= ‘ S Gras char ee + 4 A. %@ xe ‘C. Ramesh Babu Durai . jay F a he BB Ee =Published by LAXMI PUBLICATIONS (P) LTD 22, Golden House, Daryaganj, i New Delhi-110002. Phe 011-23 26 23 68 ones | 011-23 26 23 70 : 011-23 25 25 72 Roses { 011-23 26 22 79 | Branches : 129/1, I1ird Main Road, IX Cross Chamrajpet, Bangalore (Phone : 080-26 61 1 61) 26, Damodaran Street, T. Nagar, Chennai (Phone : 044-24 34 47 26) St. Benedict’s Road, Cochin (Phone : 0484-239 70 04) Pan Bazar, Rani Bari, Guwahati (Phones : 0361-254 36 69, 251 38 81) 4-2-453, Ist Floor, Ramkote, Hyderabad (Phone : 040-2475 02 47) Adda Tanda Chowk, N.D. 365, Jalandhar City (Phone : 0181-222 12 72) 106/A, Ist Floor, S.N. Banerjee Road, Kolkata (Phones : 033-22 27 37 73, 22 2452 47) 18, Madan Mohan Malviya Marg, Lucknow (Phone : 0522-220 95 78) 128A, Block 3, First Floor, Noorani Building, L.J. Road, Mumbai (Phone : 022424 46 39 98) Radha Govind Street, Tharpagna, Ranchi (Phone : 0651-230 77 64) | EMAIL : colaxmi@hotmail.com | WEBSITE : www.laxmipublications.com ©All Rights Reserved with the Author and the Publishers. | IEDS-0652-125-DIGITAL SIGNAL PROCESSING First Edition Price : Rs. 125.00 Only. C—-9743/04/12 Laser Typesetied at : Goswami Printers, Delhi-110053 Printed at : Ajit Prinfers, Delhi-110053Contents Chapters Pages DL Titra isis eect encscansnainanmnnsesinaiessnianrn 12S 1.1 Classification of Signals . 1.2 Multi Channel 1.3 Molti Dimensional Signals 1.4 Continuous-time Versus Discrete-time Signals 1.5 Frequeney Concept is Continuous Time and Discrete Time Signals... 1 Continuous-time Sinusoidal signals 2 Discrete-time Sinusoidal Signals 3 Harmonically Related Complex Exponentials .... 1.6 Energy and Power Signals (Continuous time-instants) L7 Singularity Functions Ea i - 1.2.1 _Unit-Impulse Function ....cccscsssssssscsssessssssnsssesesnnssssaceneccessnanes 1.7.2 Unit-Step Function 17.3 Unit-Ramp Function. 3 LLZ.4 Unit-Pulse Function vucsssscssssesssssssesessesssasssssnsnsenissscsssssnissssssssssnsssssesseee LT 1.8 Energy Signals and Power Signals (Discrete-time instants) 1.9 Signal Processing. seu 1.10 Analog Versus Digital Signal Processing Review Question: Exercii eo 20 loo be 2.__Applications of Digital Signal Processing. 2.1 Introduction ......... 2.2 Application to Speech Processing ...... 2.2.1 Vocal Mechanism .. 2.2.2. Speech Technology 2.2.3 Parameters of Speech ... 2.2.4 Speech Analysis .. .2.5 Speech Coding: 2.3 Application to Image Processing se 2.3.1 Image Formation and Recording .. 2.3.2 Image Sampling and Quantization2.3.4 Tmage Restoration 2.3.5 Image Enhancement Review Questions .. 3. Discrete Time S: 3.1 Discrete-time Signals and Systems .... 1.1 Definition 3.1.2 Representations 5 3.13 Some Blementary Sequence 4! 3.14 Representation of Arbitrary Sequence wn 43 3.2 Classification of Discrete-time Signal ... 44 3.3. Samplin; 2 46 3.4 Real and Compler Sequenc 46 3.5 Finite and Infinite Sequence 47 3.6 Types of Infinite-length Sequence . oxen 8.7 Operations on Sequences 48 3.8 Sampling Rate Alteration ... 50 $.9 Classification Based on Symmetry Problem... 50 3.9.1 Periodic Conjugato-symmetric Part and Periodic Conia Anti-symmetric Part 3.10 Sampling Process... : 3.11 Classification of Discrete-time Sytem 8.12 Time-domain Characterization stents 3.12.1 Representation of a Discrete-time Signal in Terms of 2.122 Discrete-time Unit Impulse Res Representation of LTI System ... 8.12 The Convolution Process ossssusscnssssnnssnssssnessnsesins 3.14 Properties of Linear Time-invariant 83 3.15 Causality and Stability Condition for LI Discrete-time Systerp ... 3.16 Classification of LTI System .. Systems Described by Difference Equation . Recursive and Non-recursive Discrete-time S Linear Constant Co-cfficient Difference Equation 3.19.1 IIR and FIR System... Solution of Linear Constant Co-efficient Equation 9.20.1 The Homogeneous Solution of a Difference Equation 3.20.2 The Particular Solution of the Difference Equation 3.20.3 The Total Solution the Difference Equation . . ‘The Impulse Response of a LTI Recursive System pulse BERBBSERERNBEB BEEBEChapters 3.22 Impulse Response Review Questions Exercises. 4,__Frequency Donain Characterization or Discrete-Time System. 4.1 Fourier Transform of discrete-time Signals .. 4.1.1 Fourier Series for Discrete-time Periodic Signal 4.12 Condition for convergence of Fourier Transform. 4.2. Frequency response of Discrete-time Systems 4.3 Properties of Frequency Response... 4.4 Polar form of Frequency Response . 4.5 Frequency Response of First order System 4.6 Properties of Frequenc rRleenee iis 47 2-Transform 4.7.1 Definition of Z-transti 4.7.2 Region of Convergence .. 4.73 Properties .... . pessoal uni Gina Sided 2 tyiaaforma Patra 108 4.8.1 The Inverse Z-transform Using Contour Integration ..... 4.82 The Inverse Z-transform by Power Series Expansion or Via Long Division . 4.8.3 The Inverse Z-transform by Partial Fraction Expansion .. 4.9 Solution of Difference Equation Using Z-Transform ... Review Questions Exercisos 5, Frequency aii gaie of Signals . 5.1 Frequency Analysis of Continuous-time (Analog) Signals 5.3 Symmetry Conditions for Periodic Signals 5.4 Exponential Fourier Series 5.4.1 Existence of Fourier Series 5.5 Fourier Spectrum 5.8 Properties of Continuous-time Fourier Series .. 5.7 Continuotis-Time Fourier Transform 5.8 Fourier Transform of a Periodic Signal .. 5.9 Properties of Continuous Time Fourier Transform 6.101 Frequency Domain Representation of Discrate Time Signal and System 5.10.1 Frequency Analysis of Discrete Time SignalsChapters 0.2 Fourier Series for Discrete Time Periodic Signals... 5.10.3 Expression for the Values of the Co-efficient a,.. * [ime Fourier Transform ... ee) 5.11.1 Inverse Discrete Time Fourier Transform ... 5.11.2 Condition for Convergence of Fourier Transform 5.11.3 Energy Density Spectrum . 5.11.4 Properties of Discrete-Time Fourier Transform Review Questions ..... 6.3 Properties of the DFT... 6.4 Linear Convolution 6.5 Circular Convolution ... rm i hued 181 6.5.1 Methods of Performing Circular Convolution GG Sectioned ConvolUtions si scscssetsscenennsnnssusssusssss 6.6.1 Overlap Add Method .. 6.6.2 Overlap Save Method .. 6.7 Computation of the DFT of Real Sequences, 6.7 N-point DFTs of Two Real Sequences using a Single N-poit DFT 6.7.2 2N-point DFT of a Real Sequence using a Single N-point DFT . 6.8 Fast Fourier Transforms Algorithms 1... sta: Biel SuiniiOn acess caecacce 68.2 Radix of FFT Algorithms. 6.9 Decimation-in-time FFT Algorithms ; . 6.10 The 8-point DFT using Radix-2 DIT FFT sess 6.10.1 Flow Graph for 8-point DIT Radix-2 PFT... . 6.11 Decimation in Frequency (DIF) Radix-2 FFT oj... 3 6.11.1 Tho 8.point DFT using Radix.2 DIF FFT i 6.12 Comparison of DIT and DIF .. 1 Introduction Sampling Process 7. - 7.2.1 Analysis of Sampling <— in Frequency Domain... 46Chapters 73 14 75 16 a 18 19 7.10 Zl 7.12 113 Tis TAS 7316 717 8. Digital Filter Structures... BL 82 82 84 8.5 Structure For IIR System .. Pages Sampling Theorem 250 Anti Aliasing Filter. 250 Signal Reconstruction .. .250 Zero-order Hold .... 253 7.6.1 Transfer Function of Zero Order Hold 253 Sampling of Band Pass Signals 257 Frequency Selective Filters and 7.8.1 Filter Specifications... Analog Lowpass Filter Design - ‘Analog Lowpass Butterworth Filter . Analog Lowpass Chebyshev Filters ...ocussuss 7.11.1 Type-I Chebyshev Approximation 7.112 Pole Locations for Chebyshev Filter .11.3 Chebyshev Type-Il Filter Analog Frequency Transformation .... " Design Procedure for Analog Butterworth Lincs Filter... Design Procedure for Analog Chebyshev Lowpass Filter .. Sample and Hold Circuit...... Analog-to-Digital Convertor 7.16.1 Flash A/D Converters ... 7162 Serial-Parallel A/D Converter 7.16.3. Successive-approximation A/D Converter 7.16.4 Counting A/D Converter 7.16.5 Oversampling Sigma-Delta A/D Converter .. Digital-to-Analog Converter 2.17.1 Weighted-Resistor D/A Converter .... 2.17.2 Resistor Ladder D/A Converter 7.17.3 Oversampling Signal-delta D/A Converter Review Questions Brercises ter Specifications. Introduction System Describing Equations ... Recursive and Non-recursive Structures Block Diagram Representations cu sci 8.4.1 First Order System Block Blazon! Bugruaall iii ‘ 8.5.1 Direct Form Structures 8.5.2 Cascade Form Structure .. Parallel Form StructureChapters 8.6 Structures For FIR Systems 51 Direct Form FIR Structure .... 8.6.2 Cascade Form FIR Structure .. 8.6.3 Linear Phase FIR Structure Review Questions . 9.2 Selection of the Filter Type... nee 9.2. R Filter Design by Impulse Invariance .. 02 Bilinear Transform Matho: ; 9.3.1 Development of Transformation ..... 9.3.2 Characteristics of Bilinear Transformation .. 9.4 Warping Effect 95 Pre-Warping Review Questions. sotnssies - . vasananee BRT Exercises ... Examination Question Papers Index.Chapters : 1. Introduction 2. Applications of Digital Signal Processing DIGITAL SIGNAL PROCESSINGIntroduction Characterization and Classification of Signals Signal A‘signal’ is defined as any physical quantity that varies with time, space and any other independent variable or variables. More precisely a signal is a function of a set of independent variables. The signal itself carries some kind of information available for observation. Processing By ‘processing’ we mean operating in some fashion on signal to extract some useful information. Digital ‘The word ‘digital’ shall mean that the processing is done with a digital computer or special purpose digital hardware. Digital Signal Processing Digital signal processing is concerned with the representation of signals by sequence of numbers or symbols and the processing of these sequence. ‘The purpose of such processing may be to estimate characteristic parameters or trans- form a signal into form which is in some sense more desirable. Application Bio-medical engineering, acoustics, radar, speech communication, data communication, image processing, nuclear science and many others. 