Beruflich Dokumente
Kultur Dokumente
SEM/YEAR: V/III
PART A
UNIT I
SIGNALS AND SYSTEMS
1. Define Signal and Signal Processing.
A signal is defined as any physical quantity that varies with time, space, or any
other independent variable.
Signal processing is any operation that changes the characteristics of a signal.
These characteristics include the amplitude, shape, phase and frequency content of a
signal.
2. What are the classifications of signals? (Nov/Dec 2010)
There are five methods of classifying signals based on different features:
(a) Based on independent variable.
(i)
(ii)
(ii)
(iii)
(c) Depending upon the certainty by which the signal can be uniquely described as
(i)
Deterministic signal.
(ii)
Random signal.
Periodic signal.
(ii)
Even signal.
(ii)
Odd signal
x ( n)
A signal x(n) is called an energy signal if and only if the energy obeys the relation
0 E . For an energy signal P=0.
Power signal:
The average power of a discrete time signal x(n) is defined as
N
1
P lim
x ( n)
N 2 N 1
n N
A signal x(n) is called power signal if and only if the average power P satisfies the
condition 0 P .
(iv)Sequence representation
13. What are the different types of operations performed on discrete time
signals?
The different types of operations performed on discrete time signals are
(1)Delay of a signal
(2) Advance of a signal
(3)Folding or Reflection of a signal
(4) Time scaling
(5) Amplitude scaling
(6)Addition of signals
(7)Multiplication of signals.
14. Represent the following duration sequence x(n)={1, 3, -1, -4} as a sum of
We can write
x( n) 1 x(k ) ( n k )
2
1 ( n 1) 3 ( n) 1 ( n 1) 4 ( n 2)
y ( n) My
and
20. Define unit sample response (impulse response) of a system and what is its
significance?
The unit sample response is defined as the output signal designated as h(n) ,
obtained from a discrete time system when the input signal is a unit sample
sequence ( unit impulse).
The output y(n) of an LTI system for an input signal x(n) can be obtained by
convolving the impulse response h(n) and the input signal x(n).
y ( n) x (n) h(n)
x ( k ) h( n k )
h(k )
x ( k ) h( n k )
(OR )
y (n)
h(k ) x( n k )
2.
Folding: Fold the h(k) about the origin and obtain h(-k).
3.
positive or
left if n is negative to obtain h(n-k).
4.
w(k)=x(k)h(n-k)
5.
the
value of output y(n).
6.
one
sample and do step4.
25. What do you meant by sampling process?
Sampling is the conversion of a continuous time signal (or analog signal) into a
discrete time signal obtained by taking samples of the continuous time signal ( or
analog signal ) at discrete time instants.
26. State sampling theorem. (April/May 2011, Nov/Dec 2012)
A sampling theorem states that the band limited continuous time signal with
highest frequency (band width) fm hertz , can be uniquely recovered from its samples
provided that the sampling rate fs is greater than or equal to 2fm samples per second.
27. Define Nyquist rate. (Nov/Dec 2008)
The Nyquist rate is defined as the frequency 2fm , which , under sampling
theorem, must be exceeded by the sampling frequency.
28. What is meant by Quantization process?
The process of converting a discrete time continuous valued signal into
discrete time discrete valued signal is called Quantization.
Pav
Pq
Votagereso lution
VFS
2n 1
where
VFS - full scale voltage and
n number of bits.
41. Define percentage resolution.
Percentage resolution is defined as
% resolution
1
100
2n
42. What are the advantages and disadvantages of counter ramp type ADCs?
Advantages of counter ramp type ADCs are
1. The principle is simple and straightforward.
2. It is very easy to construct this ADC.
3. This basic principle is employed in many advanced ADCs.
Disadvantages counter ramp type ADCs are
1. Only increasing voltages can be measured.
2. The system is very slow.
3. This may be mainly used to read DC voltages.
4. Practically the system comes to rest only when Vd > Va . This set an
error in readings.
