Explore topic-wise InterviewSolutions in .

Right choice is (d) Code excited LPC

Best explanation: Code excited LPC USES code book. In this method, the coder and decoder have a PREDETERMINED code book of stochastic (zero mean WHITE Gaussian) EXCITATION signals.

Correct answer is (d) Sub-band coding

Easy EXPLANATION: Sub band coding (SBC) is a method where the speech SIGNAL is subdivided into several frequency BANDS and each band is digitally encoded SEPARATELY. The audible frequency spectrum 20Hz-20 KHz is divided into frequency sub-bands using a bank of finite IMPULSE response (FIR) filter and output of each filter is sampled and encoded.

Right option is (c) Similarity, time

Best explanation: The AUTOCORRELATION function (ACF) gives a QUANTITATIVE measure of the closeness or similarity between samples of a speech signal as a function of their time SEPARATION. In every sample of speech, there is a large component that is easily PREDICTED from the VALUES of the previous samples.

Correct option is (d) Uniform probability distribution of SPEECH amplitude

Easy explanation: Speech waveforms have a NUMBER of useful properties that can be EXPLOITED when DESIGNING efficient coders. They are non uniform probability distribution of speech amplitude, non-zero autocorrelation between successive speech samples, the nonflat nature of the speech spectra and quasiperiodicity of voiced speech signals.

The correct option is (c) Regular pulse EXCITED long term prediction

The explanation is: The original speech CODER used in the pan European digital cellular STANDARD GSM GOES by a RATHER grandiose name of regular pulse excited long term prediction (RPE-LTP) codec. This codec has a bit rate of 13 kbps.

Correct ANSWER is (b) False

To explain I would say: The RPE-LTP codec combines the ADVANTAGES of the EARLIER French proposed RELP codec with those of the multipulse excited long term PREDICTION (MPE-LTP) codec proposed by Germany.

The correct answer is (a) Signal COMPRESSION techniques

Explanation: SPEECH CODERS are categorised on the basis of signal compression techniques. Speech coding is an art of compressing and then ENCODING speech signals.

Right answer is (b) False

To EXPLAIN: There is a nonflat characteristic in power spectral density of speech. It makes it possible to obtain significant compression by CODING speech in the FREQUENCY domain.

Correct OPTION is (b) False

For explanation I would SAY: Multipulse EXCITED LPC does not require pitch detection and the prediction residual is better APPROXIMATED by SEVERAL pulses per pitch period. This is the reason for better speech quality.

The CORRECT answer is (a) Weighted sum of PAST samples

To EXPLAIN: The linear predictive coder USES a weighted sum of p past samples. Using this technique, the current sample can be WRITTEN as linear sum of the immediately precoding samples.

Correct answer is (b) False

Best explanation: Channel VOCODERS are frequency domain vocoders. They DETERMINE the envelope of the speech signal for a number of frequency bands and then SAMPLE, encode and multiplex these samples with the ENCODED OUTPUTS of the other filters.

Right option is (b) Quantization ERROR increases

Easiest explanation: ADM REDUCES SLOPE error, at the EXPENSE of increasing quantizing error. This error can be reduced by using a low-pass filter.

Right option is (d) Vector sum excited LPC

The explanation: The US DIGITAL cellular SYSTEM, IS-136 USES a vector sum excited linear predictive coder (VSELP). This coder operates at a RAW data rate of 7950 bits/s and a total data rate of 13 kbps after channel coding.

The correct choice is (a) CONTINUOUS, DISCRETE

Easiest explanation: Quantization is a PROCESS of mapping a continuous RANGE of amplitude of a signal into a finite set of discrete amplitudes. Quantizers are the devices that REMOVE the irrelevancies in the signal.

The correct answer is (d) Increasing function of AMPLITUDES between these extremes

Best explanation: There is a non-uniform PROBABILITY distribution of speech amplitude. The PDF of a speech signal is in general characterized by a very high probability of non-zero amplitudes, a SIGNIFICANT probability of very high amplitudes, and a MONOTONICALLY decreasing function of amplitudes between these extremes.

The correct choice is (a) Waveform coders and source coders

The explanation: SPEECH coders can be CATEGORISED into waveform coders and source coders. Waveform coders can further be categorised into TIME domain and FREQUENCY domain. Source coders can be classified into linear predictive coders and VOCODERS.

Right option is (b) False

To EXPLAIN I WOULD SAY: The choice of speech coder depends on the cell size used. When the cell size is sufficiently small such that high spectral efficiency is achieved through frequency reuse, it may be SUFFICIENT to use a simple high rate speech codec.

Right answer is (c) ARITHMETIC MEAN, geometric mean

Best explanation: Spectral flatness measure is defined as ratio of arithmetic to geometric mean of the samples of the PSD taken at UNIFORM intervals in frequency. Spectral flatness measure is a qualitative measure of the theoretical maximum coding GAIN that can be obtained by exploiting the nonflat characteristics of speech spectra.

Correct answer is (a) Long AREA ratio transform

For explanation: Long area ratio (LAR) transform is GENERALLY USED to improve the coding of reflection coefficient. This non linear TRANSFORMATION reduces the sensitivity of reflection coefficients to quantization errors. LAR performs an inverse HYPERBOLIC tangent mapping of reflection coefficients.