11 CLASSIFICATION OF SIGNALS ‘There are five methods of classifying signals based on different features : (a) Based on independent variable. () Depending upon the number of independent variable. (©) Depending upon the certainity by which the signal can be uniquely described. (d) Based on repetition nature. (e) Based on reflection.Digital Signal! Processing (a) Based on independent variable. Independent variables can be ¢pntinuous or dis- crete. | 1. Continuous Time Signal. It is also referred as analog signal i.e., thp signal is repre- sented continuously in time. In simple words, a signal x(t) is said t# be a continuous time signal if it is defined for all time. 2. Discrete Time Signal. Signals are represented as sequence at diserefe time intervals. ‘Thus, the independent variable has discrete values only. xt). x(n) 4 y ° ‘ r (a) Continuous time signal (6) Discrete time signal Fig. 1.1 e.g. Speech signal is an example of analog signal. A discrete time signal which diserete-valued represented by a finite number of digits is referred to as a “digital signal”. e.g. Digitized music signal stored in CD-ROM disk. (6) Depending upon the number of independent variable. (i) 1-D Signals. It is a function of a single independent variable. e.g. (a) speech signal-independent variable is time. (b) music signal (ii) 2-D Signal. It is a function of two independent variables. e.g. Photographie image signal-two independent variables aye the two spatial vafiables. | Each frame of a black and white video signal is a 2D-image sighal that is a func- tion of two discrete spatial variable, with each frame occurring squentially at discrete instants of time. (iii) M-D Signal. It is a function of ‘M’ independent variable in tin eg. Video signal. The black and white video signal can be considered an example of a 39 signal where the three independent variables are two spatial variables and time. A colour video signal is a three-channel signal composed of three 3-D kignals represent- ing the three primary colours : red, green and blue (RGB). For transmission purpose, the RGB television signal is transformed isto another type of 3-channel signal is composed of luminance component and two chrominange components. (c) Depending upon the certainity by which the signal can be unjquely described as (i) Deterministic Signal. Asignal that can be uniquely determine by a well-defined process such as a mathematical expression or rule, or table Jook-up is called a deterministic signal.Introduction e.g. (a) A sinusoidal signal can be represented as, x) v(t) = V,, sin ot fort 20. (6) A square signal can be defined as a xf)=A for O0 forn<0 forn=0 forn>0. Thus, Even [x(n)] = Flin) + xm A138) Oda (xin)} = Ftxin) — 2m 1.4)Digital]Signal Processing Properties of even and odd signal : 1. The sum of two even signals are even signal. 2. The sum of two odd signals are odd. 4. The product of two even signal is even. 5, The product of two odd signal is even. 6. The product of even signal and an odd signal is odd. 1.2 MULTI CHANNEL signal can be generated by a single source or by multiple sourcesjor multiple sensors. In the former case, it is a (single) scalar signal and in the later case it is 4 vector signal, often called a multichannel signal. These type of signals can be represented in vector form as, Fxy(e) aft) = | 2, (6) a(t) Equation represents a 3-channel signal. e.g. In clectrocardiography [ECG] for example 3-lead and 12-leadjelectrocardiographs are often used in practice, which result in 3-channel and 12-channel sigafls. (1.5) 1.3 MULTI DIMENSIONAL SIGNALS Ifa signal is a function of a single independent variable, then it is Falled as one-dimen- sional signal, Similarly, if signal is a function of N-independent variables, it is called as N- dimensional signal. eg. « Picture signal is a two dimensional signal, since the intensity ‘wo independent variables x and y. ‘¢ Black and white television picture is an example of 3-dimensfonal signal because brightness I(x, y, £) is a function of three independent variablesfr, y and ¢ (time). @ It isalso possible to have multichannel and multidimensional sighals simultaneously. For example, a colour TV picture is described by three intengity functions of form 1,4, y, t) fredl, I, (x,y, #) [green], and I, (x, y, £) {blue}. Hence colour TV picture is a three dimensional and three channelfignal, which can be represented by the vector. ,y) is a function of L(x, 9,0) T(x, 90) I,(x, ¥, 8) lay, 0) (1.6) 1.4 CONTINUOUS-TIME VERSUS DISCRETE-TIME SIGNA (1) Signals can be further classified into different categories depehding on the charac- teristics of the time (independent) variables and the values they take. Continuous * Continuous-time signals or analog signals are defined for evel they take on values in the continuous interval (a, b). value of time andIntroduction [ 9 ] where — acan be bcan be +. Mathematically, these signals can be described by functions of a continuous variable. www Ww e.g. Speech signals x,(t) = cos nt xft)=eltl, -n0 xn) o| + 2 an Continuous time signal x(t)=e for n<0. x) (i) Random Signal. A signal that is generated in a random fashion 4nd cannot be predicted ahead of time is called a “random signal”. eg. Speech signal, ECG signal, EEG signals. 1.5 FREQUENCY CONCEPT IS CONTINUOUS TIME AND DISCH SIGNALS | We know that the frequency is closely related to a periodic motion which iq described by ETE-TIME nature of frequency accordingly. 1.5.1 Continuous Time Sinusoidal Signals A continuous time signal is mathematically described as, x, ()=Acos(Qt+8); -a 0) if and only if x(n +N) =x(n) =A1.9) The epallaet value ofN is called the fundamental period. = cos [2n f,n + 8} ‘The above relation is true if and only if, there exists an integer & On f.N = Onk. k N Therefore, the discrete-time sinusoids are periodic only if its pressed as rational number (ratio of two integers). | To determine the fundamental period N of a periodic sinusoid, } and N is eqn. 1.10 should be relatively prime. Then the fundamental period of sinusoid is ¢yual to N. For exam- ple, de, fo= (1.10) juericy fy can be ex-introduction 13 | 21 ie, Aa then fundamental period N, is 40 and if 20_1 f= 4072 then fundamental period N, is 2. We observe that a small change in frequency may result in a large change in the period. (2) Discrete-time sinusoids whose frequencies are separated by an integer multiple of 2x. are identical. Proof : Consider a sinusoid cos (wy n + 6). If the frequencies are separated by 2n, then, cos [(wp + 2x) n +6] =cos (cw, n + 2an +O) = cos {ay n + 6]. Therefore, all the sinusoid signals, x(n) = Acos (w,n +8); k=0,1,2... where, w, = w, + 2nk, are identical (distinguishable). Conclusion. The discrete-time sinusoids with frequencies | | or |f | > V/2, are identical to the sequence obtained from the sinusoid with frequency | @ | ot a in on 4x0) Gen 1 + oe 1 15.3 Harmonically Related Complex Exponentials Sinusoidal signals and complex exponentials play a major role inthe lysis of signals and systems. In some cases we deal with sets of harmonically related compl exponentials (or that are multiples of a single positive frequency. Continuous-time exponentials. The basic signals for continuous- related exponentials are, St) = ett = gi p= 0,2 122 dip we note that for each value of k, S,(¢) is periodic with fundamental period 4 = + - or funda- mental frequency RF. | Since a signal that is periodic with period T /k is also periodic with iod|k (T/A = T, for any positive integor k, we see that all of the S,(t) have a common period fT, From the basic signals in eqn. (1.11), we can construct a linear combinption of harmoni- cally related complex exponentials of the form,Introduction [ 5 ] =, = a CS, = x Cyeitnt (12) bane where, C,,# =0,#,1#2......0re arbitrary canmplex constants, Nv ‘The signal z,(¢) is periodic with fundamental period T, = 1/F, and its representation in terms of eqn. (1.12) is called the Fourier series expansion for x,(t). ' Discrete-Time exponentials Since a diserete-time complex exponential is periodic if its relative frequency is a ra- tional number, we choose f, = I/N and we define the sets of harmonically related complex exponentials by Sy(n) = ef?ekfon jk=0,81,42,23.... +118) In contrast to the continuous-time case, we note that, Sycayltt) = oP NN = of2MS,(n) = S,(n) ‘This means that, consistent with eqn. (1.9) [ie.,2(N +n) =x(n)] there are only N distinct periodic complex exponential is the set described by eqn. (1.13). Furthermore, all members of the set have a common period of N samples. Clearly, we can choose any consecutive N complex exponentials say from k =n, tok =n +N-1 to form a harmonically related set with fundamental frequency fy = 1/N. For our convenience, we choose the set that corresponds to n, = 0, that is the set, S, (2) = oN k= 0,1,2.....N—1. As in the case of continuous-time Sgnals it is obvious that the linear combination, x(n) = s C,Sy(n)= > Cyel2nnN | m a results in a periodic signal with fundamental period ‘N’. The sequence of S,(n) is called the k** harmonic of x(n). 