Scribd
Upload
104 views

Speech Processing

This document is a 3 page exam paper on speech processing from the University of Pune. It contains 12 questions divided into 2 sections. Section 1 has multiple choice questions on topics like the LTI model of speech production, pitch extraction methods, features of speech sounds, and linear predictive coding. Section 2 covers speech enhancement techniques, hidden Markov models, speech recognition using dynamic time warping, speaker identification and verification systems, and text-to-speech conversion. Students must answer 4 questions, with 2 from each section. Diagrams should be included where needed.

Uploaded by

Prem Kumar Apk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
104 views

Speech Processing

This document is a 3 page exam paper on speech processing from the University of Pune. It contains 12 questions divided into 2 sections. Section 1 has multiple choice questions on topics like the LTI model of speech production, pitch extraction methods, features of speech sounds, and linear predictive coding. Section 2 covers speech enhancement techniques, hidden Markov models, speech recognition using dynamic time warping, speaker identification and verification systems, and text-to-speech conversion. Students must answer 4 questions, with 2 from each section. Diagrams should be included where needed.

Uploaded by

Prem Kumar Apk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

www.vidyarthiplus.

com

[Total No. of Questions: 12] [Total No. o f Pr i n t ed Pag es : 3]

UNIVERSITY OF PUNE
[4364]-585
B. E. (E&TC) Examination - 2013
SPEECH PROCESSING
(2008 Pattern)
[Time: 3 Hours] [Max. Marks: 100]

Instructions:
1 Answer Q1 or Q2, Q3 or Q4, Q5 or Q6 from section I and Q7 or
Q8, Q9 or Q10, Q11 or Q12 from section II.
2 Answers to the two sections should be written in separate
answer-books.
3 Figures to the right indicate full marks.
4 Draw neat diagrams must be drawn wherever necessary.
5 Assume suitable data, if necessary.

SECTION -I
Q.1 A Explain LTI model for speech production system? What [8]
are their limitations?
B What is pitch? Explain the method of finding pitch of [8]
speech signal using AMDF method? What are
advantages of finding pitch period using AMDF method
over autocorrelation methods?
OR
Q.2 A Explain the terms [8]
1) Vowels 2) Diphthongs. 3) Semivowels 4) nasals
5) fricatives 6) Stops 7) Affricates 8) Consonants
B Explain the technique to separate voiced and unvoiced [8]
region of speech signal using energy and ZCR.

Q.3 A Explain Levinson –Durbin recursive algorithm for [8]


calculation of predictor coefficients.
B Explain the basic principal of linear predictive analysis? [8]
Explain the method of finding LPC coefficients using
autocorrelation method.
OR
Q. 4 A Explain the method to determine pitch period of speech [8]
signal using LPC analysis?
B Explain the basic principal of linear predictive analysis? [8]
Page 1 of 3

www.vidyarthiplus.com
www.vidyarthiplus.com

Explain the method of finding LPC coefficients using


covariance method.

Q. 5 A What is MFCC? Explain the method to calculate MFCC [10]


using block diagram?
B With the help of block diagram explain homomorphic [8]
speech processing?
OR
Q. 6 A Explain the method to evaluate the formants o speech [8]
signal using cepstrum analysis. What is importance of
formants in speech processing?
B What is complex ceptrum? Explain the procedure to [10]
evaluate cepstrum of a speech signal? State properties of
complex ceptrum.

SECTION II
Q. 7 A What is speech enhancement? Explain the spectral [8]
subtraction method of speech enhancement technique in
detail.
B Compare different speech enhancement techniques. [8]
OR
Q. 8 A What is wiener filter? Explain the method of speech [8]
enhancement using wiener filtering.
B Explain the method of speech enhancement by re- [8]
synthesis and compare with other methods.

Q. 9 A What are three basic problems associated with HMMs, [6]


Give their solution
B Draw and explain, HMM based isolated word speech [8]
recognition system.
C Write the compact form representation of HMM model [2]
(ߣ).
OR
Q. 10 A Draw and Explain, isolated digit recognition system [8]
using Dynamic time warping.
B Compare Speaker Identification system with Speaker [4]
Verification system.
C State different distortion measures used for Automatic [4]
Speech Recognition system.

Q. 11 A Draw and Explain Speaker Identification system. Also [9]


explain different performance measurement parameters
used for speaker recognition.

Page 2 of 3

www.vidyarthiplus.com
www.vidyarthiplus.com

B List out different features used for speaker recognition [9]


system and explain.
OR
Q. 12 A Explain Text-to-Speech conversion system with block [9]
schematic? State different applications of TTS.
B Explain different issues and challenges for speaker [9]
recognition system? What will be the probable solutions
to resolve issues, Explain.

Page 3 of 3

www.vidyarthiplus.com

You might also like