43. What are the advantages and disadvantages of SAADC?
Advantages of SAADC are
1. It is more accurate than the stair case (counter ramp) ADC.
2. It maintains a high resolution.
3. It is much faster.
4. Its conversion time is much less.
Disadvantages of SAADC are
1. It requires a complex register called the successive approximation
register
2. It is costly, as it contains more components.
44. Write down the formula for conversion time of SAADC?
Conversiontime
n
f
where
n number of bits and
f clock frequency.
10
x ( n) z
11
z a1 x1 (n) a 2 x2 (n) a1 X 1 ( z ) a 2 X 2 ( z )
m 1
z
x
(
n
m
)
z
X
(
z
)
x(i ) z m i
2. Shifting: (a)
i 0
(b) z x (n m) z m X ( z )
3. Multiplication: z n m x(n) z
dz
X ( z)
4. Scaling in z- domain: z a n x( n) X a 1 z
5. Time reversal : z x( n) X ( z 1 )
6. Conjugation: z x (n) X ( z )
7. Convolution: z
H ( z ) R( z )
m0
X ( z)
8. Initial value: z x(0) zLt
(1 z 1 ) X ( z )
9. Final value: z x( ) zLt
1
52. State Parsevals relation in z-transform. (April/May 2011)
Parsevals relation in z-transform state that
12
x ( n) x
1
2j
( n)
1 1
v dv
v
(v ) X 2
53. What is the relationship between z-transform and DTFT? (Nov/Dec 2010)
The z-transform of x(n) is given by X ( z )
x ( n) z
.(1)
Where z re j
Substituting z value in eqn (1) we get,
x ( n) r
X (re j )
e jn ..(2)
X ( e j )
x ( n )e
j n
.(3)
Eqn(2) and Eqn(3) are identical , when r=1. In the z-plane this corresponds to the
locus of points on the unit circle
z 1.
For X (e j ) to exist , the ROC of X(z) must include the unit circle.
54. What are the different methods of evaluating inverse z transform?
1. Long division method.
2. Partial fraction method.
3. Residue method.
4. Convolution method.
55. Find the convolution of the following using z- transform.
Solution:
13
Z x(n) h(n) X ( z ) H ( z )
X ( z )(1 z 1 z 2 )
H ( z ) (1 2 z 1 z 2 )
X ( z ) H ( z ) (1 3 z 1 4 z 2 3 z 3 z 4 )
x(n) h(n) 1,3, 4,3,1
56. Define system function.
Let x(n) and y(n) is the input and output sequences of an LTI system with
impulse response h(n). Then the system function of the LTI system is defined as
the ratio of Y(z) and X(z), i.e.,
H ( z)
Y ( z)
X ( z)
where,
Y(z) is the z transform of the output signal y(n)
X(z) is the z transform of the input signal x(n)
UNIT II
14
FREQUENCY TRANSFORMATIONS
1. Define Fourier transform of a discrete time signal. (April/May 2008)
The Fourier transform of a discrete time signal x(n) is defined as
F x (n) X ( )
x ( n)e
jn
1
2
X ( )e
jn
15
H ( )
is symmetric and H ( ) is
antisymmetric.
iii) If h(n) is complex then the real part of H ( ) is antisymmteric over the
interval 0 2 .
iv)The frequency response is a continuous function of .
8. Write short notes on the frequency response of first order system.
A first order system is characterized by the difference equation
y ( n) x( n) ay (n 1)
16
X p (k ) x p (n)e j 2kn / N
n 0
11. What are the two basic differences between the Fourier transform of a
discrete time signal with the Fourier transform of a continuous time
signal?
1. For a continuous signal , the frequency range extends from to .
On the other hand , the frequency range of a discrete time signal extends from
to ( or 0 to 2 ) .
x ( n ) e j n
n 2
e j 2 e j 1 e j e j 2
1 2 cos 2 cos 2
17
j n
X ( k ) x ( n)e j 2kn / N
k= 0, 1, 2..N-1.
n 0
x ( n)
N 1
X ( k )e
j 2k / N
n= 0, 1 , 2 ..N-1 .