1.6 ENERGY AND POWER SIGNALS (CONTINUOUS TIME-INSTANTS) Signals can also be classified as those having finite energy or finite average power. However, there are some signals which can neither be classified as energy signals nor power signals. Consider, a voltage source u(i), across a unit resistance R, conducting, a current i(t). ‘The instantaneous power dissipated by the resistor is v(t) P(t) = v(2) it) = Since R= 1, we have, Plt) = v%(t) = i? (#). The total energy and the average power are defined as the limits. =P(e)R. 7 B= jim |? Wade joules (14) and P= din kf ? (de watts (115) The total energy and the average power normalised to unit resistance of any arbitrary signal x(¢) can be defined as,B= lim [" |x(e)/? at Joul (4.16) = Jim [lef de Joules. wd. Continuous 0 faoa-{ ,t<0? Since, the area of the impulse function is all concen- Fig. 1.9 (). Discrete ti trated at ¢ = OP me signal.Introduction Gy for any value oft < 0, the integral becomes zero and for ut) ‘ t> of BU) dt =1. The integral of the impulse function is also a sin- 1 gularity function and called the unit-step function and is represented as, 0, t<0 : OPa a w= fh iS0 Fig. 1.10 (a). Continuous time ‘The value att = 0 is taken to be finite and in most unit step signal. cases it is unspecified. The discrete-time unit-step sig- uta) nal is defined as (oO , n 2So- 1.7.3 Unit-Ramp Function Fig. 1.10 (6). Discrete time unit The unit-ramp function, r(¢) can be obtained by in- step signal. «a tegrating the unit-impulse function twice or integrating the unit-step function once, ie, rt) = f _fs@acaa, i fi u@ac. no= fide (Oo, <0 Thatis, ro={ Es A ramp signal starts at ¢ = 0 and increases linearly with time ‘t. In discrete-time domain, the unit-ramp sig- nal is defined as, Fig. 1.11 (a). Continuous time ramp signal. "(n) ee ic + r0° Fig. 1.11 (). Discrete time 17.4 Unit-Pulse Function ramp signal. An unit-pulse function, x(¢), is obtained from unit-step signal as shown below. x(t) = ule + 1/2)— u(e— 172)Fig. 1.12. Unit pulse signal ‘The signal u(t + 1/2) and u(t — 1/2) are the unit-stop signals shifted by f/2 units in the time axis towards the left and right respectively. is made up of straight line segments can be represented in terms of step and Advantage. The advantage of the singularity function is that any arbitfary signal that functions, Properties of &(t) @ [soa =1 (2) [xe) 8) dt = x(0) | Proof for (2): [” x(t) lim 5,() dt 5 )= lim 3,() LO oe 2, re a . 1 zt. = jim fi x) Py dt = dim {Pr} lim Pre) = lim af x(t)dt =x(0). | According to pulse funtion property, Py(t)= 1 @) [) #8t- 4) dt = xe) & f2@se-ar <2) 1 (5) Sat) = + &(e) (6) x(0) &¢ - £4) = att) la} (7) x(ty) B(¢ - ty) =x(,) ©) [* x@B"G- 4) dt = CIP). mn Proof for (8): & Lele) Be — tg)] = alt) 8 (- 29) + 4 (H Bt ty) = x{t)B(t — tg) + (ty) Bt ty), ty < ty 0 Evaluate the real and imaginary components of x(¢),aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Applications of Digital Signal Processing 2.1 INTRODUCTION Because of the availability of high resolution spectral analysis, DSP has various appli- cation areas, which requires high speed processors to implement the FFT algorithm. It is also popular due to availability of custom made DSP chip which is highly reliable. Speech process- ing, Audio processing, Radar signal processing and Image processing would be discussed in this chapter, 2.2 APPLICATION TO SPEECH PROCESSING ‘The signals of speech are one dimensional. DSP is applied to a wide range of problem in speech such as channel vocoders, spectrum analysis etc. Problems in speech processing can generally be divided into three classes, first is the speech analysis. The speech analysis is performed to extract some desirable information of speech. This system starts with analysis of speech waveform and the desired result is used for speech recognization and speaker indentification. Second type of problem is speech synthesis. Init, input is in written text form and the outputis a speech signal. For example, an automatic reading machine for which the input is written text and the output is speech. Finally the third type is speech compression which involves speech analysis followed by speech synthesis. If the speech is transmitted by simply sampling and digitizing, the data rate required isin the order 0f 90,000 bits per second of speech. Through the use of appropriate coding this can be reduced by factor of 50, depending on the type of system used. 2.2.1 Vocal Mechanism Production of speech. The two important part responsible for human speech are (a) vocal cord and (b) vocal tract. (a) Vocal cord. It has two bands of tough, elastic tissue, which is located at the opening of the larynx. It vibrates when the air from the lungs passes between them producing sound waves which are emitted from the lips and to some extent from the nose ; these are sound waves heard as speech. (8) Vocal tract. It includes larynx, the pharnx and the nasal cavity. 29Kinds of Sounds \ @ Voiced sound (Gi) Unvoiced (fricative) sound. || Voiced sounds are produced by quasi-periodic pulses of air exciting the vocal tract. Unvoiced sounds are produced at some point along the vocal tract, usually towards the mouth. ‘There are some important speech technology areas. viz,, speech coding, speech enhance- 2.2.2 Speech Technology (a) Speech coding. “Speech Coding” is the process of capturing the speech of a person and processing it to transmit over a communication channel. ‘The application of “speech coding” is in the area of telephony, narroy-band cellular radio, military communication ete. (0) Speech enhancement. This is the process of minimizing the derogatory effects of noise on the performance of speech communication, source coding etc. The application of ‘speech enhancement’ is in the areas where the perfornance of equip- ment is improved in noisy atmosphere like factories ete. (©) Speech analysis and synthesis. Analysing speech is done by studying its spee- trum and extracting time-varying parameters from the signal for the productipn of speech. Synthesizing speech lies in creating speech like waveforms from textual words or sym- bols, using a model for speech production and time-varying parameters. ‘The application of this are in voice alarms, reading machines for the Rumb or blind, data-base enquiry services etc. (d) Speech recognition. The process of deriving the meaning from p speech input whereby a request can be made for information or service from a machinery by conversing with it. Application of“speech recognition” could be Banking from distant locatipn, information retrieval systems ete. | (©) Speaker recognition. It means to recognize a particular person's identity with the sample speech dipping. 2.2.3 Parameters of Speech (i) Pitch : Corresponds to frequency of sound (in Hz). (ii) Loudness : This relates to intensity of sound (in 4B) (iii) Quality : This relates to harmonic constant of sound (in timbre). ‘Phonemes’ are the smallest unit of sound that are recognized by contrast with their environment, these are forming the basic units of speech. ‘Dipones’ are sourgls that stretch from the middle of one phoneme to the centre of the next, there by spanning the transition region. 2.2.4 Speech Analysis ‘The most common methods of speech analysis are as follows (a) Short-time fourier analysis (©) Linear prediction. (c) Homonorphic filtering. Let us discuss about these three methods of speech analysis.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Applications of Digital Signal Processing additive recombination of the set of sub-band signals, the original speech signal can be generated. Each band is separately quantized and coded using pulse code modulation and transmitted. The schematic is shown in Fig. 2.5. 2.3 APPLICATION TO IMAGE PROCESSING Any function which bears two-dimensional information is called an image. Image can be represented by an array of real or complex (real and imaginary) numbers with finite number of bits with respect to speech signal (which are one-dimensional signals), image signals are two dimensional. Image can be divided into picture elements or pixels (smallest element of image). Manipulation of two-dimensional signal with the help of digital computer is called “{mage Processing”. Its purpose is to improve the visual appearance of Image. A Digital Image is digitalization of picture. Normally two-dimensional imaye nas reso- lution 128 x 128, 256 x 256, 512 x 512. So image can be processed using two-dimensional signal processing. The image processing including the following steps : (a) Image Formation and Recording. (6) Image Sampling and Quantization. (c) Image Compression. (d) Image Restoration. (e) Image Enhancement. All the operations are possible on advanced software artificial intelligence and high- tech digital computers, Let us discuss all the operations one by one. 2.3.1 Image Formation and Recording ‘The two-dimensional signal of image can be expressed by image function as By) L £ lz ~ xy, ¥- yy) Fey, yy) dey dy --(2.15) Eqn. (2.15) governs a 2D linear time invariant system. Here system impulse response function M(x ~x,y -y,) is commonly referred as point-spread function which is usually associ- ated with optical image. The function flx, y) is the accumulation of energy from the objects radiant energy distribution. ‘Two major technologies are used for image sensing and recording, which are photo- chemical recording and photo-electronic recording. Both of the technologies are exemplified by readily available products which are photo-graphic films and television respectively. (Here “television” is used in generic sense not commercial broad casting television). 