DIGITAL SIGNAL PROCESSING

1. Compare DIT-FFT and DIF-FFT.

2. Compare FIR and IIR filter.

Basis for Comparison FIR Filter IIR Filter
Stands for Finite Impluse Response Infinite Impulse Response

Nature Non-recursive Recursive

Computational Efficiency Less Comparatively more

Usage Difficult Quite easy

Feedback Absent Present
Stability More Less
Requirement to generate current Present and past samples of Present and past samples of input along with
output input. past output.
Delay offered High Comparatively lower
Transfer function Only zeros are present. Both poles and zeros are present.

Memory requirement More Less

Sensitivity Less Comparatively more
Resolution offered at low Less More
frequencies
Controllability Easy Quite Difficult

3. In what condition Z-transform reduces to fourier transform?

Ans- If we replace the complex variable z by e –jω, then z transform is reduced to Fourier transform.
 4. What do you mean by Region of convergence (ROC)? What are the Properties of ROC?
Ans- Region of Convergence (ROC) is defined as the set of points in s-plane for which the Laplace
transform of a function x(t) converges. In other words, the range of Re(s). (i.e.,σ) for which the
function X(s) converges is called the region of convergence.
Properties of ROC of Z-Transform
The region of convergence (ROC) of Z-transform has the following properties −
1. The ROC of the Z-transform is a ring or disc in the z-plane centred at the origin.
2. The ROC of the Z-transform cannot contain any poles.
3. The ROC of Z-transform of an LTI stable system contains the unit circle.
4. The ROC of Z-transform must be connected region. When the Z-transform X(z) is a rational, then its
ROC is bounded by poles or extends up to infinity.
5. For x(n)=δ(n), i.e., impulse sequence is the only sequence whose ROC of Z-transform is the entire z-
plane.
6. If x(n) is an infinite duration causal sequence, then its ROC is |z|> a , i.e., it is the exterior of a circle of
the radius equal to a.
7. If x(n) is an infinite duration anti-causal sequence, then its ROC is |z|< b , i.e., it is the interior of a circle
of the radius equal to b.
8. If x(n) is an infinite duration two-sided sequence, then its ROC is a<|z|<b , i.e., it consists of a ring in
the z-plane, which is bounded on the interior and exterior by a pole and does not contain any poles.
9. If x(n) is a finite duration causal sequence (i.e., right-sided sequence), then its ROC is the entire z-plane
except at z = 0.
10. If x(n) is a finite duration anti-causal sequence (i.e., left sided sequence), then its ROC is the entire z-
plane except at z = ∞
11. If x(n) is a finite duration two-sided sequence, then its ROC is the entire z-plane except at z = 0
and z=∞
12. The ROC of the sum of two or more sequences is equal to the intersection of the ROCs of these
sequences.

5. Compare between circular and linear convolution.

6. What are the various FIR filter design method? Which window is most commonly used for FIR
filter design and why?
Ans- There are essentially three well-known methods for FIR filter design namely:
(1)The window method
(2)The frequency sampling technique
(3) Optimal filter design methods
The side lobe of Gaussian window, Hamming window, Kaiser window and Blackman window are -57.2, -42.5,
-58.3, and -58.1 respectively. So, we can say that that Kaiser window is better than the other window.
 7. State and prove any Four properties of discrete fourier transform.

1. Periodicity
Let x(n) and x(k) be the DFT pair then if
x(n+N) = x(n) for all n then
X(k+N) = X(k) for all k
Thus periodic sequence xp(n) can be given as

2. Linearity
The linearity property states that if

DFT of linear combination of two or more signals is equal to the same linear combination of DFT of individual
signals.
3. Multiplication
The Multiplication property states that if

It means that multiplication of two sequences in time domain results in circular convolution of their DFT s in
frequency domain.
4. Time reversal of a sequence
The Time reversal property states that if

It means that the sequence is circularly folded its DFT is also circularly folded.
 8. Compare Rectangular and hanning window with the help of required equations.