K 0
then
x(n N ) x (n) for all n
X (k N ) X (k ) for all k .
(b)Linearity
If
X1(k)=DFT[x1(n)] and
X2(k)=DFT[x2(n)]
then
DFT[a1x1(n)+a2x2(n)]=a1X1(k)+a2x2(k)
( c ) Time reversal of a sequence
If DFT {x(n)}=X(k),
then
DFT{x((-n))N}=DFT{x(N-n)}=X((-k))N=X(N-k)
(d)Circular time shifting of a sequence
If DFT {x(n)}=X(k),
then
DFT{x((n-l))N}=X(k) e j 2kl / N
18
13. IF N-point sequence x(n) has N- point DFT X(k) then what is the DFT of the
following?
(i ) x (n)
(ii ) x ( N n)
Solution:
(i ) DFT {x ( n)} X ( N k )
(ii ) DFT {x ( N n)} X ( k )
(iii ) DFT {x ((n l )) N } X ( k )e j 2kl / N
(iv ) DFT {x ( n)e j 2 ln/ N } X ((k l )) N
forN 16
Solution:
X (k )
N 1
x ( n)e
j 2kn / N
K=0, 1 , 2..N-1
n0
n 0 4
15
e j 2kn / 16
1 j k / 8
e
n 0 4
15
16
1
j 2k
e
4
1
1 e jk / 8
4
DFT[x(n)]=X(k),
then
DFT[x((n-m))N]= e j 2kn / N X (k )
19
N 1
X ( k ) x ( n)e j 2kn / N
k=0,1,2,..N-1.
n 0
x(n)e j 2 kn /4
k 0,1, 2 , 3.
n 0
X (0) x(n) {1 1 0 0} 2
n 0
3
X (1) x( n)e j n /2 {1 j 0 0 } 1 j
n 0
3
X (2) x(n)e j n {1 1 0 0 } 0
n 0
3
X (3) x( n)e j 3 n /2 {1 j 0 0 } 1 j
n 0
X (k ) {2,1 j , 0,1 j }
DFT[x(n)]=X(k),
Then
DFT[x(n) e j 2 ln/ N ] X ((k l )) N
23. What is zero padding? What are its uses? (May/June 2009)
20
(ii)
DFS x
( n) X
DFS x1 p (n) X 1 p (k )
2p
2p
and
(k )
(1)
If X 3 p (k ) X 1 p (k ) X 2 p (k )
21
Given two sequences x1 ( n) and x 2 (n) , the circular convolution of these two
sequences x3 (n) x1 ( n) Nx 2 (n) can be obtained by using the following steps.
1. Graph N samples of x1 ( n) , as equally spaced points around an outer circle
in counter clockwise direction.
2. Start at the same point as x1 (n) graph N samples of x 2 (n) as equally
spaced points around an inner circle in clock wise direction.
3. Multiply corresponding samples on the two circles and sum the products to
produce output.
4. Rotate the inner circle one sample at a time in clock wise direction and go to
step3 to obtain the next value of output.
5. Repeat step4 until the inner circle first sample lines up with the first sample of
the exterior circle once again.
27. Distinguish between linear and circular convolution of two sequences.
Sl.NO
Linear Convolution
1.
If x(n) is a sequence of L number of
Circular convolution
If x(n) is a sequence of L number of
2.
contain N=L+M-1
Linear convolution can be used to find
3.
response of a filter.
h( 2)
h(0)
h(2)
1
2
h(1)
2
1
2
2
h(1) x (0)
y (0)
h( 2) x (1) y (1)
y ( 2)
h(0) x (2)
1
3
2 2
1
1
2
1
2
y ( n) 3, 2, 1
31. How will you obtain linear convolution from circular convolution.
Consider two finite duration sequences x9n) and h(n0 of duration L samples and
N samples respectively. The linear convolution of these two sequences produces an
output sequence of duration L+M-1 samples , whereas , the circular convolution of
x(n) and h(n) give N samples where N=Max(L,M) . In order to obtain the number of
samples in circular convolution equal to L+M-1, both x(n) and h(n) must be appended
with appropriate number of zero valued samples. In other words, by increasing the
length of the sequences x(n) and h(n) to L+M-1 points and then circularly convolving
the resulting sequences we obtain the same result as that of linear convolution.
