ACOUSTICSACOUSTICS Leo L. Beranek 1984 Edition Acoustic Laboratory Massachusetts Institute of Technology Bolt Beranek and Newman, Inc. 1993 Edition 975 Memorial Drive, Suite 804 Cambridge, MA 02138Library of Congress Catalog Card Number: 86-70671 International Standard Book Number: 0-88318-494-X Copyright©1954, 1986, 1990, 1993, 1996, by the Acoustical Society of America All rights reserved. No part ofthis publication may be eproduced, stored in a retrieval system, or transmitted, in any form or by any means (electronic, mechanical, photocopying, recording, or otherwise) without the prior written permission of the publisher. Published by the Acoustical Society of America though the American Institute of Physics, Ine '500 Sunnyside Blvd, Woodbury, New York 11797 Printed in the United States of America PREFACE Acoustics isa most fascinating subject. Music, architecture, engineer- ing, science, drama, medicine, psychology, and linguistics all seek from it answers to basic questions in their fields. In the Acoustics Laboratory at MLL-T. students may be found working on such diversified problems as auditorium and studio design, loudspeaker design, subjective percop- tion of complex sounds, production of synthetic speech, propagation of sound in the atmosphere, dispersion of sound in liquids, reduction of noise from jetaireraft engines, and ultrasonic detection of brain tumors. The annual meetings of the Acoustical Society of America are veritable five- ring shows, with papers and symposia on subjects in all the above-named fields, Opportunites for employment are abundant today because man ‘agement in industry has recognized the important contributions that acoustics makes both to the improvement of their products and to the betterment of employee working conditions. There is no easy road to an understanding of present-day acoustics. First the student must acquire the vocabulary that is peculiar to the subject. Then he must assimilate the Iaws governing sound propagation and sound radiation, resonance, and the behavior of transducers in an Acoustic medium, Last, but certainly not of least importance, he must Jearn to understand the hearing characteristics of people and the reac- tions of listeners to sounds and noises. ‘This book is the outgrowth of course in acoustics that the author has taught to seniors and to first-year graduate students in electrical engineering and communication physies. The basic wave equation and some of its more interesting solutions are discussed in detail in the first part of the text. The radiation of sound, components of acoustical sys- tems, microphones, loudspeakers, and horns are treated in sufficient detail to allow the serious student to enter into electroacoustic design. ‘There is an extensive treatment of such important problems as sound in enclosures, methods for noise reduction, hearing, speech intelligibility, and psychoacoustic criteria for comfort, for satisfactory speech intelligi Dility, and for pleasant listening conditions The book differs in one important respect from conventional texts on acoustics in that it emphasizes the practical application of electrical circuit theory in the solution of a wide variety of problems. Wherever possible, the background of the electrical engineer and the communica- tion physicists utilized in explaining acoustical conceptsPREFACE ‘The high-fidelity expert will find the chapters onloudapesker enclosures, horus, and rooms particularly interesting because they show how the per formance of loudspeakers either in baffles oF attached to horns may be accurately and simply calculated. ‘These chaptere alco ilustrate the necessity of considering in design the overall system, including the amplifier, the loudspeaker, the baflle or horn and considering also the room in which they are to be operated. Numerical examples and sum- rary charts are given to facilitate application ofthis material to music- reproduction systems, Tn view of the incressed interest in noise control, the author has pt this subject in mind in waiting Chapters 1, 2, 4, and 10 to 18. ‘These chapters served as the basis of a special summer program on noise reduction at MILT. in 1958. ‘The material of Chapters 11 and 1 is new, and it is hoped that it will be of value to those interested in noise and its effect on human beings, Tn short, the engineer or scientist who wishes to practice in the field of