9. State Nyquist rate.

Ans- The theoretical minimum sampling rate at which a signal can be sampled and still can be reconstructed
from its samples without any distortion is called the Nyquist rate of sampling.
Mathematically,
NyquistRate,fN=2fm

10. Write the conditions to define stability of ROC.

[1] System z-plane pole(s) lie outside the unit circle: System impulse response increases, with time, toward
±infinity. System frequency response does not exist. System is unstable.

[2] System z-plane pole(s) lie on the unit circle: System impulse response remains non-zero and finite for all
time. System frequency response exists but contains infinite-magnitude value(s). System is conditionally
stable.

[3] System z-plane pole(s) lie inside the unit circle: System impulse response decreases, with time, toward
zero. System frequency response exists and contains no infinite-magnitude value(s). System is stable.

11. Differentiate between causal and non-causal systems.

12. What do you mean by BIBO stability?
Bounded input, bounded output (BIBO) stability is a form of stability often used for signal processing
applications. The requirement for a linear, shift invariant, discrete time system to be BIBO stable is for
the output to be bounded for every input to the system that is bounded.
13. Illustrate the proof of the convolution property of Z transform.
Statement - The convolution in time domain property of Z-transform states that the Z-transform of the
convolution of two discrete time sequences is equal to the multiplication of their Z-transforms.
Therefore, if,
𝒁𝑻
𝒙𝟏(𝒏) ↔X1(z); ROC = R1
𝒁𝑻
𝒙𝟐 (𝒏) ↔ 𝑿𝟐 (𝒛); 𝑹𝑶𝑪 = 𝑹𝟐
Accorcding to the convolution property,
𝒁𝑻
𝒙𝟏 (𝒏) ∗ 𝒙𝟐 (𝒏) ↔ 𝑿𝟏 (𝒛)𝑿𝟐 (𝒛); 𝑹𝑶𝑪 = 𝑹𝟏 ∩ 𝑹𝟐

Proof
The convolution of two sequences is defined as,
x1(n)∗x2(n)= ∑∞ 𝒌 ∞ 𝐱𝟏(𝐤)𝐱𝟐(𝐧 − 𝐤)

Now, from the definition of Z-transform, we have,

∞
𝒏
𝒁[𝒙(𝒏)] = 𝒙(𝒏) 𝒛
𝒏 ∞

[𝐱𝟏(𝐧) ∗ 𝐱𝟐(𝐧)] 𝐳 𝐧
𝐙[𝐱𝟏(𝐧) ∗ 𝐱𝟐(𝐧)] = 𝐗(𝐳) =
𝒏 ∞
∞ ∞
𝐧
𝐗(𝐳) = [ 𝐱𝟏(𝐤)𝐱𝟐(𝐧 − 𝐤)] 𝐳
𝒏 ∞ 𝐤 ∞
∞ ∞

𝑿(𝒛) = 𝒙𝟏(𝒌)𝒙𝟐(𝒏 − 𝒌) 𝒛 𝒌 𝒛 (𝒏 𝒌)

𝒏 ∞𝒌 ∞
Rearranging the order of summations, we get,

∞ ∞
𝒌 (𝒏 𝒌)
𝑿(𝒛) = 𝒙𝟏(𝒌) 𝒛 𝒙𝟐(𝒏 − 𝒌) 𝒛
𝒌 ∞ 𝒏 ∞

Substituting (n−k)=m in the second summation, we have,

∞ ∞
𝒌 𝒎 𝑿𝟏(𝒛)𝑿𝟐(𝒛)
𝑿(𝒛) = 𝒙𝟏(𝒌) 𝒛 𝒙𝟐(𝒎) 𝒛
𝒌 ∞ 𝒎 ∞
𝒁[𝒙𝟏(𝒏) ∗ 𝒙𝟐(𝒏)] = 𝑿𝟏(𝒛)𝑿𝟐(𝒛)
Or it can also be represented as,
𝒁𝑻
𝒙𝟏(𝒏) ∗ 𝒙𝟐(𝒏) ↔ 𝑿𝟏(𝒛)𝑿𝟐(𝒛); 𝑹𝑶𝑪 = 𝑹𝟏 ∩ 𝑹𝟐
14. What is the condition for checking BIBO stability.
A system is said to be input-output stable, or BIBO stable, if the poles of the transfer function (which is
an input-output representation of the system dynamics) are in the open left half of the complex plane. A
system is BIBO stable if and only if the impulse response goes to zero with time.