32. What is meant by sectioned convolution?
23
If the data sequence x(n) is of long duration , it is very difficult to obtain the
output sequence y(n) due to limited memory of a digital computer. Therefore, the data
sequence is divided up into smaller section. These sections are processed separately
one at a time and combined later to get the output.
33. What are the different methods used for the sectioned convolution?
The two methods used for the sectioned convolution are (i)the overlap-add
method (ii)Over lap-save method.
34. Differentiate (i) overlap-add method (ii) overlap save method.
Sl.No
Overlap add method
1.
In this method the size of the input
2.
3.
4.
block is L.
Each data block is L points and we
DFT.
In this no corruption due to aliasing as
convolution is employed.
To form the output sequence the first
circular convolution.
To form the output sequence, the last M-
succeeding block.
2.
DFT
Obtained by performing sampling
DTFT
Sampling is performed only in time
domain.
frequency domains.
Discrete frequency spectrum
Continuous function of
24
N
log 2 N
2
respectively.
38. How many multiplications and additions are required to compute N-point
DFT directly? (April/May 2008)
The number of multiplications and additions required to compute N-point DFT u
are N ( N 1) and
N2
respectively.
25
N
64
log 2 N
log 2 64 192
2
2
4096
21.33
192
Linear filtering
(ii)
Correlation
(iii)
Spectrum analysis
UNIT III
IIR FILTER DESIGN
1. What are the different types of structures for realization of IIR systems?
The different types of structures for realization of IIR system are
27
28
29
H (z)
b z
k 0
1 ak
k 1
12. Give the magnitude of Butterworth filter. What is the effect of varying order
of N on magnitude and phase response?
The magnitude function of the Butterworth filter is given by
1
H ( j )
1
2
2N
N 1, 2 , 3..............
Where N is the order of the filter and c is the cut off frequency. The magnitude
response of the Butterworth filter closely approximates the ideal response as the order
N increases. The phase response becomes more non-linear as N increases.
13. Give any two properties of Butterworth lowpass filters.
1. The magnitude response of the Butterworth filter decreases monotonically as
the frequency increases from 0 to .
2. The magnitude response of the Butterworth filter closely approximates the
ideal response as the order N increases.
3. The Butterworth filters are all pole designs.
4. The poles of the Butterworth filter lies on a circle.
5. At the cut off frequency c , the magnitude of normalized Butterworth filter
is
1
2
30
1
c
1 2 C N2
where
C N
31
20. How the order of the filter affects the frequency response of Chebyshev filter.
From the magnitude response of Type -1 Chebyshev filter it can be observed that
the magnitude response approaches the ideal response as the order of the filter is
increased.
21. How will you determine the order N of Chebyshev filter.
The order N of the Chebyshev filter is given by
cosh 1
N
cosh 1
Where
10
0.1 p
100.1 1
s
1
1 2
at the cutoff
frequency c .
23. Compare the Butterworth and Chebyshev Type -1 filter. (May/June 2009)
Sl.No
Butterworth filter
1.
All pole design
2.
The poles lie on a circle in s-plane
3.
The magnitude response is
maximally flat at the origin and
Chebyshev filter
All pole design
The poles lie on a ellipse in s-plane
The magnitude response is equiripple in
pass band and monotonically decreasing in
32
4.
of .
The normalized magnitude
5.
1
at the
2
value of
1
1 2
c .
function.
function.
24. What are the different types of filters based on the frequency response?
The filters can be classified based on frequency response. They are (i) low
pass filter (ii) high pass filter (iii) Band pass filter (iv) Band reject filter.
25. Distinguish between FIR and IIR filter. (Nov/Dec 2008)
Sl.No
1.