2.3.2 Image Sampling and Quantization After formation and recording of an image, it is sampled and quantized for the suitabil- ity of digital processing. In a system project, a spot of light with intensity I, incident on a film and intensity I, reflected from the film and collected by photo-multiplier. The transmittance is defined by 2.16)ax.y= f. f. Agile 1,9 — I) 8 (Xr, Ys) dey dy Here h, is the intensity profile of the spot of light projected an film. f is the image on film and finally g, is actual sampled image. The sample matrixg,(# A x, ! Ay) ps the sampled or digital image. (217) 2.3.3 Image Compression In a digital image 105 to 108 data are there. The processing of thest higher value of image data is a very stupendous task. But a digital image has large numbpr of redundancy which can be reduced by image compression. So we can say that image comprdssion isa science of efficiently coding a digital image to reduce the number of bits, which requfred to represent it. Uncompressed image consumes memory space in a large amount so t intreases com- plexity in computational and need a very large transmission bandwidth. A chmpressed image reduces the redundancy in image. There are mainly three types of redunflancy which are discussed one by one. ‘The first type redundancy is Spatial Redundancy which arises due fp correlation be- tween neighbouring pixels. Second type is Spectral Redundancy which is corgelation between various colour plans. And finally is Temporal Redundancy is the correlation Hetween different frames in an image sequence. In an Image Compression System, the original continuous time image signalis fed to A/D (Analog to Digital) converter, which converts it into digital signal. Now a seria to parallel (S/P) converter decomposed signal into parallel channels which fed to a quantigpr. The S/P con- verter is linear transformer or filter banks are used. This quantized output isjeoded by the use of lossless coding device, whose output is compressed digital image signal] A simple block diagram of image compression system is shown in Fig. 2.6. coe Decomposed Quantized signal signal into eutput parallel channels Input (orginal continuous ime image signal) AD sr *| Converter *) Converter >| Quantizer >| Aigital image signal) Fig. 2.6. Image compression system, Applications of digital signal processing image compression system ‘There are mainly three types of compression technique based on the dancy detection : method of redun-Applications of Digital Signal Processing [7] (@ Direct data compression method (6) Transformation method (c) Parametric extraction method. 2.3.4 Image Restoration ‘The process of image restoration is used for correcting imaging effect to recover an original signal. This type of effect (imaging effect) is due to variety of intermixing factors, which are defocusing imaging camera, relative motion between object and camora, noise in sensors ete., All types of imaging effects deteriorate image quality. ‘The process of image restoration is to attempt a image which should be sharp, clean and free from the degradation. The restoration process is also called Image Deblurring. The proc- ess of image formation and recording can be modelled as. sane aff frac =n yyy fly x) dey on] +nlxy) of 2.18) Hereg(x, y) is the actual image, R is the response characteristic of the recording process and a(x, y) is additive noise source. In the restoration of digital image following equation can be expressed in discrete form : . N-1N-1 aaa= Y, Y fi Mh(p-ig-) (2.19) ia fo A large set of simultaneous linear equations can be solved by DSP techniques such as linear filters and FFT algorithms which are computationally efficient tools for solving these 2.3.5 Image Enhancement This technique improves the appearance of image for human perception by choosing some image features like edges or contrast etc. Its main application is in biomedical engineer- ing field for computer aided mammographies studies. In image enhancoment spatial filtering is mainly used whose operation is done on im- age to reduce noise contamination of the image signal. Image enhancement is composed of a variety of methods whose suitability depends upon the goals at hand when enhancement is originally applied REVIEW QUESTIONS Give the areas in which signal processing find its application. Explain the various stages in voice processing. How is a speech signal generated ? Give the model of speech production system ? What is the need for short time spectral analysis ? What is a vocoder ? Explain with a block diagram ? Describe how targets can be detected using radar. Give an expression for the following parameters related to radar (a) beam width, and (6) maximum unambiguous range. PRPS PS9. 10. 11, 12. 13, 14, 15. 16. Explain with the block diagram the modern radar system. Give the various image processing applications. Give the various coding techniques for images. ‘What is the need for image compression ? Give the block diagram of basic restoration process. What is sub-band coding ? Explain the process of digital FM stereo signal generation, Explain how privacy can be acheieved in telephone communications. jignal_ ProcessingChapters : 3. Discrete Time Systems 4. Frequency Domain Characterization of Discrete-time Systems DIGITAL SIGNAL PROCESSINGDiscrete Time Systems 3.1 DISCRETE-TIME SIGNALS AND SYSTEMS 3.1.1 Definition 1. A discrete-time signal is a sequence, that is a function defined on the positive and negative integers. 2. A discrete-time system is a mapping from the set of acceptable discrete-time signals called the input set, to a set of discrete-time signals called output set. 3. A discrete-time signal whose values are from a finite set is called a digital signal. 4. A digital system is a mapping which assigns a digital output signal to every accept- able digital input signal. 3.1.2 Representations 1.Graphical. In digital signal processing, signals are represented as sequence of num- bers called samples. A sampled value of typical discrete-time signal or sequence is denoted by x(n) which is a function of independent variable that is an integer. It is graphically repre- sented in Fig. 3.1. (0) =n I | 3-2-1012 3 | | | n x(6) Fig. 3.1. Graphical representation. “1jignal Processing It is important to note that x(n) is defined only for integer values of nland undefined for non-integer values of n. Inthe signal we have assumed that a discrete-time sequence is defind! forlevery integet value ofn for -e x. = BD (3.8) ~A3.9)Sol. (a) To find y(n) = X(n + 1) In this case the system ‘advances’ the input one sample into the 0) = x(1). The response of this system to the given input is, yn) - 0, 3,2, 1, 0, 1,2, 3, 0 tT (b) To find y(n) = V3 [x(a + 1) +.x(n) + a(n - UI). ‘The output of the system at any time is the mean value of the presd past and the immediate future sample. For example, the output at time n = 0 is, 0) = 3 fx 1) + (0) + x(1)] =into+n KO; Repeating this computation for every value of n, we obtain, y(n) = 0, 1, 5/3, 2, 1, 2/3, 1, 2, 6/3, 1, 0. T yn= LY xe) ’ This system is basically an accumulator that computes the running input values upto present time. The response of the system to the given inp y(n) = {...... 0,3, 5, 6, 6, 7, 9, 12, 12, .......). 3.3 SAMPLING In some applications a discrete-time sequence x(n) is generated by pling a continuous time signal z,(¢) at uniform time intervals. a(n) =x,0)|) yp =%, (nT) 32 = -2,-1,0,1,2 ‘The spacing T between two consecutive samples in eqn. (3.10) interval or sampling period. ‘The reciprocal of the sampling interval T, denoted as F,, is called the sa priodically sam- +--(8.10) callbd the sampling pling frequency. Fp 3 -(eycle/see or Hertz). 3.4 REAL AND COMPLEX SEQUENCE Real sequence. If X(n) is real for all values of n, then (x(n)) is a real sequence. Complex sequence. If the n‘* sample value is complex for one or m@re values of n, then it is complex sequence. It can be defined as,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems [es] The device implementing the delay operation in) ‘ waio) by one sample is called a “Unit delay” and its § —————) =" schematic representation is shown in Fig. 3.8 Fig, 3.8. Unit delay. w,(n) = x[n - 1] ‘The schematic representationof the unit advance —(") 2 (9) operation is shown in Fig, 3.9 Fig. 3.9. Advance operation. ela) = xin + 1) (v) Time-reversed or Folding. The time reversal operation, also called the folding operation, is another useful scheme to develop a new sequence. we(n) = x(-n) --(B.20) which is the time-reversed version of the sequence a(n). (vi) Pick-off node, Itis used toprovide multiple copies ofa. ai sequence. Problem 3. Consider the following two sequence of length 5 defined for 0)]. We observe that, ut[(- 0) utl(- u* (-2)] =u%(2)=4+j2 u*((- 3),) =u*(1) =-2-J3.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems 53] where, 43.0) It is the normalised angular frequency of the discrete-time signal x(n). Units The unit of the normalised digital angular frequency w, is radians per sample. While, the unit of the normalised analog angular frequency 9, is radians per sample and the unit analog frequency fy is hertz is the unit of the sampling period T is in seconds. Problem 8. Consider the three sequence generated by uniformly sampling the three co- sine functions of frequencies 3Hz, 7Hz and 13Hz respectively : g,(t) = cos( 6m), g,(t) = cos (1471), and g,(t) = cos (2621) with sampling rate of 10 Hz. i., with T = 0.1 sec. Find the derived sequence or discrete sequence. Sol. a(t) = 008 (Qy 1); g(r) = cos (2 nT) (7) = cos (6 mn x T) (7) = cos (0.6 mn) Similarly, y(n) = cos (1.460) and By(r) = cos (2.6 nn). Problem 9. Determine the discrete time signal v(n) obtained by uniformly sampling at sampling rate of 200 Hz., a continuous time signal v,(t) composed of a weighted sum of five sinusoidal of frequencies 30 Hz, 150 Hz, 170 Hz, 250 Hz and 330 Hz as given below : v, (t) = 6 cos (6Ont) + 3 sin(300nt) + 2 cos (340nt) + 4 cos (500nt) + 10 sin (660nt). Sol. To find the sampling period (T) : 1 1 T= 5 = app = 0.008 sec. The generated discrete-time signal u(n) is given by, v(n) = 6 cos (0.3nn) + 3 sin (1.5nn) +2 cos (1.7nn) + 4.c0s (2.5nn) + 10 sin (3.3nn). =6 cos (0.3nn) + 3 sin [(2n —0.5n)n] + 2 cos [(2n - 0.3n)n] +4 cos [(2n + 0.5n)n] + 10 sin [(4n - 0.72] =6 cos [0.3m] — 3 sin [0.5 nn] + 2 cos [0.3nn} + 4.cos [0.5nn] — 10 sin(0.7an} v(x) = [8 cos (0.3nn) + 5 cos (0.5xn + 0.6435) — 10 sin (0.7nn)] ‘The discrete-time signal vin) is composed of a weighted sum of three. sinusoidal signals of normalised angular frequencies : 0.3n, 0.5 and 0.77} 3.11 CLASSIFICATION OF DISCRETE-TIME SYSTEMS Discrete-time systems are classified according to their general properties and charac- teristics. ‘They are (1) Static and Dynamic systems. (2) Time-variant and time-invariant systems.