Speech Recognition MCQs

1) Commands that make speaking to a speech recognition system almost as natural as talking to a person are known as NLT, an acronym for ______.

A) Natural Lab Technology         B) Natural Language Technology

C) Neuron Language Technology D) Neutral Language Technology

Correct Ans: Natural Language Technology

2) Set of mathematical formulas that analyzes human voice patterns and evaluates the probability of spoken words matching words in the speech recognition vocabulary ______.

A) Acoustic model B) Diagram model C) Gram model D) Speech model

Correct Ans: Acoustic model

3) The first step when using the speech recognition software is to ______.

A) complete the initial enrollment B) properly adjust microphone volume level

C) save the document D) say your name

Correct Ans: complete the initial enrollment

4) What kind of signal is used in speech recognition?

A) Electromagnetic signal B) Electric signal C) Acoustic signal D) Radar

Correct Ans: Acoustic signal

5) What is called as the properties of the signal that extend over interval?

A) Hops B) Rates C) Frames D) All of the mentioned

Correct Ans: Frames

6) What is the name of elements present in patterns?

A) variables B) patterns C) pattern variables D) pattern elements

Correct Ans: pattern variables

7) Where does the Hidden Markov Model is used?

A) speech recognition B) understanding of real world

C) both A & B D) none of the mentioned

Correct Ans: speech recognition

8) Which algorithm works by first running the standard forward pass to compute?

A) smoothing B) modified smoothing C) HMM D) depth-first search

Correct Ans: modified smoothing

9) Which data structure is used to give better heuristic estimates?

A) Forwards state-space B) Backward state-space  

C) Planning graph algorithm D) None of the mentioned

Correct Ans: Planning graph algorithm

10) Which of the following are the types of Spectral analysis models?

A) Acoustic phonetic model B) Pattern recognition model

C) both A & B D) None of the mentioned

Correct Ans: both A & B

11) The actual number of filters used in the filter bank is _____.

A) Q < = N - 2 B) Q < = N + 2 C) Q <  N - 2 D) Q <  N + 2

Correct Ans: Q < = N - 2

12) The vector quantization is the intension of ______.

A) LPC derived vector B) Scalar quantization C) LPC analysis vector D) LPC

Correct Ans: Scalar quantization

13) Mean absolute log spectral distortion is defined by ______.

A) P = 1 B) P = 2 C) P < 1 D) P > 2

Correct Ans: P = 1

14) Typical constraints defined by Time Normalization constraint :

A) Endpoint constraint B) Monotonicity condition

C) Global path constraint D) All of the mentioned

Correct Ans: All of the mentioned

15) Adverse conditions in speech recognition : 

A) Noise     B) Distortion C) Articulation Effect D) All of the mentioned

Correct Ans: All of the mentioned

16) We can apply single enhancement method to ______ the noise before applying recognition algorithm.

A) suppress B) compress C)detect D)decompress

Correct Ans: suppress

17) HMM is a statistical method of characterizing the ______.

A) random probability distribution     B) spectral properties of frames

C) classical pattern classifier     D) none of the mentioned

Correct Ans: spectral properties of frames

18) In coin tossing experiment, the problems are ______.

A) what are the states in the model B) How many states should be in the model

C) what are the state transition probabilities D) All of the mentioned

Correct Ans: All of the mentioned

19) The lower constraint line in multiple level is represented as ______.

A) L(m) = (m + 1 )/2 B) L(m) = (m - 1 )/2 C) L(m) = (m + 1 )*2 D) L(m) = (m - 1 )*2

Correct Ans: L(m) = (m + 1 )/2

20) "_______ is sounds of speech through the analysis of the speech waveform and spectrum"

A) phonetics B) auditory phonetics C) acoustic phonetics D) articulatory phonetics

Correct Ans: acoustic phonetics