FIR filter
These filters can be easily
IIR filter
These filters do not have linear phase.
phase.
FIR filters can be realized
3.
4.
kind of filters.
The round off noise in IIR filters are more.
minimax design.
27. What do you understand by linear phase response?
33
For a linear phase filter ( ) . The linear phase filter did not alter the shape of
the original signal. If the phase response of the filter is non linear the output signal
may be distorted one. In many cases a linear phase characteristic is required
throughout the passband of the filter to preserve the shape of a given signal with in
the pass band. IIR filter cannot produce a linear phase. The FIR filter can give linear
phase, when the impulse response of the filter is symmetric about its mid point.
28. For what kind of application , the antisymmetrical impulse response can be
used?
The antisymmetrical impulse response can be used to design Hilbert transformers
and differentiators.
29. For what kind of application, the symmetrical impulse response can be used?
The impulse response, which is symmetric having odd number of samples can be
used to design all types of filters , i.e., lowpass, highpass , bandpass and bandreject.
The symmetric impulse response having even number of samples can be used to
design lowpass and bandpass filter.
30. How can you design digital filters from the analog filters?
1. Map the desired digital filter specifications into those for an equivalent
analog filter.
2. Derive the analog transfer function for the analog prototype.
3. Transform the transfer function of the analog prototype into an equivalent
digital filter transfer function.
31. Mention any two procedures for digitizing the transfer function of an analog
filter.
The two important procedures for digitizing the transfer function of an analog
filter are
1. Impulse invariance method.
2. Bilinear transformation method.
34
32. What are the requirements for a digital filter to be stable and causal?
i.
The digital transfer function H(z) should be a rational function of z and the
co-efficient of z should be real.
ii.
iii.
33. What are the requirements for a analog filter to be stable and causal?
i.
ii.
iii.
iv.
34. What are the advantages and disadvantages of digital filters?
Advantages:
1. High thermal stability due to absence of resistors , inductors and capacitors.
2. The performance characteristics like accuracy, dynamic range, stability and
tolerance can be enhanced by increasing the length of the registers.
3. The digital filters are programmable.
4. Multiplexing and adaptive filtering are possible.
Disadvantages:
1. The bandwidth of the discrete signal is limited by the sampling frequency.
2. The performance of the digital filter depends on the hardware used to
implement the filter.
35. What is impulse invariant transformation?
The transformation of analog filter to digital filter without modifying the impulse
response of the filter is called impulse invariant transformation (i.e., in this
transformation the impulse response of the digital filter will be sampled version of the
impulse response of the analog filter.)
35
38. What is the relation between digital and analog frequency in impulse
invariant transformation.
The relation between analog and digital frequency in impulse invariant
transformation is given by
Digital frequency, T
Where,
36
right half of s-plane is mapped into exterior of unit circle in z-plane . The Bilinear
mapping is a one to-one mapping and it is accomplished when
s
2 1 z 1
T 1 z1
40. What is the relation between digital and analog frequency in Bilinear
transformation?
In Bilinear transformation, the digital frequency and analog frequency are related
by the equation,
1
Digital frequency, 2 tan
Analog frequency
T
2
or
tan
T
2
Where,
- Analog frequency
T - Sampling time period
37
In IIR filter design using bilinear transformation the specified digital frequencies
are converted to analog equivalent frequencies, which are called prewarp frequencies.
Using the prewarp frequencies, the analog filter transfer function is designed and then
it is transformed to digital filter transfer function.
44. Compare the impulse invariant and bilinear transformations.
Sl.No Impulse Invariant transformation
1.
It is many to one mapping
Bilinear transformation
It is one to one mapping.
2.
3.
frequency is nonlinear.
There is no problem of aliasing and so
limited.
The magnitude and phase response of
limited.
Due to the effect of warping, the phase
sampling frequency.
4.
UNIT IV
FIR FILTER DESGIN
1. What is the condition for the impulse response of FIR filter to satisfy for
constant group and phase delay and for only constant group delay?
