TWO MARKS
DIGITAL SIGNAL PROCESSING
UNIT I
1. Define DFT of a discrete time sequence.
The DFT is used to convert a finite discrete time sequence x (n) to an N point
frequency domain sequence denoted by X (K). The N point DFT of a finite duration
sequence x (n) is defined as
N-1
X (K) = x (n) e-j2 nk/N for K=0, 1, 2,N-1
n=0
2. Define IDFT.
The IDFT is used to convert the N point frequency domain sequence X (K) to an
N point time sequence. The IDFT of the sequence X (K) of length N is defined as
N-1
x (n) =1/N X (K) e+j2 nk/N for n=0, 1,2,N-1
K=0
3. List any four properties of DFT.
Let DFT{x (n)} =X (K), DFT{x1 (n)} =X1 (K), DFT{x2 (n)} =X 2(K)
i)
ii)
iii)
iv)
divide and conquers approach. This is based on the decomposition of an N point DFT
into successively smaller DFT.
In an N point sequence if N can be expressed as N= r m then the sequence can be
decimated into r- point sequences. For each r- point sequence, r point DFT is computed.
From the results of r- point DFTs the r 2 - point DFTs are computed. From the results of
r 2- point DFT the r3- point DFT are computed and so on, until we get r m point DFT.
Hence the number of stage of computation is m. The number r is called the radix of the
FFT algorithm.
7. What is radix- 2 FFT?
The radix- 2 FFT is an efficient algorithm for computing N point DFT of an N
point sequence. In radix- 2 FFT the N point sequence is decimated into 2 -point sequence
and the 2 point DFT for each decimated sequence is computed. From the result of 2point DFT the 4- point DFT are computed. From the result of 4- point DFT the 8- point
DFT are computed. From the result of 4- point DFT, the 8 point DFT are computed and
so on until we get N point DFT.
8. How many multiplication and addition are involved in radix- 2 FFT?
The total number of complex addition is N log N and the total numbers of
complex multiplication are N/2log N.
9. How many multiplication and addition are involved in DFT?
The total numbers of complex additions are N (N-1) and the total number of
complex multiplications are N 2
10. What is twiddle factor (or) phase factor?
The complex number WN is called phase factor or twiddle factor.
WN=e-j2 /N
11. Draw and explain the basic butterfly diagram (or) flow graph of DIT radix-2
FFT?
The basic butterfly diagram of DIT radix-2 FFT is shown below. It performs the
following operations .Here a & b are the input complex number and A&B are output
complex number.
i) Input complex number b is multiplied by the phase factor WN K.
ii) The product bWNK is added to the input complex number a to from a new
complex number A.
iii) The product bWNK is subtracted to the input complex number a to from a new
complex number B.
12. What is DIT radix-2 FFT?
The DIT (Decimal in Time) radix-2 FFT is an efficient algorithm for computing
DFT. In DIT radix-2 FFT, the decimal N-point sequence is decimated into 2-point
sequences. The results of 2-point DFT are used to compute 4-point DFTs. Two numbers
of 2-point DFT are combined to get a 4-point DFT. The results of 4-point DFTs are used
to compute 8-point DFTs. Two numbers of 4-point DFTs are combined to get an 8-point
DFT. This process is continued until we get N-point DFT.
13. What is DIF radix-2 FFT?
The DIF radix -2 FFT is an efficient algorithm for computing DFT. In this
algorithm the N point time domain sequence is converted to two numbers of N/2 point
sequences. Then each N/2 point sequence is converted to two number of N/4 point
sequences. This process continued until we get N/2 numbers of 2 point sequences. Now
the 2 point DFT of N/2 number of 2 point sequence will give N samples, which is the N
point DFT of time domain sequence .Here the equation for forming N/2 point
sequences,N/4 sequences etc., are obtained by decimation of frequency domain
sequences. Hence this method is called DIF.
14. Draw and explain the basic butterfly diagram or flow graph of dif radix2 FFT?
The basic butterfly diagram of DIF radix2 FFT is shown below. It performs the
following operations .Here a & b are the input complex number and A&B are output
complex number.
i) The sum of the complex number a and b are computed to form a new complex
number A.
ii) The complex number b is subtracted from a to get the difference a-b.
iii) The difference term a-b is multiplied with phase factor WNK to from a new complex
number B.
15. Compare the DIT and DIF.
DIT radix-2 FFT
1.When the input is bit reversed order, the
output will be in normal order .
2.In each stage of computation the phase
factor are multiplied before add and
subtract operation
3.The value of N should be expressed such
that N=2 m and this algorithm consists of
m stage of computation.
4.Total number of arthemetric operations is
N log N complex addition and N/2logN
complex multiplications.
