answer all the questions in the attached file

INSTITUTE OF SPACE TECHNOLOGY

Digital Signal Processing

BS-EE

Final Exam

Fall 2018 12 January 2018

Time Allowed: 3 Hrs Max Marks: 120

Instructions:

  • Attempt all Questions.

  • You may refer to your text books during the exam.

  • Please show your complete reasoning to receive credit.

  • In case of confusion, go on to the next question until the instructor comes to you.


  1. Answer the following Questions. Show your reasoning. (60 points)

    1. A continuous-time signal, is sampled at a sampling frequency Hz. Find the sampled discrete-time sequence . [1 point]

    2. A continuous-time signal is given as . Is this signal periodic? If it is, find the fundamental period. [1 point]

    3. A discrete-time signal is given as . Is this signal periodic? If it is, find the fundamental period. [1 point]

    4. A discrete-time signal is given as . Is this signal periodic? If it is, find the fundamental period. [2 points]

    5. The non-zero values of a discrete-time signal are given as and . Decompose into conjugate symmetric (even) and conjugate anti-symmetric (odd) signals. [3 points]

    6. The input-output relation of a discrete-time system is given as . Is this system: [4 points]

      1. Linear?

      2. Time-invariant?

      3. BIBO stable?

      4. Memoryless?

    7. The input-output relation of a discrete-time system is given as . Is this system: [4 points]

      1. Linear?

      2. Time-invariant?

      3. BIBO stable?

      4. Memoryless?

    8. A discrete-time system is given as , where is a real scalar constant. Find:

      1. The impulse response of the system. [3 points]

      2. The range of values of for which the system is BIBO stable. [1 point]

    9. Two discrete-time systems with impulse responses and are combined to form a discrete-time system h.

      1. Find the impulse response of if the two systems are connected in parallel. [2 points]

      2. Find the impulse response of if the two systems are connected in cascade. [3 points]

    10. Sketch suitable pole-zero diagrams for following filters:

      1. Stable Notch filter with notch frequencies and . [2 points]

      2. Stable Digital resonator with resonant frequencies and . [2 points]

      3. All pass filter with two zeros at and . [2 points]

    11. An unstable discrete-time system is described by the transfer function

      1. Sketch the magnitude of the frequency response, from . [3 points]

      2. If, for , the input to the system is

find the frequencies and for which the output will be unbounded. [4 points]

      1. If, for , the input to the system is

find the frequencies and for which the output will be zero. [4 points]

    1. The communications channel in a particular scenario can be modeled as , where . Find the difference equation of the channel equalizer (the inverse system that cancels the effect of the channel). [3 points]

    2. The impulse response of an ideal low pass filter with cut-off frequency is given as

Use this result to answer the following questions.

      1. Find the impulse response of an ideal high pass filter with cut-off frequency . [1 point]

      2. Find the impulse response of an ideal band pass filter with pass band frequencies from to . [2 points]

      3. Find the impulse response of an ideal band stop filter with stop band frequencies from to . [2 points]

    1. A discrete-time signal is given as

,

where is the window. How would you change and so that the two frequencies in the spectrum estimate are clearly distinguishable? [3 points]

    1. A discrete-time signal is given as

,

where is the window. How would you change and so that the two frequencies in the spectrum estimate are clearly distinguishable? [3 points]

    1. Two periodic signals with period are described as

Find the result of convolution of and . [3 points]

    1. For fractional sampling rate conversion, will you use an up-sampler first or a down-sampler? [1 point]

  1. (Sampling of Continuous-time signals, Multirate Signal Processing) (20 points)

A Continuous-time signal, , is defined as

.

A discrete-time signal, , is obtained by sampling at a rate . The sampled signal, , is then input to an upsampler followed by a downsampler. The output of the downsampler, , is then input to an ideal interpolator (Digital to Analog converter) to obtain the continuous-time signal , as shown in figure.

w[n]

y[n]

x[n]

D/A converter

Sampler


Analyze this system by answering the following questions.

  1. Find an equation for the sampled signal, and sketch the magnitude of its DTFT, from . [4 points]

  2. Find an equation for the signal, and sketch the magnitude of its DTFT, from . [7 points]

  3. Find an equation for the signal, and sketch the magnitude of its DTFT, from . [7 points]

  4. Find an equation for the continuous time signal at the output of the system. [2 points]


  1. (Hilbert Transformer) [10 points]


h1 [n]

h2 [n]


Consider the Discrete-time system shown in the diagram above. The filters, are are given as

Note that the second term in and is the Hilbert transformer. In this question, assume

    1. Find the equation of the discrete-time signal and plot the DTFT from . [2 points]

    2. Find the equation of the discrete-time signal and plot the DTFT from . [2 points]

    3. Find the equation of the discrete-time signal and plot the DTFT from . [2 points]

    4. Draw a discrete-time system that can be used to recover , from . [4 points]

  1. (Correlation, Discrete-Fourier Transform) (15 points)

    1. A continuous-time signal is sampled at a sampling rate . A 60 point Discrete Fourier Transform, , is then computed for the sampled signal. For what value of would you expect to observe a peak in ?[9 points]

    2. A length 4 sequence, , is to be correlated with a length 5 sequence . Draw a block diagram to show how this linear correlation can be implemented using FFT algorithm. (Note: You do not need to draw the butterfly diagram in this part. Just assume that you have a block that takes the FFT. Clearly mention the number of points in FFT block). [2 points]

    3. Derive the butterfly diagram for 4 point radix 2 FFT algorithms. [4 points]


  1. (Design of Digital filters) (15 points)

    1. The impulse response of an FIR low pass filter is given as . Find the phase of this filter. [4 points]

    2. A low pass FIR filter is to be designed using windowing technique. The filter is required to have a minimum stop band attenuation of 50dB and a maximum transition bandwidth of up to . Answer the following questions using the table given below: [5 points]

      1. Which window will you select to meet the specifications with minimum filter order?

      2. What is the length of the filter?

Window

Minimum Stop band attenuation

Transition Bandwidth

Rectangular

20.9dB

Hann

43.9dB

Hamming

54.5dB

Blackman

75.3dB

Page 5/4