For linear phase FIR filter to have both constant group delay and constant phase
delay.
( )
N 1
2
h ( n) h( N 1 n)
N 1
2
ii)
iii)
iv)
4. What is the necessary and sufficient condition for the linear phase
characteristic of a FIR filter? (May/June 2006)
The necessary and sufficient condition for the linear phase characteristic of a FIR
filter is that the phase function should be a linear function of
, which in turn
The following conditions have to be satisfied to achieve constant group and phase
delay.
Phase delay,
N 1
( i.e., phase delay is constant )
2
Group delay,
6. What are the possible types of impulse response for linear phase FIR filters.
There are four types of impulse response for linear phase FIR filters
1. Symmetric impulse response when N is odd.
2. Symmetric impulse response when N is even.
3. Antisymmetric impulse response when N is odd.
4. Antisymmetric impulse response when N is even.
7. List the well known design techniques for linear phase FIR filter. (May/June
2006)
There are three well known methods of design techniques for linear phase
FIR filters. They are,
i)
ii)
iii)
8. Write the two concepts that lead to the Fourier series or window method of
designing FIR filters.
The following two concepts lead to the design of FIR filters by Fourier series
method.
i.
ii.
9. Write the procedure for designing FIR filter by Fourier series method.
(April/May 2008)
i)
ii)
1
2
( )e jn d
40
iii)
sequence h(n) .
Take Z transform of h(n) to get a noncausal filter transfer function
iv)
Multiply H (z ) by
v)
N 1
N 1
h(0)
N 1
2
h( n) z
n 1
response is truncated at n=
will lead to oscillation both in passband and in stopband. This phenomenon is known
as Gibbs phenomenon.
12. Write the procedure for designing FIR filter using windows.
i)
ii)
( )e jn d
41
iii)
iv)
N 1
h(0)
N 1
2
h( n) z
n 1
13. What are the desirable characteristics of the window? (May/June 2009)
The desirable characteristics of the window are
1. The central lobe of the frequency response of the window should contain most
of the energy and should be narrow.
2. The highest side lobe level of the frequency response should be small.
3. The side lobe of the frequency response should decrease in energy rapidly as
tends to.
14. What is window and why it is necessary? (April/May 2008)
The way of finding an FIR filter that approximates H (e j ) would be to truncate
N 1
. The abrupt truncation of the series will
2
42
iii).The attenuation in the stop band is fixed for a given window , except in case of
Kaiser window where it is variable.
16. Write the procedure for FIR filter design by frequency sampling method.
1. Choose the desired frequency response H d ( ) .
2. Take N samples of H d ( ) to generate the sequence
3. Take inverse DFT of
~
H ( k ).
~
H ( k ).
WR ( )
N
2
sin
2
sin
4
N
v)
vi)
vii)
2n
N 1
for
=0
( N 1)
( N 1)
n
2
2
Otherwise.
2n
N 1
=0
for
( N 1)
( N 1)
n
2
2
Otherwise.
2n
4n
0.08 cos
N 1
N 1
for
=0
( N 1)
( N 1)
n
2
2
Otherwise.
2n
N 1
for
=0
( N 1)
( N 1)
n
2
2
Otherwise.
44
2n
N 1
I 0 ( )
wk ( n) I 0
for n
=0
Where
( N 1)
2
Otherwise.
is an independent parameter.
I 0 (x) is the zeroth order Bessel function of the first kind
1 x
k 1 k! 2
I 0 ( x) 1
i)
ii)
iii)
27. Why the triangular window is not good a good choice for designing FIR
filters?
In FIR filters designed using triangular window the transition from passband to
stopband is not sharp and the attenuation in stopband is less when compared to filters
designed with rectangular window. For the above two reasons the triangular window
is not a good choice.
28. List the features of hanning window spectrum.
i) The mainlobe width is equal to
8
.
N
45
8
.
N
Hanning window
The width of mainlobe in window
4
N
spectrum is
8
N
2.
3.
increasing
In FIR filter designed using
attenuation is 44dB.
