Scribd
Upload
2 views

Advanced Digital Signal Processing (ECE - 504) (Makeup)

This document outlines the examination details for the Advanced Digital Signal Processing course at Manipal Institute of Technology, including instructions for candidates and a list of questions covering various topics such as cepstrum, homomorphic systems, optimum predictors, decimation filters, and PSD estimation. Candidates are required to answer any five full questions from the provided topics. The exam is scheduled for May/June 2013, lasting three hours with a maximum score of 50 marks.

Uploaded by

paula
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

Advanced Digital Signal Processing (ECE - 504) (Makeup)

This document outlines the examination details for the Advanced Digital Signal Processing course at Manipal Institute of Technology, including instructions for candidates and a list of questions covering various topics such as cepstrum, homomorphic systems, optimum predictors, decimation filters, and PSD estimation. Candidates are required to answer any five full questions from the provided topics. The exam is scheduled for May/June 2013, lasting three hours with a maximum score of 50 marks.

Uploaded by

paula
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Reg. No.

MANIPAL INSTITUTE OF TECHNOLOGY


Manipal University
SECOND SEMESTER M.Tech. (DEAC) DEGREE END SEMESTER EXAMINATION
May/June 2013

SUBJECT: ADVANCED DIGITAL SIGNAL PROCESSING (ECE - 504)

TIME: 3 HOURS MAX. MARKS: 50

Instructions to candidates
• Answer ANY FIVE full questions.
• Missing data may be suitably assumed.

1A. Define complex cepstrum of a signal and explain the properties . Find and plot the
complex cepstrum of the signal x(n) = cos(πn/3) u(n).

1B. Define real cepstrum of a signal. Give the relationship of the same with complex
cepstrum. Giving relevant mathematical relations, explain how complex cepstrum is
estimated from real cepstrum. (5+5)

2A. With block diagram explain the Homomorphic system for de-convolution. Discuss
the implementation of the same using DFT.

2B. Voiced speech s(n) is assumed to be s(n) = v(n) * p(n) where v(n) is the impulse
response of vocal tract and p(n) is the source signal. With relevant mathematical
analysis, explain how the fundamental frequency of voicing can be estimated using
cepstral evaluation. (5+5)

3A. Explain the concept of optimum predictor. Obtain solution for tapped delay line
filter coefficients for one step optimum predictor.

3B. Derive LMS algorithm for updating the tap weights of tapped delay line filter.
Discuss on the convergence of LMS algorithm. (5+5)

4A. Certain signal x(n) band limited to π/M is decimated by a factor M. Obtain and plot
the spectrum of the decimated signal. Assume suitable spectrum for x(n). Explain
how the original signal is recovered from the decimated signal.
4B. Explain the multistage implementation of decimation filter for decimation by large
factor. Discuss in detail how the number of stages and order of these stages are
considered. (5+5)

ECE 504 Page 1 of 2


5A. Explain IFIR technique for designing decimation filter. Design IFIR filter for the
decimation of 8 KHz signal by 50. The filter has pass band deviation of 0.01 and stop
band magnitude less than 0.001. The signal occupies a band width of 0 to 80 Hz.

5B. Derive type-1 poly-phase structure. Implement decimation filter using poly phase
components and discuss about the computational efficiency. (5+5)
6A. Obtain filter bank for DFT. Plot the frequency response of these filters. Implement
using polyphase and discuss about the computational advantage.

6B. Discuss the spectral leakage and resolution problems involved in estimation of PSD.

6C. Explain the principles of parametric method of PSD estimation. (5+3+2)

*************************

ECE 504 Page 2 of 2

You might also like