coustios and who does not intend to confine his efforts to theoretical tatters must know the material of this text Problems for each chapter are included at the end ofthe text for use by thestudent. References to collateral reading in English are given in the text, although no attempt has been made to give a bibliography of the primary sources of material. Suggestions to instructors for best use of the text aro given immodintely after this preface. ‘The author wishes to express his deep appreciation to Francis M. Wiener and Rudolph H. Nichols, Jr for their assistance in the detailed review and editing of the text and the preparation of some original material. Many members of the Acoustics Laboratory at M.L:T. have read one or more chapters and have given valuable assistance to the author. Of these, particular mention is made of Mary Anne Summer- field, Walter A. osenblith, Kenneth N. Stevens, Jerome R. Cox, Jordan J. Baruch, Joanne J. English, and Norman Doeling ‘The illustrations are due to the highly capable and untiring efforts of Clare Twardsik. ‘The author is deeply indebted to his typist, Eliasbeth H. Jones, to his secretary, Lydia Bonazzoli, and to his wife, Phylis, who ‘made it possible for him to complete the text within a reasonably short span of time. Leo L. Beranex SUGGESTIONS FOR INSTRUCTORS ‘This text is divided into thirteen chapters, comprising thirty-two parts. Bach patt is intended to he approximately 1 week's work, although this will vary among students owing to differences in their previous training. If the entire class expects to take a full year of acoustics, the parts should be taught in sequence, with the exception of Part XXVIII, Meas- urement of Acoustic Levels, which may be referred to in associated labor ratory experiments and demonstrations throughout the course. If only ‘8 purt of the eluss plans to continue throngh both terms, thefundamental ‘material should be taught in the first term and the more applied material in the second, One suggested division, in this case, is as follows: Birst Term Second Term Part I. Introduetion Part V. Energy Density ‘and Intensity Part I. ‘Terminology Part IX. Circuit’ Theorems, Part HI. ‘The Wave Bqua- ergy and Power Part XI. Diroctivity Index Part IV tions of the and Directivity Wave Equation Factor Part VI. Mochanieal Cire Patt XIV, General Characters ae isties of Micro- — phones Pare VIt. Acoustical Part XV. Pressure gin) Microphones, Part VIL “Pransducers Part XVI. Gradient and Com- Part X.Directivity bination Micro Patterns phones oe Part XVII. Design Factors Af Part XL aiaion Impede leas aera ve ator Loudspeakers Part XIU. Acoustic Part XX. Bass Reflex Enclo- Hlements cae Part XVI. Basie Theory of Part XI Hora Driving Direot-radiator Units Lanulspeakers Part NXUL Hornsvill SUGGESTIONS FOR INSTRUCTORS First Term Second Term, Part XIX. Simple Enclosures Part XXV, Sound Transmis- sion through Walls Part XXII, Sound Fields in between Enclosures Small Regularly Part XXVI. Noise Control Pro- cedures and Noise Shaped Brelonures ‘ Part XXIV. Sound Fields in Pitt XXVIT Acmatic Tanemi Large Irregularly ‘at Seen Shaped Bnelowures PAC XXIX. Reciprocity of Transducers Part XVII. Measurement of Par XXXI. Speech Acoustic Levele Intelligiblity Part XXXII. Psychoacoustie Part XXX, Hearing Criteria A course in acousties should be accompanied by a set of well-planned JIaboratory experiments. For example, the material of the first few chap- ters will be more significant if accompanied by a laboratory experiment, (on noise measurement. ‘This will familiarize the student with the meas- urement of sound pressure and with the use of a frequency analyzer He will appreciate more fully the meaning of sound pressure, sound intensity, decibels, sound energy density, and power level; and he will understand the accuracy with which noise can be measured, A suggested minimum of 10 experiments, listed both numerically for a year's course and by term, is as follows: First Term Second Term No. 1. Noise measurement, No. 3. Free-feld calibration of ‘microphones No. 5. Design and testing of a loudspeaker baffle No. 8. Prediction and control of noise in a ventilating No.2. Measurement of the cone stants of an electro- ‘mechanical transducer No.4. Measurement of free-field response of a loudspeaker system No. 6. Study of sound fields ina No, 9, Audiometsic testing of small rectangular hearing enclosure No. 