15. What id FFT? And Its advantages.

FFT is nothing but computation of discrete Fourier transform in an algorithmic format, where the
computational part will be reduced.
Adavntages:
1. The FFT is much faster than the DFT but requires more memory.
2. The FFT can be used to compute the DFT of a sequence that is not a power of two, while the
DFT can only be used to compute the DFT of a sequence that is a power of two.
3. The FFT is used in many applications, including image, audio, and signal processing.
16. State and prove convolution property of DFT.
The Circular Convolution property states that if

It means that circular convolution of x1(n) & x2(n) is equal to multiplication of their DFT s. Thus circular
convolution of two periodic discrete signal with period N is given by

Multiplication of two sequences in time domain is called as Linear convolution while Multiplication of two
sequences in frequency domain is called as circular convolution. Results of both are totally different but are
related with each other.
There are two different methods are used to calculate circular convolution
Graphical representation form
Matrix approach
17. Advantages of digital filters over analog filters.
Digital filters have the following advantages compared to analog filters:
 Digital filters are software programmable, which makes them easy to build and test.
 Digital filters require only the arithmetic operations of addition, subtraction, and multiplication.
 Digital filters do not drift with temperature or humidity or require precision components.
 Digital filters have a superior performance-to-cost ratio.
 Digital filters do not suffer from manufacturing variations or aging.
18. What is GIBBS Phenomenon?

The GIBBS phenomenon was discovered by Henry Wilbraham in 1848 and then rediscovered by J.
Willard Gibbs in 1899.
For a periodic signal with discontinuities, if the signal is reconstructed by adding the Fourier series,
then overshoots appear around the edges. These overshoots decay outwards in a damped oscillatory
manner away from the edges. This is known as GIBBS phenomenon and is shown in the figure
below.
19. what are the properties of chebyshev filter?
Chebyshev filters have the following characteristics:
 Minimization of peak error in the passband
 Equiripple magnitude response in the passband
 Monotonically decreasing magnitude response in the stopband
 Sharper roll-off than Butterworth filters
20. Difference between butterworth and chebyshev filters.

21. State and prove the time shifting property of Z-transform.

The Time shifting property states that if z x(n)

Thus shifting the sequence circularly by „k samples is equivalent to multiplying its z transform by 𝑧 .
 22. Derive the time reversal property of DFT.
For a continuous-time function 𝑥(𝑡), the Fourier transform of 𝑥(𝑡) can be defined as,

𝒋𝝎𝒕
𝑿(𝝎) = 𝒙(𝒕)𝒆𝟎 𝒅𝒕

Statement – The time reversal property of Fourier transform states that if a function 𝑥(𝑡) is reversed in time domain,
then its spectrum in frequency domain is also reversed, i.e., if
𝑭𝑻
𝒙(𝒕) ↔ 𝑿(𝝎)
Then, according to the time-reversal property of Fourier transform,
𝑭𝑻
𝒙(−𝒕) ↔ 𝑿(−𝝎)
Proof
Form the definition of Fourier transform, we have,

𝒋𝝎𝒕
𝑭[𝒙(𝒕)] = 𝑿(𝝎) = 𝒙(𝒕)𝒆 𝒅𝒕

𝒋𝝎𝒕
∴ 𝑭[𝒙(−𝒕)] = 𝒙(−𝒕)𝒆 𝒅𝒕

Replacing 𝑡 by (−𝑡) by (-t) in RHS of the above equation, we get,

𝒋𝝎𝒕
𝑭[𝒙(−𝒕)] = 𝒙(𝒕)𝒆 𝒅𝒕

𝒋( 𝝎)𝒕
=> 𝐹[𝒙(−𝒕)] = 𝒙(𝒕)𝒆 𝒅𝒕 = 𝑿(−𝝎)