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems 35] Sol.) x(n) +x(n- 1). ‘We know that, Then), (n) + x(n - 1). If the input is delayed by & units in time, we have, (0, B) = Thelen —)] yn, b) = an —b) + x(n —k - 1) If we delay the output by & units in time then, y(n —k) = a(n —k) a(n - k= 1) (2) (1) = (2) Here, yn, k) = y(n -) So, the system is time-invariant. Gi) ya) = x(-n) If the input is delayed by & units in time and applied to the system, we have, y(n, k) = Ten —B)) = ln — Bd of) If the output is delayed by # samples, y(n —k) =al-(n—k)} = 21-0 +h) 44) (3) #(4) Here, yn, k)# yn—R) So, the system is time-variant, (iii) y(n) = x(2n) ‘The system is described by input output equation, yn) = Thx(n)]} yn) = x(2n) If the input is delayed by # unit in time and applied to the system, on, k) = Thetn — W)) y(n, k) = x(2n — k) Dd Now, if we delay the output y(n) by & unit in time, the result will be on — k) = x[2(n —k)] = xI2n — 2k] --(2) Since y(n, k) #(n —k), the system is time variant. xia) xin) 1K) = x21) »[ Dey [System POO Ser y(o) = x@2n) ->| System +[_Datay Gv) yn) = x) sin @, n If the ip is delayed by & unit in time and applied to the system, y(n, k) = a(n -h) sin on. I we delay the output by & unit in time, then yin - k) = x(n ~ k) sin @, (n - k) Since y(n — k) #y(n, k). So the system is time variant.3. Causal and Non-causal System | In a discrete-time system the n,th output sample yo] depends only on Input samples x{n] for n ng. This means tht, a system is said to be causal if the output of the system at any time n depends only on present and past inputs, but does not depend on future inputs. This can be expressed mathematfcally as, yn) = Plata), x(n - 1), x (n — 2)... | Ifa system depends not only on present and past inputs but also on futufe inputs, then it is said to be a non-causal system. 4. Stable System There are various definitions of stability we define a discrete-time systey and only if, for every bounded input, the output is also bounded. This imp response to x(n) is the sequence y(n) and if, tobe stable if es that, if the [x(n < B, or | xin) | ) x) h(n 2) i = Det ay? =p > 24 aay" & me = cup 5, 2% mo = (1/2) [1 + 274 244 2 +.. +(n+ D terms) yah i =(ay (a J =] [4 = ‘| (27-1 nt y(n) = (1/2)* [I] Problem 20. Find the impulse response of the cascade systems, if hy (n) = (- 12)" u(r) hy (n) = (1/2)" u(r). Use the convolution sum to find the response to x(n) = (1/4)" u(n). Sol. (i) To find the impulse response of the system. hin) =h,(n)*h(n) > cascade connection. = Swmin-m =| ¥ cant oar un) a ae o = [oa > cary onl 1)" % (12)* (3) 2% cyt WS", Assume that u(n) = 1.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems Te] Let us consider an LI system having an o/p atn = 1g y709) = Sy A) (ng - 2) ae The above systems can be subdivided into the two terms, one having present and past values (n < nq) and other having future values (n 27). Thus a = 1%) = [ sonar] & 1G x(n -»| & = = [h(0) a(n,) + h(1) x(n, - 1) +. + [Al 1) x(ngt 1) + A 2) x(nrg +2) +... yng) = 1 +A. ‘The first term in the sum with x(n,) x(7tg~ 1)... are the present and pastinput. Whereas the term (II) in the sum with x(n + 1), x(n, + 2)... are the future values of input. Now if, h(n)=0 forn<0. the output »(n) depends on present and past values of inputs but does not depend on future values of inputs because the second term in sum becomes zero once h(n) = 0 forn <0. Hence an LTI system is causal if and only if its impulse response is zero for n < 0. The limit of summation of the convolution formula may be modified. Thus we have, J (3.50) y(n) = ¥ hk) x(n-k) putn-k=m ma > xm) A em) y(n) = S x(k) h(n =) (3.51) If the input to a causal linear time invariant system is causal i.e., x(n) = 0 forn <0, the limit on the convolution formula can be further modified. Thus we obtain, yn)= x AM-W =D eDAM-B), mS yn) =) Wk)z (2-2) m It is clear from above eqn. that the response of the causal system to a causal input sequence is causal because y(n) = 0 for n <0. Stability. A discrete time systemis to be stable, bounded input, bounded output stable, if the output sequence y(n) remains bounded for all bounded input sequence x(n). We now develop the stability condition for an LT discrete-time system. We shall dem- onstrate that an LTI system is BIBO stable if and only if its impulse response is absolutely summable, i.c.,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems 7 ohn) = 5 tnyln=1) + a0) n 1 Yn) = Tom - + al) --(3.60) Eqn. (3.60) suggested that the computation of y(n) requires two multiplication, one addition and one memory location. The block diagram representation is shown in Fig. 3.18. This system is known as recur- sive system and )(n — 1) is called the initial condition of the system. yin) —> Fig. 3.18. Block diagram representation. In general the o/p of a causal can be expressed as, yn) =f ln - 1), y(n 2) y(n -N), x(n), x(n - 1)...... x(n - MD} ‘The block diagram representation of causal system is shown in Fig. 3.19. 40)__P yg) x01) x(n] ) Fig. 3.19. Block diagram representation of causal system. Ify(n) is only function of present and past inputs, then yn) =f a(n), x(n - 1) x(n -m)] Such a system is known as non-recursive system. The block diagram representation of causal is shown in Fig. 3.20 xo 9) «(3.61) --(3.62)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.crete Time Systems [ey] Distinct roots. For the case where eqn. (3.70) has N distinct roots 0, Op, -.-.., hy, then the most gencral solution is of the form, gyn) = Alo + A, oy 4. where, A,, A,, Ay are weighting co-efficient. Those co-efficients are determined from the initial condition specified by the term. Problem 26. Determine the homogeneous solution of the system described by the first order difference equation. y(n) + 3y(n - 1) = x(n), with initial condition y(—L) = 1. Sol. For the homogeneous solution, x(n) = 0 thus, y,(n) + 3y, (n - 1) = 0. ‘We assume solution of the form of y,(n) = oP a" + 30-120. on-lia+3]=0 + Ay Oy" ABV) a=-3 Thus, the general form of solution of homogeneous difference equation is, yy, (0) = Ao” = A-3y". ‘Using the initial condition y(- 1) = 1. We have, y(n) = - By, (2-1) Putn =0 (0) = - 39, (0-1) =- 359, 0) =A A=-3 Therefore the homogeneous solution is given by, y(n) = 3 3)" = (~8)8+1, Problem 27. Determine the homogeneous solution of 2"4 order difference equation. y(n) -y(n- D-y(n-2)=0 with initial condition y(0) = 0, y(1) = 1. Sol. Let us assume the solution of the form y(n) = 0” cet orl ot? a? (o? -o- 1 Therefore, the roots are, 1+J5 1-J5 2° 2° ‘The general solution to the homogeneous eqn. is. Mah, a4) +g e4y: a= Using the initial condition ‘y(0) = 0,9(1) =1, we have, Ath,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems [es] 4 Therefore, Jpn) = zn Buln), The total solution is, y(n) =A, 2 4A, (-17 +4 n2* u(r) ) where the constant A, and A, are determined such that the initial conditions are satisfied. To accomplish this, evaluate the given eqn. at n = 0, 1. (0) —y(— 1) ~ 29-2) = (0) = 1. y¥Q=1 and atn= 1. ‘y() = y(0) - 2y(- 1) = x(1) + 2x(0) = 24 2= 4. y=5 Using the value of (0) and y(1) in eqn. (1), we have A\+A,=1 8 2A, ~A, + 5 = 5. z “Thus the final solution forx(n) = 2"u(n) is given by, These two eqn. give A, Ay endo aye y tae yn)= ae ac: Dr +gnt". 3.21 THE IMPULSE RESPONSE OF A LTI RECURSIVE SYSTEM Impulse response of the linear time-invariant system was defined as a response of the system to a unit impulse i.e, x(n) = B(n). Now consider the problem of determining the impulse response h(n) given a linear con- stant co-efficient difference equation. In the proceeding subsection, we have described that the total response of the system to any input consists of solution to the homogeneous equation plus the particular solution. In case when the input is an impulse, then the particular solution is zero because x(n) = 0 forn > Oi.e., 3p(n) = 0. Therefore, the response of the system to an impulse consists of homogeneous solution. Problem 31. Determine the impulse response of the system described by, y(n) - ay(n - 1) = x(n) with f- I = 0. Sol. For x(n) = 8(n) above equation reduced to, yn) —ay(n - 1) = Bn), for n > 0, this reduces the homogeneous eqn. i.e., yn) —ay(n — 1) = 0.