UNIT II
1. What are the types of digital filter according to their impulse response?
i)IIR (Infinite Impulse Response) filter
ii)FIR ( Finite Impulse Response) filter
2. What are the advantages of FIR filter?
1. FIR filter have exact linear phase.
2. FIR filters are always stable.
3. FIR filter can be realized in both recursive and recursive structures.
4. Excellent design methods are available for various kinds of FIR filter.
5. FIR filter are free of limit cycle oscillation, when implemented on a finite word
length digital system.
3. What are the disadvantages of FIR filter?
i) Memory requirement and execution time are very high.
ii The implementation of narrow transition band fir filters is very costly, as it
requires considerably more arithmetic operation and hardware components
such as multipliers, adders and delay elements.
4. What is the necessary and sufficient condition for the linear phase characteristic
of a FIR filter?
The phase function should be a linear function of w, in which requires constant
group delay and phase delay.
(w)=-w
for satisfying above condition
h (n)=h(N-1-n)
i.e. The impulse response must be symmetrical about =(N-1)/2
If only constant group delay is desired then
(w)=-w
for satisfying above condition
h (n)=-h(N-1-n)
i.e. The impulse response must be symmetrical about =(N-1)/2
5. Distinguish IIR and FIR filters
FIR
Impulse response is finite
IIR
Impulse Response is infinite
6. List the well known design technique for linear phase FIR filter design?
1. Fourier series method
2. Window method
3. Frequency sampling method.
7. What is Gibbs phenomenon?
OR
What are Gibbs oscillations?
One possible way of finding an FIR filter that approximates H d(e j)would be
to truncate the infinite Fourier series at n= (N-1/2).Abrupt truncation of the
series will lead to oscillation both in pass band and is stop band .This
phenomenon is known as Gibbs phenomenon.
8. What are the desirable characteristics of the windows?
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 lobes of the frequency response should decrease in energy
rapidly as w tends to
9. Under what conditions a finite duration sequence h(n) will yield constant group
delay in its frequency response characteristics and not the phase delay?
If the impulse response is anti symmetrical, satisfying the condition
h (n)=-h(N-1-n)
The frequency response of FIR filter will have constant group delay and not the
phase delay .
10. Give the equation specifying rectangular window function.
-(N-1)/2n(N-1)/2
otherwise
- (N-1)/2n (N-1)/2
otherwise
Rectangular Window
Kaiser window
There are a few terms used to describe the behavior and performance of FIR filter
including the following:
Filter Coefficients - The set of constants, also called tap weights, used to multiply
against delayed sample values. For an FIR filter, the filter coefficients are, by
definition, the impulse response of the filter.
Impulse Response A filters time domain output sequence when the input is an
impulse. An impulse is a single unity-valued sample followed and preceded by
zero-valued samples. For an FIR filter the impulse response of a FIR filter is the
set of filter coefficients.
Tap The number of FIR taps, typically N, tells us a couple things about the filter.
Most importantly it tells us the amount of memory needed, the number of
calculations required, and the amount of "filtering" that it can do. Basically, the
more taps in a filter results in better stopband attenuation (less of the part we want
filtered out), less rippling (less variations in the passband), and steeper rolloff (a
shorter transition between the passband and the stopband).
Multiply-Accumulate (MAC) In the context of FIR Filters, a "MAC" is the
operation of multiplying a coefficient by the corresponding delayed data sample
and accumulating the result. There is usually one MAC per tap.
UNIT III
1.. How can you design a digital filter from analog filter?
Digital filter can de designed from analog filter using the following methods
1. Impulse invariant method
2. Bilinear transformation
2. State the steps to design digital IIR filter using bilinear method.
Substitute s by 2/T (z-1/z+1), where T=2/ (tan (w/2) in H(s) to get H (z).
3. Give the bilinear transform equation between s plane and z plane
s=2/T (1-z-1/1-z+1)
Where N is the order of the filter and c is the cutoff frequency. The magnitude
response of Butterworth filter closely approximates the ideal response as the order N
increases. The phase response becomes more non linear as N increases.
11. Give any properties of butterworth filters.
i)The magnitude response of butterworth filter closely approximates the ideal
response as the order N increases
ii) The magnitude response of butterworth filter decreases monotonically as the
frequency increases from 0 to
iii)the Poles of the butterworth filter lies on a unit circle.
17. What are the disadvantages of IIR filters (compared to FIR filters)?
They are more susceptable to problems of finite-length arithmetic, such as
noise generated by calculations, and limit cycles. (This is a direct
consequence of feedback: when the output isn't computed perfectly and is
fed back, the imperfection can compound.)
They are harder (slower) to implement using fixed-point arithmetic.
They don't offer the computational advantages of FIR filters for multirate
(decimation and interpolation) applications.