∴ 𝐹[𝑥(−𝑡) = 𝑋(−𝜔)
Or, it can be represented as,
𝑥(−𝑡) 𝑋(−𝜔)
23. Write the advantages of digital signal processing over analog signal processing.
ADVANTAGES OF DSP OVER ASP
1. Physical size of analog systems is quite large while digital processors are more compact and light in
weight.
2. Analog systems are less accurate because of component tolerance ex R, L, C and active components.
Digital components are less sensitive to the environmental changes, noise and disturbances.
3. Digital system is most flexible as software programs & control programs can be easily modified.
4. Digital signal can be stores on digital hard disk, floppy disk or magnetic tapes. Hence becomes
transportable. Thus easy and lasting storage capacity.
5. Digital processing can be done offline.
6. Mathematical signal processing algorithm can be routinely implemented on digital signal processing
systems. Digital controllers are capable of performing complex computation with constant accuracy at high
speed.
7. Digital signal processing systems are upgradeable since that are software controlled.
24. Define steps to obtain same results from linear and circular convolution.
Ans: to obtain the same results from both convolutions, the following steps are used:-
(i) Using equations
(ii) We calculate the value of N that means number of samples contain in linear
convolution. Let us assume it is 15.
(iii) By doing Zero padding , we make the length of every sequence equal to 15 .
This means that in this case, we need to add seven zeros in x(n) as well as
h(n).
(iv) Then we perform the circular convolution. The result of circular convolution
and linear convolution will be same.
24. what are different design techniques available for IIR filter?
There are three main methods of design IIR filter, the impulse invariant method, the backward
difference method, and the bilinear z-transform.
25. Compare between Blackman and Hamming Window.

26. Compare between Rectangular and Hamming Window.

 27. .Compare Hamming window with Kaiser Window.

28. Write the procedure for FIR filter design by frequency sampling method.
1. Choose the desired frequency response Hd(w).
2. Take N-samples of Hd ( W) to generate the sequence H (K)
(Here H bar of k should come)
3. Take inverse of DFT of H (k) to get the impulse response h (n).
4. The transfer function H (z) of the filter is obtained by taking z-transform of impulse response.
29. List the characteristic of FIR filter designed using window.
a) The width of the transition band depends on the type of window.
b) The width of the transition band can be made narrow by increasing the value of N where N is the length
of the window sequence.
c) The attenuation in the stop band is fixed for a given window, except in case of Kaiser Window where it
is variable.

30. DIFFERENCE BETWEEN OVERLAP SAVE AND OVERLAP ADD METHOD

Short Notes
1. Sampling Theorem: The sampling theorem states that a continuous
continuous-time
time signal needs to be uniformly sampled at a
minimum rate in order to recover or reconstruct the original signal.
If x(t) is a low-pass continuous-time
time signal with a band limit such ththat
𝑥(ω)=0 for ω≥ωmax
is represented in the form of its samples.
Then x(t) can be recovered in its original form if the sampling frequency is greater than or equal to twice the maximum
frequency of the message signal x(t).
If ωs≥2ω𝑚𝑎𝑥 (Nyquist sampling ratee condition);
x(nTs) = x(t), n=0, ±1, ±2, ±3, ……
Here Ts is the sampling period (sec/sample).
The Nyquist sampling rate condition can also be written as 𝑓𝑠=1𝑇𝑠≥𝜔𝑚𝑎𝑥𝜋
Here fsis is the sampling frequency (sample/second).
If the Nyquist sampling rate condition
ndition is satisfied, then the original signal x(t) can be recovered by passing the sampled
signal through an ideal low pass filter with the frequency response H(ω)=T
H(ω)=Ts; when -ωs/2<ω<ωss/2and equal to zero
elsewhere.
We can do the sampling in different ways like
Pulse amplitude modulation (PAM), and
Ideal impulse sampling.
2. Gibbs Phenomenon: The GIBBS phenomenon was discovered by Henry Wilbraham in 1848 and then
rediscovered by J. Willard Gibbs in 1899
1899.
For a periodic signal with discontinuities, if the sig
signal
nal is reconstructed by adding the Fourier series, then
overshoots appear around the edges. These overshoots decay outwards in a damped oscillatory manner away
from the edges. This is known as GIBBS phenomenon and is shown in the figure below.