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Time Systems [eo] Determine its values and sketch the signal x(n) Determine x(4—n) [ameicrsind={...02,2,41410] @xt4—m= {02 4442.20] t 1 . From the two sequences x(n) = (3) 190) = [- aT Prove that (@) x(n) +n) = 14 ay (8) x(n) y(n) = [ Express the sequence defined by x(n) = as a weighted sum of unit-sample sequences Ans, x(n) = 4&(n + 2) —25(n + 1) —2Kn) — 250 — 1) + 45(n - 2)) Prove or disprove that x(n) =5 sin (2n+2) is period. If the sequence is periodic, determine its period. fAns. Not periodic] A discrete time system can be (1) Static or dynamic (2) Linear or non-linear (3) Time invariant or time variant (4) Causal or non-causal (5) Stable or unstable with respect to above properties examine the following systems (a) x(n) = cos [y(n)] (Ans. Static, non-linear, time-invariant causal, stable) (6) y(n) = a(n) sin (@y n) {Ans. Static, linear, time variant, causal, stable) (ec) y(n) =2{-n +3) {Ans. Dynamic, Linear, time-invariant, non-causal, stable] (@) y(n) = | an) | (Ans, Static, non-linear, time-invariant, causal, stable} (e) y(n) = x(n) ua) (Ans. Static, linear, time-invariant, causal-stable) Dy) =e, (Ans. Stable, causal, non-linear, time-invariant) Compute the convolution y(n) = x(n) * h(n) of the following signals : (a) x(n) = (0,1,-2,3,-4) [ass fags By BO 4] (6) x(n) =(0,0,1,1,1, 0) T h(n) = (1,-2, 3) (ans. (0, 0, 1,- 1,2, 2, 1,311 t fa” -asns5 © xn = {2 fheche 4 in) = {5 Sense [ane ¥ w-» -sensa] wise msaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characterization of Discrete-time System: [es] a, oho oN” ‘The fourier series representation of x(n) co1 functions. k 1, 2, ene N=. sts of N harmonically related exponential ate xin)= Doe ™ w(44) ino Now multiply e~ 2" to both sides, wea ehtems!N in) = Sy ay ofS go jtamN 0 Nea NaiN-1 2, — FR mn Ym n= DY Yiave ¥ nao a=0 kao Noi Not (2 Fe -mn =D De® 4.5) ied neo if a=1 Here, = if ae -{ otherwise 2e, Qn an Wn cog hk ns jsink 2 PRY = 008k Fn +jsink $n. =1+0. (fork =0, +N, £2N Bernt nn of k-m=0,4N,,+2N, fh v a otherwise Eqn. (4.5) is reduces to (4.6) Sat, Synthesis eqn. xin)= Dae ® io Mot 2x i a in Analysis eqn. aay Lameaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characterization of Discrete-time Systems [7] Or Plot the magnitude and phase response of a system whose impulse response h(n) =a” u(n) for a=0.5. Sol. Given xn) = ay(n— 1) +n) ‘Take fourier transform on both sides, Ye) = aeF" Ye) + Xle) yle] [1 - ce] = X(e*) _¥e*)__ 1 Xe) 1-ae Another method : Consider the difference eqn. of a 1" order systems, kn) = ayn — 1) +n), with K(n) = om ‘The particular solution to this eqn. is in the form »,{n) = Co, Substituting this in original complete eqn. 2) ~ aygin ~ 1) = xin) Ceien — qoeiein=1 Therefore, = Thus, the steady state solution of the system is, 1 ; IM) = ae om, This solution is of the form H¢ei®) eit, ‘Therefore, the frequency response of this first order system is, joy —_1__ He) = To plot H(e), we find the magnitude and phase term as, 1 He") = Toe . 1 1 1-alcos—jsina)] 1-acos@+ jasino ‘The magnitude response, | He™ | =aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characte a = = = = Yeo Y por |= Year y ante a mo If X(z) converges in some region of the complex plane, both summations in eqn. (4.12) must be finite. If the first sum of eqn. (4.12) converges, there must exist values ofr small enough for x(—n)r" to be absolutely summable. Hence the ROC for the first sum consists of all points in a circle of radius r, as shown in Fig. (4.2) where r, >r. x(n) 7 +(4.12) a) wy Fig. 4.2. ROC for >, 1- r* |, Fig. 43. ROC for > “m a If the second sum of eqn, (4.12) converges, there must exist large values ofr for which x(n)/r* is absolutely summable. Hence the ROC for the second sum consists of all points in a circle of radius, r, as shown in Fig, (4.3) where r, xn)2™ co = « = Doren at aa mm = @tay" = Y @tz"-1 i io Using infinite geometric series sum, we get X@=—t_- t-alz X@ =. a-zaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characterization of Discrete-Time Systems [v7] Sol. (a) Since the ROC is exterior of a circle, we except x(n) to be a causal sequence. ‘Thus we divide so as to obtain a series in negative power of z. Carrying out the long division, we obtain, 14421 +72? +102 + 122442 )i4227 1224427 421-2? 421-827 442% Tz? 428 42 ~142% +724 1023-724 102? - 2024 + 102% ‘Thus X(z) = 1+ 4z1 4 72? + 1003 + By comparing the relation, we have, x(n) = (1,4, 7, 10, Sn+1 t (6) When the ROC is the interior of the circle, the signal x(n) is anticausal signal. Thus we divide so as to obtain a series in power of z as follows. 22 +52? +829 +1124 221 41)22041 82-52? 82 - 162° + 82° 112? - 82° Liz? - 2225 + 112! 1425 - 1124 | Thus, X(z) = 22 + 527+ 829 + 1124+. | In this casex(n) =0 forn 2 0, thus by comparing the result with eqn. X(z)= ) x(n) 2, we get, 11,8, 5, 2, 0). +aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characterization of Discrete-Time Systems [rey] 1 O° ae z XO° Ce Xt) Zz “2 "@+De-D? XG) Ai, Aa, _ Aa z z+1 (@-D (@-1 A= XG) (+) = 0.25 z a=] 4,= 22 e-n?| =05 z lena d[X@) ‘| d 2 2 =o|e- =4|—__@-» a el 2 8 WN aelespe-a 7 _ af 2 aEtWee-z? | _(+)x2-1_3 _ 9 oe “dz|z+ij) . @+DF a+p? 40" X@) _ 0.25 0.75 0.5 z 241 2-1 @-1 0.252 0.752, 0.52 Med= yi 2-1 Go x(n) = 0.25(- 1 u(n)+ 0.75 u(n)+0.5n(1)" for 220. Problem 17. Determine the inverse z-transform of the following z-domain functions : Be? +2241 2-04 (@) Xt) = 3 (Xe) = Fae 2-4 OX@) = ye 3 2 SE ry Sol. (a) Kis SE Zee #24 282! +2241 2° 3242 82 By partial expansion, we get Xie)=34 A414 Ae z-1 Avsaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Frequency Domain Characterization of Discrete-Time Systems 725] z-1 z+081 =- 0.616. l= 031 a(n) = [0.404 (— 0.81)" - 0.154 (0.31) uln). System function : Let us defined the function He) by, Sas WO H@)= (4.19) ko From the eqn. (4.19) Yee) = Hz) X@) The function H(z) is known as the transfer function of a system or system function. When the input to the system is impulse signal x(n) = &(n) then, Yee) = He) He) =2th(n)). Problem 20. Determine the system function H(z) of i (a) y(n) + 2 yn-D+ ae y(n -2) = x(n) + x(n - 1). 3 (6) yn) = Slr) + Bet. Sol. (a) Taking z-transform of both sides, ve) + Set ve) +224 We) = Xie) +27 Xe) Ye) te Xe) "7, 3,1,1 Wie +57 Hiz)= (b) yn) yn — 2) + 2x(n) Taking z-transform of both sides, Ye)= ; 21 Y¥@) + 2X) 2 He) = 3 1-2 2 Taking inverse z-transform, h(n) =2 (- 5) un).aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.3. Find the fourier transform of x(n) = @y u(n- 1). [ Hint. Xw)= = stn) eo -£(@) yen ee [14+ 2+ =e + Zev Je —} 1-5e 3° 5.11.4 Properties of Discrete-time Fourier Transform 1. Linearity. Fla, x, (n) + a, x,(n)] =e, X(e*) +a, X, (ei), 2. Periodicity. The discrete-time FT X(e®) is periodic in w with period X(e*) = X{ei + 9) for any integer k. 3. Time shifting. If F[x(n)] = Xlel, then Fan - )) =e Xfem], Proof. Fixin- = dy x(n— eo {Let n-k=P a n=P+k] a ame BES asa > ap ao ) : [Amplitude spectrum docs not change, The phase spectrum is|changed by-cok] 4, Frequency shifting property: Ir Xle*] = Fix(n)], then Flatn) ef] = Xfe/@""?), Proof. = Yayo orion = xhaye Kom" _ xpeilo-e0)), 5, Time reversal : i Fha(n)] = Xle), then Fix(—n)] = Xie“). 6. Differentiation in frequency : If x(n) 4 Xo) nal) Ej AO.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform 6.1 INTRODUCTION Frequency analysis of a discrete time signal is usually performed on a fligital signal processor. To perform a frequency analysis on a discrete time signal, we convert time domain ‘sequence x(n) to an equivalent frequency domain representation X(«). We know that such a representation is given by the Fourier transform X(o) of the signal x(n). A difficulty encounter with the direct application of Fourier transform to discrete time signal is that the resulting representation X(«) becomes continuous function}of frequency. Hence, it is unsuitable for digital processing. In this section we consider the representation of a discrete time signal x(h) by samples of its spectrum X(o). Such a frequency domain sampling leads to the discrete fourjer transform, a second transform domain representation that is applicable only to a finite length sequences. 6.2: THE DISCRETE FOURIER TRANSFORM : (FOURIER REPRES| OF PERIODIC SIGNALS) Before defining the DFT, we consider the sampling of Fourier transform discrete time sequence. Hence, we establish the relationship between the sai transform and DFT. To start, let us consider an aperiodic discrete time sequ Fourier transform, INTATION ‘an aperiodic X)= ze Let us sample X(w) periodically in a frequency at a spacing Aw radian} successive samples. As we know that X(co) is periodic with period 2n, we require the samples mental frequency range. We take N equidistance samples in the internal of 0 A.D) between the 2 spacing Aw = = which is shown in Fig. (6.1). ak Let us evaluate eqn. (6.1) at @ = =~, we have, 158aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Nast Xt) = azo 1 | Ni2 k=O | Xt) = Saw k= odd 0 k=even Ok XG). then x(n +N)=x(n) then X(k +N) = X(k) | (6.15) Ry | Proof. Xth) = J) me" ® a N-1 Pah), | Xh+N)= J xe” * | ao | Nol jie =D xe Nc LS xine 8 a0 aso Xk +N) = X(). (3) Circular shift of a sequence : This property is analogous to the time shifting property of the DTF1, but with some difference. Lot us consider a sequence x(n) of length N which is defined for 0N. For any arbitrary int¢ger k, the shift sequence x,(n) = x(n — k) is no longer defined for the range 0 Xb) Da AQ) xn) > 2 XX) ‘The proof of this property is similar to circular convolution. (9) Circular correlator: For complex valued sequence x(n) and y(n). per If a(n) <> Xb). DFT an) > Yb) « (6.32) (6.33)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.2. - Y@)= LY xine wo Yik)=X(k) OSkS15 Bo 10:11 2 13 14 15 Therefore 16-point DFT of interpolated signal y(n) contains two copies x(n). Since for ¥(k), N = 8 therefore ¥() is periodic with period 8. Problem 6. Compute the DFT of sequence defined by x(n) = (— * for fa)N=3 (b)N=4 (c)N=even(d)N = odd. Nel -itmn Na} Sol. Xk)= Y xnje = LW", wo no Ne = Devt we m0 Net Nz1 =D (eax Wy" = DY - Wytl" so aso wick wail SU TsWy + Wye (a)for N=3 X(k)= 1 1+ X= Te jan 3 7n3 1-c (6)forN=4 = X(k)= >} =O for W,'#-1 or k#2 14 we of 8 point DFT ofaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform Problem 12. Perform circular convolution of the two sequences x(n) =(2,1,2,1) and xy (n)= (1, 2, 3,41. Sol. Method 1: Graphical method of computing circular convolution The circular convolution of x(n) and x,(n) is