10. Application of psycho- acoustic eriteria in the design of an auditorium, No.7. Study of sound fields in a large irregular enclosus An assignment of two problems per week should provide sufficient application of the material of the text, ‘The short list of problems for each chapter should he supplemented by timely problems derived from the instructor's experience, CONTENTS PREFACE SUGGESTIONS FOR INSTRUCTORS CHAPTER 1. INTRODUCTION AND TERMINOLOGY rant I Introduction pane IL Terminology CHAPTER 2. THE WAVE EQUATION AND SOLUTIONS vant IIL The Wore Equation vant IV Solutions of the Wave Equation pant V_ Energy Density and Intenaty CHAPTER 3. ELECTRO-MECHANO-ACOUSTICAT. CIRCUITS Pant VI" Mechanieal Cirewite pat VIL Acoustical Cireuite part VIII Transducers rant IX Circuit Theorems, Bnergy, and Power CHAPTER 4, RADIATION OF SOUND vane X- Directiity Patterns vant XI Directivity Index and Directivty Factor CHAPTER 5. ACOUSTIC COMPONENTS pax XIT Rodiation Impedances vant XIII Acoustic Blomenta CHAPTER 6, MICROPHONES rant XIV. General Characeritice of Microphones pane XV Preseure Microphones vant XVI Gradiont and Combination Microphones CHAPTER 7, DIRECT-RADIATOR LOUDSPEAKERS vant XVII Basic Theory of Divect-radtotor Loudspeakers Pant XVIII Design Factors Affecting Direc-radiator Lowdepeater Performance 6 16 a a 62 0 1 a o 109 16 16 128, 14 a 150 18, 183 1832 CONTENTS JAPTER 8, LOUDSPEAKER ENCLOSURES vane XIN Simple Bnolonures pant XX Baswreler Bnelosures CHAPTER 9, HORN LOUDSPEAKERS. pant XXE_ Horm Driving Wits rawr -XXIL Horns CHAPTER 10. SOUND IN ENCLOSURES ant XXII Sound Fields in Small Regularly Shaped Enclosures pant XXIV Sound Fielda in Large Irregularly Shaped Bnelosures rant XXV_ Sound Tronemiseion throuph Walls between Enclosures CHAPTER 11, NOISE CONTROL pant XXVI Procedures and Sources vant XXVIL Acoustic Pronemission Pathe CHAPTER 12, ACOUSTIC MEASUREMENTS pane XXVIT Measurement of Acoustic Lele vant XXIX Reciprocity Calibration of Transducers CHAPTER 13. HEARING, SPEECH INTELLIGIBILITY, AND PSYCHOACOUSTIC CRITERIA pane XXX Hearing Pant XXX Speech Inteligibility rant XXXII Poychoacoustic Criteria PROBLEMS. DIX J, DECIBEL CONVERSION ‘TABLES APPENDIX Il, CONVERSION FACTORS ‘APPENDIX IL APP! INDEX 208 208 239 259 250 268 285 285 298 at 392 saz 45 361 361 317 388 388 406 az 431 464 469 479 PREFACE TO THE PAPERBACK EDITION With the advent ofthe compact dise, with miniature high-fidelity systems ambulating everywhere, and with emphasis on combination voice and | in'a medium Gee Fig: 2 where the incremeefromlefciorahtatacpace rate | wound bremure» cheng in apece at 2 a op/or (re Pag 2) pace rato gradp = 128 4 522 4 2 where i,j and ae unit velar in the and diectins, respectively, and Pit fe remire ats pone Assume that the sides of the box are completely frictionless; i.e, any viscous drag between gas particles inside the box and those outside is negligible. ‘Thus the only forces acting on the enclosed gas are due to the pressures at the faces of the box. The difference between the forces acting on the two sides of our tiny’ box of gas is equal to the rate at which the force changes with distance times the ineremental length of the box: Forge seting bo acoerat he box Inthe | Fore ating ta aelrte the box in the positive x dretion = ~ (32.2) ay a | postive dieton = ~ [1 (222) ay ae te) | 4 (2249) see + (22 2) ae ay] an) Note that the positive gradient causes an acceleration of the box in the negative direction of 2 + Nonvector derivations of the wave equation are given in Rapleigh, “Theory of Sound,” Vol. 2,pp. I-15, Macmillan & Co.,[ih, London, 1806; P.M. Morse, “Vibra tion and Sound," 2d ed, pp. 217-225, MeCiraw-Hill Book Company, Ine., New York, 1948; L. B. Kinsler and A. R. Frey, “Fundamentals of Acoustic” pp. 118-137, John Wiley & Sons, Ine, New York, 1950; R. W. B. Stephens and A. F, Bave, “Wave Motion and Sound,” pp. 32-13, 400-406, Edward Amold & Co,, London, 1950; and other pes, A vector derivation of the wave equation ie given in two papers that must be read fogethor: W.J. Cunningham, Application of Vector Analysis tothe Wave Equation, J Acoust Soe. Amer, 22: 61 (1950); nd R. Vo Hartley, Note on “Application of ‘Vector Analysis to the Wave Equation,” J. Acous, So. Amer, 22: 511 (1950).18 THE WAVE EQUATION AND SOLUTIONS — (Chap. 2 f--2 aw fener a By Newton's aw, the foree per nit volume (/7) of Ba. (2) must be equal tothe tine rie