The amount of the overshoots at the discontinuities is proportional to the height of discontinuity and
according to Gibbs, it is found to be around 9% of the height of discontinuity irrespective of the number
of terms in the Fourier series. The exact proportion is ggiven by the Wilbraham-Gibbs
Gibbs Constant
𝟏 𝝅 𝒔𝒊𝒏𝒕 𝟏
. ∫𝟎 𝒅𝒕 − = 𝟎. 𝟎𝟖𝟗𝟒𝟖𝟗 …
𝝅 𝒕 𝟐
It may also be noted that as more number of terms in the series are added, the frequency increases and the
overshoots become sharper, but the amplitude of the aadjoining
djoining oscillation reduces, i.e., the error between the
original signal x(t) and the truncated signal xn(t) reduces except at edges as the n increases. Hence, the
truncated Fourier series approaches the original signal x(t) as the number of terms in approximation increases.

Effects of GIBBS Phenomenon

Following are some consequences of the GIBBS phenomenon −

 In signal processing, the GIBBS phenomenon is undesirable since it causes clipping from the overshoots
and ringing artifacts from the oscillations.
 In MRI, the GIBBS phenomenon causes artifacts in the presence of adjacent regions of significantly
differing signal intensity.
 The GIBBS phenomenon demonstrates a cross-pattern artifact in the discrete Fourier transform of an
image, where the images have a sharper discontinuity between boundaries at the top-bottom and left-right
of the image.
4. Auto correlation and Cross Corelation: Auto-correlation is the comparison of a time series with itself at a
different time. It aims, for example, to detect repeating patterns or seasonality.
Cross-correlation is the comparison of two different time series to detect if there is a correlation between
metrics with the same maximum and minimum values.
Cross-Correlation
To detect a level of correlation between two signals we use cross-correlation. It is calculated simply by
multiplying and summing two-time series together.
In the following example, graphs A and B are cross-correlated but graph C is not correlated to either.

Auto-Correlation
Auto-correlation is very useful in many applications; a common one is detecting repeatable patterns due to
seasonality.
The following graph clearly shows repeating patterns every 8 data points. Indeed, looking at the R code, it’s a
repeatable sequence of the numbers 1 through 8 with some random noise in the mix.
 5. Short note on overlap save and overlap add.
Overlap-Save The overlap-save procedure cuts the signal up into equal length segments with some
overlap. Then it takes the DFT of the segments and saves the parts of the convolution that correspond to
the circular convolution. Because there are overlapping sections, it is like the input is copied therefore
there is not lost information in throwing away parts of the linear convolution.

Overlap-Add The overlap-add procedure cuts the signal up into equal length segments with no overlap.
Then it zero-pads the segments and takes the DFT of the segments. Part of the convolution result
corresponds to the circular convolution. The tails that do not correspond to the circular convolution are
added to the adjoining tail of the previous and subsequent sequence. This addition results in the aliasing
that occurs in circular convolution.

6. Butterfly Diagram: In the context of fast Fourier transform algorithms, a butterfly is a portion of the
computation that combines the results of smaller discrete Fourier transforms (DFTs) into a larger DFT,
or vice versa (breaking a larger DFT up into subtransforms). The name "butterfly" comes from the shape
of the data-flow diagram in the radix-2 case. In the case of the radix-2 Cooley–Tukey algorithm, the
butterfly is simply a DFT of size-2 that takes two inputs (x0, x1) (corresponding outputs of the two sub-
transforms) and gives two outputs (y0, y1) by the formula :
Y0=x0+x1
Y1=x0-x1
If one draws the data-flow diagram for this pair of operations, the (x0, x1) to (y0, y1) lines cross and resemble the
wings of a butterfly.
 7. Rectangular Window: The (zero
(zero-centered) rectangular window may be defined by

where is the window length in samples (assumed odd for now). A plot of the rectangular window
appears in Fig for length . It is sometimes convenient to define windows so that their dc gain is 1, in

which case we would multiply the definition above by .