18. Butterworth filter
no gain ripple in pass band and stop band, slow cutoff
19. Chebyshev filter(Type I)
no gain ripple in stop band, moderate cutoff
20. Chebyshev filter(Type II)
no gain ripple in pass band, moderate cutoff
21. Bessel filter
- no group delay ripple, no gain ripple in both bands, slow gain cutoff
22. Elliptic filter
gain ripple in pass and stop band, fast cutoff
23. How do I skip needless calculations?
First, if your filter has zero-valued coefficients, you don't actually have to calculate those
taps; you can leave them out. A common case of this is "half-band" filter, which have the
property that every-other coefficient is zero.
Second, if your filter is "symmetric" (linear phase), you can "pre-add" the samples which
will be multiplied by the same coefficient value, prior to doing the multiply. Since this
technique essentially trades an add for a multiply, it isn't really useful in DSP
microprocessors which can do a multiply in a single instruction cycle. However, it is
useful in ASIC implementations (in which addition is usually much less expensive than
multiplication); also, some newer DSP processors now offer special hardware and
instructions to make use of this trick.
The impulse response is a mathematical concept that can be approximated in the real
world. It is the output of a circuit when an ideal impulse (zero width pulse with unit area)
is applied to the input. The Laplace transform deals with this.
The spectrum of an ideal impulse is flat and so the frequency shape of the resulting
output of the network is the frequency response of the network
UNIT IV
7. What are the three-quantization errors to finite word length registers in digital
filters?
1. Input quantization error
2. Coefficient quantization error
3. Product quantization error
8. How the multiplication & addition are carried out in floating point arithmetic?
In floating point arithmetic, multiplication are carried out as follows,
Let f1 = M1*2c1 and f2 = M2*2c2.
Then f3 = f1*f2 = (M1*M2) 2(c1+c2)
That is, mantissa is 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 numbers
are equal and then adding the mantissas.
9. What do you understand by input quantization error?
In digital signal processing, the continuous time input signals are converted into
digital using a b-bit ACD. The representation of continuous signal amplitude by a fixed
digit produce an error, which is known as input quantization error.
10. What is product quantization error?
Product quantization error arises at the output of a multiplier. Multiplication of b
bit data with a b bit coefficient result a product having 2b bits. Since b bit register is used
the multiplier output must be rounded or truncated to b bits, which produce an error. This
error is known as quantization error.
11. What is coefficient quantization error?
The filter coefficient is computed to infinite precision theory. But in digital
computation the filter coefficient are represented in binary and are stored in registers.
If b bit register is used the filter coefficient must be rounded or truncated to b bit which
produce an error. Due to quantization of coefficient the frequency response of filter may
differ appreciably from the desired response and some time 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.
12. What are the different quantization methods?
The common methods of quantization are
1. Truncation 2. Rounding
13. What is the relationship between truncation error e(n) and the bits b for
representing a decimal into binary?
For a 2's complement representation, the error due to truncation for both positive and
negative values of x is
-2-b e (n) 0
Where b is the number of bits
14. What is meant rounding? Discuss its effect on all types of number
representation?
Rounding a number to b bits is accomplished by choosing the rounded result as the b
bit number closest to the original number unrounded.
For fixed point arithmetic, the error made by rounding a number to b bits satisfy the
inequality
-2-b
2-b
----- e (n) -------2
2
for all three types of number systems, i.e., 2's complement, 1's complement & sign
magnitude.
For floating point number the error made by rounding a number to b bits satisfy the
inequality
-2-be (n) 2-b
where e (n) =xT-x
15. What is truncation?
Truncation is a process of discarding all bits less significant than least significant
than least significant bit that is retained.
0.00110011 to 4 bit 0.0011.
16. What is meant by A/D conversion noise?
A DSP contains a device, A/D converter that operates on the analog input x (n) to
produce xq(n) which is binary sequence of 0s and 1s.
At first the signal x(t) is sampled at regular intervals to produce a sequence x(n) is of
infinite precision. Each sample x(n) is expressed in terms of a finite number of bits given
the sequence xq(n). The difference signal e (n)=xq(n)-x(n) is called A/D conversion
noise.
17. What is the effect of quantization on pole location?
Quantization of coefficients in digital filters lead to slight changes in their value.
This change in value of filter coefficients modify the pole-zero locations. Some times the
pole locations will be changed in such a way that the system may drive into instability.
26. Explain briefly the need for scaling in the digital filter implementation.
To prevent overflow, the signal level at certain points in the digital filter must be
scaled so that no overflow occurs in the adder.
27. Why the limit cycle problem does not exist when FIR digital filter is realized in
direct form or cascade form?
In case of FIR filter there are no limit cycle oscillations, if the filter is realized in
direct form or cascade form since these structures have no feedback.