of change othe momentum prunit vlan othe box. We have already assured tat our box i dorm packet thatthe sas of fhe gas within it lays constants "That is fo Neyo % sa a Dg Vat ae send FB an average velocity of the gus | where qis the average vector velocity of Jn the "box" in the direction, oi the | the gas in the box,” 9” is the average space average of the instantaneous den- | density of the gas in the box, and sity of the gas inthe box, and MC = p'V | AF = p'V isthe total mass of the gas in fi the otal mess ofthe gas in the bor, | the box, D/Dt ia not a simple partial erivative but represents the total rate of the change of the velocity of the par- ticular bit of gas in the box regardless of its position, Pa 29 ee) where tg and gar the component She set paride vest hanes in density ofthe gas due | "iF te veto prt vlsty nema nd wove isemal enough, then | enough, the stee! change of womestam stares danty 9h appro | fe pres n the ean be apa tnsnly ual tbe average dency, | indy the ate of change of omens Thee, Samat dned pointy Dq/DP = 9/95 and the instantaneous" ewity. pea be throsinated bythe average dealy po thn, ae “at p e 2 (2a) ‘The approximations just given are generally acceptable provided the sound pressure levels being considered are below about 110 db re 0.0002 microbar. Levels above 110 db are so large as to create hearing dis- comfort in many individuals, as we shall see in Chap. 13 at the end of this book, The Gas Law. If we assume an ideal gus, the Charles-Boyle gas law applies to the box. It is Py =RT es) where P is the total pressure in the box, V is the volume equal to Ar By 42, T is the absolute temperature in degrees Kelvin, and R is a ‘constant for the gas whose magnitude is dependent upon the mass of gas, Pact 11 THE WAVE EQUATION 0 chosen. Using this equation, we can find a relation between the sound pressure (excess pressure) and an incremental change in V for our box. Before we can establish this relation, however, we must know how the temperature 7’ varies with changes in P and V and, in particular, whether the phenomenon is adiabatic or isothermal ‘At audible frequencies the wavelength of a sound is long compared with the spacing between air molecules. For example, at 1000 eps, the wavelength 4 equals 0.34 m, as compared with an intermolecular spacing of 10° m, Now, whenever a portion of any gas is compressed rapidly, its temperature rises, and, conversely, when it is expanded rapidly, its temperature drops. At any one point in an alternating sound field, therefore, the temperature rises and falls relative to the ambient, tem- perature. This variation oceurs at the same frequency as that of the sound wave and is in phase with the sound pressure, ‘Let us assume, for the moment, that the sound wave has only one fre- quency. At points separated by one-half wavelength, the pressure and the temperature Quctuations will be 180° out of phase with each other. Now the question arises, Is there sufficient time during one-half an alternation in the temperature for an exchange of heat to take place between these two points of maximally different temperatures? Tt has been established? that under normal atmospheric conditions the speed of trave) of a thermal diffusion wave at 1000 eps is about 0.5 1n/se ‘and at 10,000 epsit is about 1.5 m/see. The time for one-half an alterna- ‘ion of 1000 eps is 0.0005 sec. In this time, the thermal wave travels a distance of only 0.00025 m. This number is very small compared with one-half wavelength (0:17 m) at 1000 eps. At 10,000 eps the heat travels, 7.5 X 10-*m, which is a small distance compared with a half wavelength (1.7 X 10? m), It appears safe for us to conclude, therefore, that there is negligible heat exchange in the wave in the audible frequency range. Gaseous compressions and expansions of this type are said to be adiabatic, For adiabatic expansions, the relation between the total pressure and the volume is known to bet PV? = constant 26) where + is the ratio of the specific heat of the gas at constant pressure to the specifie heat at constant volume for the gas. This equation is 130 4 mass of gas i choren so that ite weight in grams is equal to ite molecular weight Genowa to chemists as the gram-molecular weight, or the mole, then the Yolume of thia mace st O°C and 0.76 m Hg is the sue for all gates and equals 0.02242 m?. ‘Then R= 8314 watt-see per degree centigrade per