given by, Noa N-1 xq(m)= Dy 2101) x9(m =m) = SY) 2462) x2, (0), Fer) ex] where p, (2) = x,(m —n). The given sequences can be represented as points on acircle as shown below, the folded sequence z,(~ n) is also represented on the circle. x(t)=1 x1) = 2 (2) = 2: K,(0) = 2 xf2) = 3: (0) = 1 x,(@)=1 x3) = 4 xQ=4 (2) = 91 x20) = 1 xi) =2 when m =0 Not 3 £40) = D1) 2, m (2) =D) x(n) x9, (0) iso when m = 1 Nat 3 x(D= x y(n) x, (1-n)= yaw 2,10), when m =2 xo 3 x)= ¥ xin) -a)= DV, a), 210) mM nsoaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.200 q n -1}] of/ij2]s |4]5] 6 8 x,(n) 4a}-—- h(n) -1 1 An) aje-l | R(G- n) = hen) 1{-1 A(T 1) = hala) 1| + A(B—n) = Ayn) =1 yim) = y x4(n) A(m -n) = > X4(N) y(n) ; m= 6, 7,8 4 whenm=6 — y,(6)= Dy *4(7)hgln) =0-440=-4 ms 8 whenm=7 —y,(1)= >, “an hy(n) 24444028 mM 8 whenm=8 —y,(8)= Dy *4(™) g(n) =0-440= For] To combine the output of the convolution of each section. It can Ye observed that tho last sample in an output sequence overlaps with the first cample of next oftput sequence. In this method the overall output is obtained by combining the outputs of th¢ convolution of each section. The overlapped portions (or samples) are added while combining the output. ‘The output of all sections can be represented in a table as shown belo m of} 1] 2}]s3]4]5] 6] 7] 8 ya) | -1 | 2] -1 ygm) ,72 | 4 | -2 yafm) \ -3 | 6 | -3 yarn) ‘ -4 | 8] -4 yon) -1 ] 2/-3 | 4/]-5 | 6|-7 ] 8] -4 ym) = x(n) * h(n) = {— 1, 2, — 3, 4,— 5, 6, — 7, 8, ~ 4}.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform [205] ‘Method 2 : In method-II the overlapping samples are placed at the end of the section, Each section of longer sequence is converted to 3-sample sequence, using the samples of original longer sequence as shown below. It can be observed that the last sample of x,(n) is placed as overlapping sample at the end of x,(n). The last sample of x,(n) is placed as overlapping sample at the end of x(n). The last samplez,(n) is placed as overlapping sample at the end ofz,(n). Since there is no previous section for x,(n), the overlapping sample of z,(n) is taken as zero. x(n)=1; n=0 | x(n)=2; n=2 s-1jnsl =-2;n=3 =-3;n=5 j mse s-ljn=4 =-2;n=6 Now perform circular convolution of each section with h(n). The output sequence ob- tained from circular convolution will have three samples. The circular convolution of each section is performed by tabular method as shown below. Here h(n) starts atn =n, = 0 x,(n) starts at n 0, y,(m) will start atn =n, +n, =0+ xn)=8; nad | x(nd=4; n=6 0 x,{n) starts at n » ym) will start atn =ngtn, = 2+ x,(n) starts at n= nge4, ym) will start atn =n, +n,=4+0=4 x,(n) starts at n = » ym) will start at. tm, =6 + Convolution of section 1: n -2 0 1 2 x(n) 1 1 0 An) -1 1 0 A(—n) =ho(n) 0 1 -1 0 1 AQ =n) =A) 0 1 =1 o AZ -n)=A,fn) 0 1 -1 9(m) =x,(n) @ hin) Sno h(m-n) oe 9m) =D x:(n)h,,(n) , m= 0, 1,2 a when m=0 ¥(0) = Y2(0) g(a) =- 14040 when m=1 yD = Dizm) h(n) = 14+14+0=2 when m =2 912) = Vix) im) = 0-140aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.6.8 FAST FOURIER TRANSFORMS ALGORITHMS 6.8.1 Introduction The Discrete Fourier transform plays an important role in many appli signal processing like linear filtering, correlation analysis and spectrum a term Fast Fourier Transform(FFT) usually refers to a cla: puting the DFT. The basic idea behind all fast algorithms for computing the FFT ip to decompose leads to a family of an efficient computational algorithms known collepti algorithms. Consider the DFT of a finite length of sequence Not 2itkin 1 X= Dixie X= ¥ alm) Wy" azo nao Similarly, the IDFT becomes, xa xn)= = YG) We", 2 =0,1, 46.41) me ‘The sequence x(n) is also assumed to be a complex value. We observe from eqn. (6.40), that direct computation of X(k) requires f of arithmetic operations. (DN complex multiplications for each value of k. (2)(N — 1) complex additions for each value of k. (3) N? complex multiplications, for N values of k. (4) NIN - 1) complex additions, for N values of k. For a complex valued sequence x(n) of N-point, the DFT can be expregsed as, Not x)= [20m cos ms + x(n) sin askn 2 --(6.42) Not Xk) =- a [nonin ao 2nkn = x(n) cos (6.43) X(b) = Xq(h) + JKR). (6.44)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Thus, we have (0) + W,° F,(0) (1) + Wat FC) X(2) = F\(2) + Wy? F,2) (3) + Wy? F,(8) (0) + Wy* F,(0) CD) + Wy Fy) (2) + Wy 8 F(2) (3) + Wy? F,(3). Figure shows this computation X(0) + x(0) von v2 point samples *@) - DFT xq) 4 X(4}0 + 82 = < point ort X(¢lo—+ xt}o—+ ‘aa N/2 point ‘samples: O) DFT, em) x(}o > 82= OFT x(7]0—> xq) In addition if we use, Wy(k + N/2) = ae BRAN, grit = g-fiehiN Writ = — Wet (6.55) We can have Table Wy! =- Wy? Using table 3 we can write, X(h) = F\(k) + Wyt Fy). (6.56) X(k + N/2) = F,(k) ~ Wyt F(A) 26.57)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform [ea ] ‘The relation between the samples of various sequences are given below : even V4, 0) = £,(0) = x(0) V2\(0) = f,(0) = a(1) Vy) = f,(2) = (4) | Va\(1) = fy(2) =x (5) V,,() = f,2n) = x(4n) odd Vip (0) = (1) = (2) V,,(1) = (8) = 2(6) V, p(n) = f,(2n + 1) = x(4n + 2). Similarly, V(r) = fy(2n) = 2(4n +1) Ven) = f(2n + 1) = x(4n + 3). The first stage of computation : In the first stage of computation the two point DFTs of the 2-point sequences are com- puted. Vox(0) = f,(1) = (8) Vin(1) = 8) = 2(7) Let, Vy) = DFTIV,(n). Va) = a) Ww" for =0,1. when k=0 : V1 1(0) Wye? +94, (D Wyyg? = 0,(0) +4,0) V, (0) = (0) + x(4) when k=1 Vj) = Vy) = 94 (0) Way? + 0D) Way? =, -,,) V,,(1) = (0) 244) Let Vk) = DFTIv fn}. na Vk) = & ato Wi for k= 0, 1. when k =0 Vist = 0490) Wy? + 05901) Way? when k =1 Vag D = 04200) Way? + Yy0(D) Wau? = 049 (0) —U,,(1) Vg) = (2) - 216)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform [27] ‘The butterfly computations are shown below : 3+3=6 Output DFT sequence = (6, 1-j,0, 1 +/,6,1-J,0, 1+J). Third stage computation : Input DFT sequence = (6, 1—j, 0,1+J,6,1-j,0,1+J1. ‘The phase factors involved in third stage computation are W,0, W,!, W,? and W,°. ‘The butterfly computations of third stage are shown below : 6+6=12 an) _p(t_it ce oe) 0-0(-))=0 spa sn =14)2414 ‘The o/p DFT sequence = {12, 1 A14, 0, 1 +j0.414, 0, 1 + j2.414) DFT x(n) = X(k) = (12, 1 -j2.414, 0, 1-j0.414, 0, 1 +j0.414, 0, 1 +j2. 414). | Xe) | = (12, 2.61, 0, 1.08, 0, 1.08, 0, 2.61). X(k) = {0, - 0.374, 0,~0.12n, 0 , 0.12n, 0, 0.37)aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Discrete Fourier Transform Using equation (6.80) the eqns. (6.75) and (6.77) can be written as; nasa G28) = LD du) Wyn! =D,,00. (6.82) ao N41 Gark+D= DY dn) Wu =D.) --(6.83) a Using eqn. (6.81) the eqns. (6.78) and (6.79) can be written as, Naot G,(2k) = YD dau(n) Wa” =D, (b) + (6.84) ex} Nast G(2k+1)= )) dan(n) Wya" =Do,(k) ---(6.88) no where, D,;(), Dyo(#), Dpy(k) and Dyo(h) are N/4 point DFTs of d,,(n), dyp(1t), day(n) and dyo(n) respectively. Butterfly computation involves the following operations (i) In each computation two complex numbers a and b are considered. Gi) The sum of the two complex number is computed which forms a new complex number (ii) Then subtract the complex number 6 from a to get the term (a — b). The difference term (a ~ b) is multiplied with the phase factor Wy" to form a new complex number B. —e———*A=atb B=(a-b)Wy Fig. 6.19 6.11.1 The 8-Point DFT Using Radix-2 DIF FFT Let 2(n) be an 8-point sequence. First stage of computation. In the first stage of computation, two numbers of 4-point sequences g,(n) and g,(n) are obtained using equations (6.69) and (6.70) respectively. g,(n) = x(n) + x(n + N/2) gyn) =x(n) +2(n+ 4) forn =0, 1,2, 3. whenn=0, — g,(0) =x(0) +x(4) whenn=1, — g,(1) =2(1) +x(5) when (2) = 22) + x(6) when n = 3, 8,63) = x(3) + x(7) g_{n) = beln) — x(n + N/2)] Ny" gon) = Len) — xin + 4)] Wy" forn = 0, 1,2,3 when n =0, G0) = fx(0) —x(4)]W,°aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.DFT of x(n) by radix-2 DIF FFT. Sol. For 8-point DFT by radix-2 FFT we require 3-stages of computafi fly computation in each stage. The given sequence is the i/p to the first stage. For other stages of| o/p of the previous stage will be the i/p for the current stage. First stage of computation : The input sequence = { 2, 2, 2, 2, 1,1, 1, 1) ‘The phase factors involved in first stage of computations are W,°, i omputations, the 1, We! and W,8, 1 We=1W= 5 (0) Fig. 6.24 ‘The sings joemoe ee = (3,3, 3, 3, 0.707, —j 0.707, —j, — 0.707 —j 0.707} Second stage of computation : ‘The ifp sequence for 2™ stage = (8, 3, 3, 3, 0.707, —j 0.707, ~j, — 0.747, ~j 0.707} 3+3=6 34926 3-320 @-AH-j) =O}aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.gnal Processing Substituting eqn. (7.9) in eqn. (7.8) yields, Fio}=q Dy XlQ- ma, ote | | 1< 2nm | FEOl= 5 Z x{{a - 2=)| | (7.10) ‘Thus the fourier transform of the sampled signal is given by an infipite sum of shifted replicas of the Fourier transform of the original signal. Now consider the signal x(¢) is band limited to/,,. Thatis the highes{ frequency compo- nent of a(t) is f,,. Then, x§%)=0 for] 21 >9,(0,, In eqn. (7.10) the term X ( ( - 2) is the shifting of XQ) from, Hence X, (2) is the sum of shifted replicas of x) centering at oa, mb 0,#1,42. The Fig. [7.5(a)] shows the plot of xe for different value of z. 4xG0) 1 =9, 0 a Fig. 7.5 (a) z T frequency spectrum of TX,(/2) in frequency range (- spectrum X(/M) can be recovered from X,(/2) by using a low pass filter which X J The same explanation holds for 7 =O, d as a result the ‘Wo sce that if = >, the replicas will not overlap as in Fig. (7.5(6)] 2) is identical to X(jp). The frequency hassharp cut off atQ=aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Processing of Gontinuous Signal [55] ‘The sampling rate, f, = 8H, Q,= 2xf, = 21(8) = 160 2,,= 2nf,, = 10n fy = BHz. Nyquist rate = 2f,, = 10 Hz. ‘The sampling rate is less than Nyquist rate. So, the original signal cannot be recovered from the samples. The frequency spectra of sampled x(t) is given by, X,Ga) x,G0)= 2 J, 1018180 + 10 — nO) + m182- 10 — 20,91) iT. an 1 where, Q,=a and 7 =f,=8 DY 10n18:0 + 10n - 16x) + 52 - 10% — 1620) X,(@)= = 80x >, [82+ 10n- 16nx) +80 - 10n— 16Kn)] ‘The plot of amplitude spectrum for | © | < 30n is shown in figure. ‘Niased component X52) ‘eo, 26x 2x “10x 6x @x 10x 2x (x Problem 3. A signal x(t) = sinc (150nt) is sampled at a rate of (a) 100 Hz (b) 200 Hz (c) 300 Hz. For each of these three cases, explain if you can recover the signal from the sampled signal. Sol. Given xt) = sine (150 x). 