28. Why rounding is preferred to truncation in realizing digital filter?
1. The quantization error due to rounding is independent of type of arithmetic.
2. The mean of rounding error is zero.
3. The variance of the rounding error signal is low.
29. What is the steady state variance of the noise in the output due to quantization of
the input for the first order filter?
y(n)=ay(n-1)+x(n)
e 2 =(2-2b/12)(1/1-a2)
30. Compare the fixed point and floating point arithmetic.
Fixed point arithmetic.
Fast operation
Slow operation
Relatively economical
31. Express the fraction 7/8 and -7/8 in sign magnitude, 2s complement and 1s
complement.
Fraction (7/8) = (0.111)2 in sign magnitude, 2s complement and 1s complement.
Fraction (-7/8)= (1.111)2 in sign magnitude
(1.000)2 1s complement.
(1.001)2 2s complement.
32. Give the expression for signal to quantization noise ratio and calculate the
improvement with an increase of 2 bit to the existing bit.
SNR=6.02b+10.79+10log10 x 2
With an increase of 2 bits, increase in SNR is approximately 12 Db
UNIT V
Sxx(f)= | x(n)e-j2Kf |2
n= -
2. What is power density spectrum?
Let x(t) be a stationary random process. The statistical autocorrelation function
for this signal is
xx()=E[x*(t)x(t+)]
The Fourier transform of the autocorrelation function of a stationary random process
gives their power density spectrum
xx(f)=F(xx())
= xx() e-j2f d
-
3. Using indirect method, how to find the energy density spectrum?
It requires two steps:
i)
First, the autocorrelation rxx(k) is computed from x(n)
ii)
The Fourier transform of the autocorrelation is computed.
r xx(k)= x*(n)x(n+k)
n= -
Sxx(f)= r xx(k)e-j2Kf
n= -
4. What is three types of non parametric methods of power spectrum estimation?
i). Bartlett methods (Averaging periodogram)
ii) Welch method (modified Averaging periodogram)
iii) Blackman and tukey methods (smoothing periodogram)
5. What are nonparametric methods?
These methods make no assumption about how the data were generated and hence
are called nonparametric methods.
6. Define Periodogram.
Schuster defined the periodogram as a method to discover the frequencies of the
hidden harmonics signal. The estimate Pxx (f) can also be expressed as
N-1
Pxx(f)=1/N | x(n) e-j2fn |2 =1/N|X(f)|2
n= 0
Where X(f) is the Fourier transform of the sample sequence x(n).This well
known form of the power density spectrum estimate is called the Periodogram.
7. What is a Bartlett method?
In this method to reduce the variance of the Periodogram,
Three steps
i) First divide the N point sequence x (n) into k non overlap subsequence of length M
ii) Find the periodogram for each sub sequence.
iii) Calculate the average periodogram of k subsequence.
8. What is the advantage of Bartlett window methods?
The effect of reducing the length of data from N point to M=N/K, result in a
window whose spectral width has been increased by a factor of K. consequently, the
frequency resolution has been reduced by a factor K. in return in resolution we reduced
the variance.
9. What is the difference between Bartlett and welchs methods?
i) First difference is that welchs method allows overlapping of data sequence. The
overlap 50% or 75%.
ii) The second difference is the data with in a sequence are windowed prior to computing
the periodogram.
10. What is a Blackman and turkey method?
Blackman and turkey proposed a method in which the sampled autocorrelation
sequence is windowed first and then Fourier transformed.
11. Define quality of nonparametric methods?
The ratio of its the square of the mean of power spectrum estimate to variance.
Qa=E[Pxx(f)]2/var[Pxx(f)]
12. Define variability.
Quality factor
1.11Nf
1.39Nf
2.34 Nf
X =E[X] =xfX(x) dx
-
where E[X] is read ``the expected value of X''. Other names for the same mean
value x or the expected value E[X] are average value and statistical average.
18. What are the two properties of power spectrum estimator?
i) Bias
ii) Variance
19. Define the bias of estimator?
Bias of estimator is defined as the true value of the parameter minus the expected
value of the estimator.
20. Define unbiased estimator.
An unbiased estimator is one for which the bias is 0. This then means that the
expected value of the estimator is the true value so that the probability density is
symmetrical then its center would be at the true value.
r xx(k)= x*(n)x(n+k)
k=0,1,2,.....
n= -
or equivalently,
r xx(k)= x*(n)x(n-k)
k=0,1,2,.....
n= -
Cross correlation
For two sequence x(n) and y(n),the cross correction function r xy(k) is defined as
r xy(k)= x*(n)y(n+k)
n= -
or equivalently,
r xy(k)= x*(n)y(n-k)
n= -
k=0,1,2,.....
k=0,1,2,.....