gram-moleculae Weight, If the mass of gas chosen ie times its molecular weight, thea e = 8.314 ‘Soe LL, Beranck, "Acoustic Measurements,” p. 49, John Wiley & Sous, Tne, Now York, 1940 "MW. Zemaasky, “leat and "Thermodynamics," 24 ed, pp. 104-114, MoGran- Hil Book Company, Inc, New York, 1,EQUATION AND SOLUTIONS — (Chap. 2 20 THE WAVE ‘obtained from the gas law in the form of Eq. (2.5), assuming adiabatic conditions. For air, hydrogen, nitrogen, and oxygen, é¢., gases with diatomic molecules, yal Expressing Eq, (2.6) in differential form, we have av an Let P=Petp V=Vetr (28) where Ps and Vs are the undisturbed pressure and volume, respectively, fand p and r are the incremental pressure and volume, respectively, owing ry ay vagesioe as ee opiet Beara oo _ o Fis, 22. Change in volume of the box with change in position. From (a) snd (6) I acen that the inesomentel change im volume of the box equals + = (96,/9¢) S to the presence of the sound wave. Then, to the same approximation as that made preceding Eq, (2.4) and because p « Po and 1 Vo, Ra-F 29) The time derivative of this equation gives Ag-72 ax ‘The Continuity Equation. The continuity equation is s mathematical expression stating that the total mass of gas in a deformable “box” must remain constant. Because of this law of conservation of mass, we are able to write a unique relation between the time rate of change of the ineremental velocities at the surfaces of the box. aot 11 THE WAVE EQUATION a Ea See a | et ibae ‘The diflerence of the two qvantitios above multiplied by the area ay 3z ives | the increment in volume + ot, Warsyar ane | Yeave eam) ved a1 Differentiating with respect to time | Diflrentiating with respict to. time veld, yields, 132) | a Fa vodivg 2.30) Where aj the instantincons particle | where it the instantaneous particle velocity velocity The Wave Equation in Rectangular Coordinates One-dimensionol Derivation Threedimensional Devnation "The one-dimensional wave equation ig | The three-dimensionsl wave equation obtained by combining the equation of | ix obtained by combining the equation of ‘mation (24a), the gus las (2.10), and the | motion (2-40), the gas law (2.10), andthe continuity equation (2.134), Combing | eontinuty equation (2.125). Combine: tom of (210) and (2.14) gives tion of (210) and (219) gives Dierotte (1 with pect o | Dierntate (2.4) wih rape to wre eet) a Piderevte a) with apes to | Ta the degen of ech sd of -% ace | 2 . ~ di gat ~ mar 220822 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 One-dimensional Derivation ‘hveeimensionat Derivation Replacing the div (grad p) by Wp, wo get vip = podiv ¥ tp = pediv 8 (a7) interchangesbilty of the 2 | where vis the operator called the Lapla- and £ derivatives, and combining (2.152) | cian. Combining (2.199) and (2.17), we tnd (2162), we get ee ~ 52% ei v essay Let us, by definition set, we eat (2.19 oe (2.19) ‘We shall see later that cis the speed of propagation of the sound wave in the medium, We obtain the onedimensional wave | We obtain the three-dimensional wave cenwation caution re) vende aay Jn rectangle coordinates vip 2 4 OF 4 OP pe ME MP any We could als ave sited p and | We sould ao have eliminated p and retained in whih ease we would have | rsttnedq, im whith ease we would have 533 (222%) oy Law where 4g = grad (div q) when there is a (2220) | Co rotation in the medium, Equations (2.20) and (2.22) apply to sound waves of “small” magni- ‘tude propagating in a source-free, homogeneous, isotropie, frictionless gas at rest, The Wave Equation in Spherical Coordinates, ‘The one-dimensional ‘wave equations derived above are for plane-wave propagation along one dimension of a rectangular coordinate system. In an anechoic (echo- free) chamber or in free space, we frequently wish to express mathe- ‘matically the radiation of sound from a spherical (nondireetional) source of sound. In this ease, the sound wave will expand as it travels away. from the source, and the wave front alvrays will he a spherical surface. To apply the wave equation to spherical waves, we must replace the ‘operators on the left side of Eqs. (2.20) and (2.22) by operators appro priate to spherical coordinates, Assuming equal radiation in all directions, the wave equation in one= dimensional spherical coordinates is ap oe (223) Pert IV] SOLUTIONS OF THE WAVE RQUATION 23 Simple differentiation will show that (2.23) can also be written