4x0) The spectrum of the signalx(t) is a rectangular pulse with a band width (maximum frequency component) of Las 150x rad/sec is shown in figure.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Processing of Continuous Signal [25] 1 1 HO) HC 9) = ag ic The above relations tell us that this function has poles in the LHP as well as in the RHP, because of the presence of two factors H(s) and H(— s). If H(s) has roots in the LHP then H(- s) has the corresponding roots in the RHP. These roots we can set by equating the denominator to zero. (7.38) ie, 1+@s9%=0 (7.39) For N odd, Eqn. (7.39) can be written as, oN a La etm (7.40) Now the roots of eqn. (7.39) can be found as, 8p ON KET, 2, cease BN. For Neven, Eqn. (7.89) reduces to, # =~ 1 =e8K-Dx, which gives, fy 2A C-DMN for K = 1,2, For N=3, Eqn. (7.39) becomes, aL K=1, =1,2, Now the roots of eqn. (7.39) can be found as, ; x © 8, =e" = cos 5 +jsin 5 = 0.5 +) 0.866 8 2 cos 2 +isin - 0.5 +j 0.866 8, = 0"3 = cosn +j sin =-1 4a s,= ott = 008 SE 4 jin SE 0.5 jo.866 oH con + jsin =0.5-j 0.866 s 8, =e?" = cos In +jsin 2n=1. All the above poles in the s-plane as shown in Fig. (7.18). It is found that the angular 360° 60" |, which in this case is equal to 60° and all the separation between the poles is given by > poles lie on a unit circle. Stability. To ensure stability, considering only the poles that lie in the left half of the s-plane we can write the denominator of the transfer function H(s) as, (+1) ((s + 0.5)? + (0.866)%) = (s+ 1) (2 +540. Therefore, the transfer function of a3" order Butterworth filter for cut off frequency 2. = Ir/sis, 1 Me)= Cee serDaa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Processing of Continuous Signals [283] Problem 6. Given the specifications a, = 1 dB ; 4, = 30 dB, 9, = 200 r/s, 2, = 600 r/s. Determine the order of the filter. Sol. Tofind A: To find k: log Vk Jog (62.115) jog@y 72758 Rounding off N to the next integer, we get, N=4 a cael oan [10° — 728 Sol. The magnitude square function of Butterworth analog lowpass filter is given by, 1HGQ) Problem 7. Prove that Q, = AD We know, [Ha = 2) ae 2) 2 ee (Z } Comparing in eqn. (1) and (2), we get a 7 a 2[ Q ve( 2) aaeer(2 fa) > (2) P (ay (a Q, Q.) * Simplifying the above eqn. by substituting e={10"* -1, weobtain (fy =10" —1,aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Processing of Continuous Sig) [277] Low pass-to-band pass transformation. Consider a band pass filter with lower band edge frequency , and upper band edge frequency, The tranaformation for converting lowpass analog filter, with band edge frequency Q,, into bandpass filter can be accomplished by first converting the lowpass filter into another lowpass filter with band edge frequency 0,’ = 1 and then performing transformation, 7 +2, 9, soa = (787) > XQ, -2)) 8 we can also obtain the same result in a single step by means of the transformation --(7.88) Thuswehave —_H,,(s)= H, (7.89) Low pass-to-band stop transformation. To convert a lowpass analog filter with band edge frequency Q, into the band stop filter, the tansformation is simply the inverse of eqn. (7.87) with addition factor: 2, serving to normalised for the band edge frequency of the lowpass filter. Thus we have sa, 40u- Ov .(7.90) 740, O, sch oi (2y - 2) which gives H,,(S)=H, a, Se= Su) (7.9) All the above four transformations are summarised in Table (7.2) Table 7.2. Frequency transformations for analog filter Filter type Transformation Band edge frequency of new filter Low pass ss Se, a” a ‘y High pass ss 2p 8p" 2," 3 2 Band ssa, 249% jand pass 9, Fe 2,9, ‘(Q, ~ Oy) Band stop 599, Teo o, 2, 2, into high pass filter with band edge frequency Q,’ and band stop filter with upper and lower band edge frequencies 2, and Q, respectively.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.6 0.106 x H(s) for Q, = 0.106 x can be obtained by substituting S > in H@). H(s) = = Dano * 4248+ 1 HO) = 93 aoe Problem 13. Obtain an analog Chebyshev filter transfer fuction that s straints 1 Yes lHOM | <1 ;0sas2 | HGQ) | <01 ;Q24. Sol. From the given data we can find that 1 vive 1-01 Vi +” Q,=2 and Q,=4, from which we ean obtaine=1 and 2=9.95. ‘We know, sh cosh"! Ve _ cosh? Fy = = 2.269 = 3. N cosh? 22 cosh" 2 Q, Finding the values of a and, poets five? =2.414. a-o [ee 2 (2.414)Y3 — (2.414)-¥9 ”. P 2 a= 0596 | UN youn us eg [em ew] _ [e410 + (241. b= Q, aw) =2 ([sae* saat 6 = 2.087aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Filter Structures 8.1 INTRODUCTION In this chapter, we consider the realization problem of causal IIR and FIR transfer functions and outline realization methods based on both the time-domain and the transform- domain representations. Here, we describe the most commonly employed methods to imple- ment the digital filter structure from either its difference equation, its unit sample response, or its z-transform. A structural representation using interconnected basic building blocks is the first step in the hardware or software implementation of an LTI digital filter. The struc- tural representation provides the relations between some pertinent internal variable with the input and the output that in turn provide the keys to the implementation. There are various forms of the structural representations of a digital filter. We review in this chapter two such representations, and then describe some popular schemes for the realization of real causal ITR. and FIR digital filters. ‘The digital filter structure determine directly from either the difference equation or the system function is called the direct form-I. An alternative view of the same equation results in the memory efficient structure, called the Direct Form-II. Digital filter structures as cascade, parallel and lattice structures are shown to have the advantages in terms of hardware imple- mentation. 8.2 SYSTEM DESCRIBING EQUATIONS ‘The equation that describe the input and output relationship, in the time and z-trans- form domain, has been defined in the previous chapters. They are repeated here for reference. ‘The linear time invariant system is described by the difference equation of form, X * y(n) =— J) ay yin-h)+ Yb x(n -») =-AB.1) mm mo or equivalently, N u Y a n-b = Yb xa-b) (8.2) io i where a, and 6, are constants with ag= 1. As we have seen by mean of z-transform such a class of LITI discrete-time systems aro also characterized by rational system function. 295aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Digital Filter Structures [205] Problem 4. Using first order section, obtain a cascade realization for Lo Vyghet te) ~tet|a-te ie) De Sol. The H(e) can be decomposed into the three section as H(z) = H,(z) H,(e) H,(e). He) = where, Hy ; 1-9 The cascade form structure is shown in Fig. (8.13). Hye) Fig. 8.13 Problem 5. Obtain the cascade form structure for the system characterized by ya) =F xn-v- Z y(n 2) + x(n) + Fain Dv. Sol. The system function of above system is given by (-i') l+ Ae". a Nee | HG) = Yee" 2-8-9] (8.34) at N-3 * N-1)_ ae in for N is even, and Hee) =4{= Fe O-PS aot] (8.35) n=0 for N is odd.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.4. Realize the following system functions using a minimum number of multipliers 3 1 (ayHeyait bate o eay dees, 2 4 2 Ans. x0) -O = 1 1 (6) H(z) = 1+ grit gets. Ans. xn) v2aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.Signal Processing 1 (@) H@) = asc : deatece? (1+ (e) H@) = i aeget cers 4 <9 OH = a2 oe = ees) 14. Find the digital network in direct and transposed form and get the tramsfer function of the system described by y(n) = x(n) + 0.5 x(n — 1) +0.4.2(n— 2)-0.6 y(n - 1).aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.aa You have either reached a page that is unavailable for viewing or reached your viewing limit for this book.About the Book This book deals with the analysis of Digital Signal Processing in a lucid and precise style. There are about 200 solved problems apart ftom exercises. This book covers the latest syllabus prescribed by the Anna University for Electrical and Electronics engineering students. Exercise problems and review questions are included at the end of each chapter. All the above aspects should make this book extremely valuable for engineering students preparing for Anna University examinations as well as for practicing engineers. About the Author C. Ramesh Babu Durai graduated from Arulmigu Kalasalingam College of Engineering, Srivilliputhur and did his post graduate studies at Hindustan College of Engineering, Chennai. He is a faculty member of the department of Electrical and Electronics Engineering, Hindustan College of Engineering, Chennai. He has more than six years of teaching experience and the college has honoured him by conferring on him “The Best Teacher Award”. His field of interest includes Control Systems, Advanced Digital Signal Ptocessing and Electromagnetic theory. TSBNSI- 7008-7368 LAXMI PUBLICATIONS (P) LTD |: | |