Correct ANSWER is (b) GOOD QUALITY of SPEECH, high complexity

Best explanation: The MPE-LTP technique produces excellent speech quality at high complexity. It is not much AFFECTED by bit errors present in the channel.

Right CHOICE is (a) Good QUALITY of speech, low complexity

For explanation I would say: The advantage of baseband RELP codec is that it provides good quality speech at low complexity. The speech quality is sometimes LIMITED DUE to tonal noise introduced by the process of high frequency GENERATION.

Right OPTION is (a) Multiple access TECHNIQUE

For explanation I would say: The type of multiple access technique used is an important FACTOR in determining the spectral efficiency of the system. It STRONGLY influences the choice of SPEECH codec.

The CORRECT option is (b) Step size

To elaborate: An adaptive quantizer varies its step size in ACCORDANCE to the input speech SIGNAL POWER. Its characteristics shrink and EXPAND in time like an accordion.

The correct option is (B) 2^n

Easiest explanation: A quantizer that USES n bits can have M = 2n discrete amplitude levels. Amplitude quantization is an IMPORTANT step in any speech coding process, and it determines to a great extent the overall distortion as well as BIT rate necessary to represent each waveform.

The CORRECT option is (c) LPC Vocoders

The best I can explain: LPC vocoder requires that the transmitter extract pitch frequency information which is often very difficult. Moreover, the phase coherence between the harmonic components of the excitation PULSE TENDS to PRODUCE a buzzy twang in the synthesized speech.

The correct option is (b) False

Easy explanation: Linear predictive coders are computationally intensive. But, they are the most popular among the class of low bit vocoders. With LPC, it is possible to TRANSMIT GOOD QUALITY VOICE at 4.8 kbps and poorer quality voice at even lower RATES.

The correct answer is (c) Vector quantization

Easiest explanation: Vector (VQ) quantization is a DELAYED DECISION coding technique. It maps a GROUP of INPUT samples called a vector to a code book INDEX. A code book is set up consisting of a finite set of vectors covering the entire anticipated range of values.

The correct choice is (b) 10-15

Explanation: LPCs USES weighted SUM of past p samples to estimate the present samples. The number of past samples used by linear predictive CODERS ranges from 10 to 15.

Right option is (a) CELP

For explanation: The VSELP SPEECH coder is a variant of the CELP type vocoders. The code BOOKS in the VSELP encoder are ORGANISED with a predefined structure such that a brute-force search is avoided.

The CORRECT answer is (d) Discrete COSINE transform

Explanation: DCT (discrete cosine transform) is one of the most attractive and FREQUENTLY used transforms for speech coding. Fast algorithms developed for COMPUTING the DCT in a COMPUTATIONALLY efficient manner are used.

The CORRECT answer is (b) False

To ELABORATE: The lower the bit rate at which the CODER can deliver toll quality speech, the more speech channels can be compressed within a given bandwidth. Thus, manufacturers are continuously in SEARCH of speech coders that provide toll quality speech at lower bit rates.

Correct choice is (b) False

The best I can explain: It is FUNCTION of block transform CODING. However, a sub band coder divides the SPEECH signal into many smaller sub BANDS and encodes each sub band separately according to some perceptual criterion.

Correct choice is (b) False

For explanation I would say: SHANNON predicted that better performance can be ACHIEVED by coding many samples at a TIME INSTEAD of one SAMPLE at a time.

The correct OPTION is (B) False

For explanation I would SAY: Vocoders are much more COMPLEX than the waveform coders. They can achieve very high economy in transmission bit rate but are LESS robust.

Correct OPTION is (a) μ law

To explain I would SAY: COMPANDING technique known as μ law is used in US. In EUROPE, A law companding technique is used.

Correct option is (a) Bandlimited

To EXPLAIN: The most basic PROPERTY of speech WAVEFORMS that are exploited by all speech coders is that they are bandlimited. A finite bandwidth means that it can be time-discretized at a finite RATE and reconstructed complexity from its samples.

Correct choice is (a) Sampling, quantizing, encoding

Easiest explanation: SEQUENCE of operation in PCM is sampling, QUANTIZATION and encoding. Sampling and quantizing operations transform an ANALOGUE signal to a digital signal. The quantizing and encoding operations are usually performed in the same CIRCUIT at the transmitter, which is called an Analogue to Digital Converter (ADC). At the receiver end the decoding operation converts the pulse back into an analogue voltage in a Digital to Analogue Converter (DAC).

Right choice is (d) Highest speech quality, MODEST computational complexity, robust to channel ERRORS

The explanation: VSELP speech coder is designed to accomplish the three goals of highest speech quality, modest computational complexity and robustness to channel errors. The CODE books used by VSELP impart high speech quality and INCREASED robustness to channel errors.

The CORRECT option is (c) All pole

Easiest explanation: The linear predictive coding system models the vocal TRACT as an all pole linear filter. The excitation to this filter is either a pulse at the pitch frequency or random white noise depending on WHETHER the speech segment is VOICED or unvoiced.

The correct answer is (b) False

Explanation: The type of modulation employed has a considerable impact on the choice of SPEECH codec. USING bandwidth EFFICIENT modulation scheme can lower the bit RATE reduction REQUIREMENTS on the speech codec and vice versa.

Right choice is (a) Time

Easiest explanation: Linear predictive VOCODERS belong to the time domain class of vocoders. This class of vocoders attempts to extract the SIGNIFICANT features of the speech from the time WAVEFORM.