eror are CaTle d product ro und aff etra
Nouquantiza1 0f he praduet of 2 no.s cawe he overflow
%the espbnse become DDnlin ear
In _ard er to oDtain mor qeneral reg ult the quantizai
nuliipkco s mode le d_a on additir e noi e
Sequen te n ) Consider an additve _made for guantiz o7_
Rrar n a hqie pale y SCm:

PageNo: 6 1 5
SEM S Date 25/1||2
Page No:

Seat No: S60 016_


Sub D iscrete lime
siqnal pro cessing
PPKareka

--3A]
-
Vn)

Additive DDise
melho d for
inale p0le tyskm

W can write dIffere nce e aSS


Vn) avn ) t X(D).t eCn)
h eekectad oFounding thepioduch avth-) i
m.odelledas an nose 9equen.c elo) it is_added fo-
acku a / pavCh-1)

nputQuan tination Érro '


The differebLe bet in1t Dtenol yeluel of quan tized
signa itp i called asquantiz at"_ eru or quanti2o_
hoise 0
2 Maz value of quanii2af error ir t S/2 where
Step ize to redute quantiz.at
at_erYo we ha Le_
reduce te p si2ebuincreauinq no 0f uaDt2gt"

eve _ The mean squ are of qugntiza eor =s*(_


(3Relat bet -n 0f quanliza" Jevels ( L 2
no of hitr pe word (N n_the tuns mdtd signal
canbe Poub d_ dello u
Be cau Se p eb quantized evel is fobe_converted into N
itdiqita oYdass Uminq a bin4r y cd dd. o/p sîgnal
Num beY OLantizat leve no. uf om bin ql s of_
N
Lit wor d
Thw N 4 bik_2r ord then no of
uantizat" Leve s 216
Pagero7/15

Page No 7/5
Pagwo:?
Date : 5 / 1 1 / 2 1
-SEMS

SeatNo:s6 0016
Sub:Discrete_Jime siqnal Protessinq Pf.karleKev

3 c -3J W

Ha ( for 0 w T2
for T/2 ` w ST
Deje rin ing fiHer coe £fiien n ) for M=1 uwinq
HeqUeN Samplinq echnique
So Heve theSample of Civen Are qu enq resDonseS faken_
uoiformls at 2 Ttk/Mfor 0Su
the value F k e 0, ForTSwS S2 k 2,3,4,5
and
far 31 w 2T E=6 Th us cGmple d
tre quentu Teaponee is gle
0,l
3,4,5-
e
The Her_caefficient b_(n) are givenbyinverse
diicret fourierIranr form
32T kn/M
baco) H CE) e
M k0
J2TTK D77
[ES e
T=0.

J2KD/ -JGTET
te

jKT (n-YT e

Tth-377 e J127T(n-3)77
1+ e
2Ttt-3)/T
Since
Since H_k)
J2r Cn-30/7
=_H*M-k)e have e
J2I{n=3)77 Jatth=3YTT
h (n) - / z 1 t e te
h 0:429t0:285 1 (oso.8 9 8(n-3)L
PagNo8_15
PageNo. 9
SEM5 Date 25 I2

Seat Na S6 0o16
Sub DTSP
PPkorlelca

4B
Suh-band Coding
DigitalSiqna s obtained bs uSing #/D converter which
maybeneeded to be stared for future useIn orderto
in.ctease the efficien.cy of the sysBem, the digitalsig.nalis
ta be startd 0Ver linit.ed bandcaidth. Ihis_ean be achieved
using
bysignal com preasion fechnique Siq nal compressiop is
qchieve d by using subhan.d coding
a Subband_Coadinq ene o din g. folla winq art th steps in valved
he ilp sianalatn) s_given to Set af filt ers wbich
deta mpase the input_Slqn.al int o set af narramhandiighal
oCuyingontig Ueu fre.q handz.
2heenaa band signals_are daun tunpled praduing
Subband_siqnals _
3. A1l narmu hand siqnals are then compresed by encader bis
redute data rate & Stargtreq.uirement af signal
4Thee camprested siqnals are muliplexed& transmitted
over the channel Thusignal Ln) ha gu.ced data rate
a sConmpare dto Cr)
E) Subband cadin decader: Steps
1 Re ceiuedsiqnal q ln) i fint demuliiplexed into set af
Subban.d stqnal2]Then_the _ubband siqn.al ts decam-
presiedby- decader 3 Decoded signal is upsampled by

ing fam plet t)Theunsampled signal{f_given ta


(2
Synthei's ilter bank G.(2) t o L
cambin ed
s] The o/p ofsynthesi filter bankart
o pro.duce o/p signal y[n)
Page No: 9 Z5
StM 5
Dul

Seat NO: S6 0 016


Sub: Discrete time sighaProcessing
PP.kor (ekqt

ii) JIR iHers fIR_S'lters


O Number of multiplic.aptions D Number of maltiplication S ot

hecessary are least mosst

Stability depend (2 Stability is Gugrapte d


Syste m_deiqn
Thee i no lin.ear phaue There is uanted inearphae
Hareware ilter contro Harduam filter contro cample
Complerity irMo derate ity imple
SSuppordts adaptive filkrin (5 6upports a deptive fjlte rin q
Requireded harduare memary Required hardware memory
LS les hfgh
Moxt compiCated (quantizat) G laost_samplicatedlquantizst2

Page:Lo5
D em ultiplex er

T7

Multpleer
PagoNo 12
SEMS
Dato 25 I 2

6eat NO: SG0016


PPkar(e lcar
Sub: DISP

H(2) Z

So Cascade Realizatid n: The ttansfer function


Canbe_ re arran ge d
H(2) ( +

1
1-1z4L z
The above s form 0F HC2) 1 6 HLz) Hi02)
H(2)
There are 2 wayof implementing th sytem dependinq-
on Comhinat oF num emtar m difh deno._
Casca de tealizaHon structure
Cn) yCo)

tt

Page No :12 15
Page No : 3

SEMS Date 25/11/21

Seat No: S6 0016


Sub DTSP
PP kartekar

Q4C) >Paallel Realizato n


H(2)
1-
Dividin 9 tbe numerator by tbe deno minator weabtain.
2'+?
-3 .

z
-2

tz t -6+ 4 7x|

- i t 5/8 z
22

H C2) z
H(a) +Haz)
where

H -an d Ha Cz)=itz

PageNo:15
N
E
Page :15
SEM 5 Date 2s/1/2

3.eqt NO' 560.01 6


Sub Dist rete Time Sianal processng PP.karle kar

Apo G5 rad sec


-No, S
Tfanu) 2 tan(o)=2:75 rad/sec
Ts

Now N Z Laq (A/£) lo94-89/0 7s)


o9 (As/0p) log(2-?S /o.65)
:299
2:3
We found ih efA tanext laryest integ
N-2
0 655 L0-25)V2
c 075 dd/seec
This i Bilin ear tans forma tton meth a d.
tekno,-
0:2 T, W20-CT A=4:89 &o-75
Saurn pling ime Is =1sec
p Ts0
dsTs 0-6 T

. N 2 1lo (aE)_la (A/Ap)


N 2_70 6
N Z2
N 22

No w 0 2t

C-75)22
c 022 rad /sec|

Page No: 15 /5

You might also like