4.
Hamming window
The width of mainlobe in window
4
N
spectrum is
8
N
2.
3.
increasing
In FIR filter designed using
attenuation is 51dB.
4.
46
Sl.No
Hanning window
1.
The width of mainlobe in window
spectrum is
Hamming window
The width of mainlobe in window
8
N
spectrum is
8
N
2.
3.
increasing
4.
high frequencies.
In FIR filter designed using Hamming
attenuation is 51dB.
Blackman window
The width of mainlobe in window
8
N
spectrum is
12
N
2.
3.
increasing
In FIR filter designed using
increasing
In FIR filter designed using Blackman
attenuation is 78dB.
The higher value of sidelobe attenuation is
mainlobe width.
4.
5.
high frequencies.
47
12
.
N
3.
4.
Kaiser window
The width of mainlobe in window
spectrum depends on the values of and
8
N
N.
The maximum sidelobe magnitude with
increasing
In FIR filter designed using
48
UNIT V
APPLICATIONS
1. What are the classification digital signal processors?
The digital signal processors are classified as
i.
ii.
49
50
51
Clock generator
(ii)
Hardware timer
(iii)
(iv)
(v)
(vi)
(vii)
(viii)
(ix)
(x)
BIO
52
hardware
Increased dynamic range
Round off errors can occur with both
additions
5.Overflow occur in addition
computers
26. What are the three quantization errors due to finite word length registers in
digital filters?
1.Input quantization error
2.Coefficient quantization error
3.Product quantization error
27. How the multiplication and addition are carried out in floating point
arithmetic?
In floating point arithmetic, multiplications are carried out as follows
Let f1=M1 x 2c1 and f2 = M2 x 2c2. Then f3=f1x f2 = (M1xM2)2(c1+c2)
That is, mantissas are multiplied using fixed point arithmetic and the exponents
are added.
The sum of two floating point numbers is carried out by shifting the bits of the
mantissa of the smaller number to the right until the exponents of the two number are
equal and then adding the mantissas.
28. Brief on coefficient inaccuracy. (or)
What is coefficient quantization error? What is its effect?
The filter coefficients are computed to infinite precision in theory. But, in digital
computation the filter coefficients are represented in binary and are stored in registers.
If a b bit register is used, the filter coefficients must be rounded or truncated to b
bits, which produces an error.
Due to quantization of coefficients, the frequency response of the filter may differ
appreciably from the desired response and some times the filter may actually fail to
meet the desired response and some times the filter may actually fail to meet the
desired specifications. If the poles of desired filter are close to the unit circle, then
those of the filter with quantized coefficients may lie just outside the unit circle,
leading to unstability
54
29. What is product quantization error (or) what is product round off error in
DSP? (Nov/Dec 2010)
Product quantization errors arise at the output of a multiplier. Multiplication of a b
bit data with a b bit coefficient results a product having 2b bits. Since a b bit register
is used, the multiplier output must be rounded or truncated to b bits, which produces
an error. This error is known as product quantization error.
30 What do you understand by input quantization error?
In DSP, the continuous time input signals are converted into digital using a b bit
ADC. The representation of continuous signal amplitude by a fixed digit produce an
error, which is known as input quantization error.
31. What is truncation? (Nov/Dec 2006)
Truncation is process of reducing the size of binary number by discarding all bits
less significant than the least significant bit that is retained. (In the truncation of a
binary number to b bits all the less significant bits beyond bth bit are discarded)
32. What is rounding?
Round is the process of reducing the size of a binary number to finite word sizes
of b bits such that, the rounded b-bit number is closest to the original unquantized
number.
33. What is Quantization step size? (Nov/Dec 2006)
In digital systems, the numbers are represented in binary. With b-bit binary we
can generate 2b different binary codes. Any range of analog value to be represented in
binary should be divided into 2b levels with equal increment . The 2b levels are called
Quantization levels and the increment in each level is called Quantization step size . It
R is the range of analog signal then,
Quantization step size, q = R/2b.
55
56
57