apr) _ 1 HG) ar Bae Ce) Itis interesting to note that this equation has exactly the same form as Eq (2.204). Hence, the same formal solution will apply to either equa tion except that the dependent variable is p(z,¢) in one case and pr(r,t) in the other case. ‘Example 21. In the steady stat, that ia, 2u/2¢ = jou, determine matheratially hhow the sound pressure in a plane progreetive sound wave fone ‘ould be determined fom ineasurement of partsle veloeity alone Solution. From Bq. (2-4a) we find in the steady state that where p and ware now rms Lively,” Written in diferent of the sound pressure and particle velocity, respme= form, 1 the particle velocity is 1 em /e0, wit 1000 radians/ste, and ais 0 em, then {8p = =30.005 X 1000 1.18 x a0 31059 newton mn? ‘We shall have an opportunity ia Chap, 6 of this test toate @ practical application ‘of these equations to the mesturement of particle vlocty hy a velocity microphone pant IV Solutions of the Wave Equation 2.3. General Solutions of the One-dimensional Wave Equation, ‘The ‘one-dimensional wave equation was derived with either sound pressure or particle velocity as the dependent variable. Particle displacement, or the variational density, may also be used as the dependent variable. ‘This can be seen from Eqs. (2.4a) and (2.18a) and the conservation of mass, which requires that the product of the density and the volume of a small box of gas remain constant. That is, 2'V = poVo = constant, (2.25) and so pave = —Vae (2.26) Let p= pte (2.27) where pis the incremental change in density. Then, approximately, from Eqs. (2.8) and (2.26), per = —Vep (2.28)a THE WAVE EQUATION AND SOLUTIONS —[Chep.2 Differentiating, ar _ Vode ee a asetl so that, from Eq. (2.132), 2 au et nese (2.29) Also, we know that the particle velocity is the time rate of change of the particle displacement. watt e220) Inspection of Eqs. (24a), (2.180), (2.29), and (2.30) shows that the pressure, particle velocity, particle displacement, and variational density are related to each other by derivatives and integrals in space and time. ‘These operations performed on the wave equation do not change the form of the solution, as we shall see shortly. Since the form of the solution is not changed, the same wave equation may be used for determining density, displacement, or particle velocity a well as sound pressure by substituting p, or &, or u for p in Eq. (2.20a) o p, &, oF q for p in (2.200), assuming, of course, that there is no rotation in the medium, General Solution. With pressure as the dependent vatiable, the wave ‘equuataon 38 ap _ lap Bet ae (2a) ‘The gencral solution to this equation is a sum of two terms, p-t-an(+3) 82) where fi and fs are arbitrary functions, We assume only that they have continuous derivatives of the first and second order. Note that hecause tand x occur together, the first derivatives with respect to z and t are exactly the same except for a factor of +e. ‘The ratio z/c must have the dimensions of time, so that e is a speed. From ¢ = yPe/pe (Eq. (2.19)] we find that 108\" 18, in airat an ambient pressure of 10° newtons/im? and at 22°C. ‘This quan- lity is nearly the same as the experimentally determined value of the speed of sound 344.8 [see Eq. (1.8)], 80 that. we recognize c as the speed at which a sound wave is propagated through the ait, From the general solution to the wave equation ‘observe two very important facts: ven in Eq, (2.82) we Pat IV] SOLUTIONS OF THE WAVE EQUATION 25 1. The sound pressure at any point x in space can be separated into ‘two components: an outgoing wave, f(t — z/c), and a backward-traveling, wave, fall + 2/0) 2, Regardless of the shape of the outward-going wave (or of the back- ward-traveling wave), it is propagated without change of shape. To show this, let us assume that, at ¢ = f,, the sound pressure at x — 0 is f(t). At a time ¢ = ty + t the sound wave will have traveled a dis- tance x equal to tye m. At this new time the sound pressure is equal to = fills + te ~ bec/e) = full). In other words the sound pressure has ropagated without change. ‘The same argument can be made for the backward-traveling wave which goes in the —z direction. It must be understood that inherent in Eqs. (2.31) and (2.92) are two assumptions. First, the wave is a plane wave, i, it does not expand laterally. ‘Thus the sound pressure is not a function of the y and coordinates but is a function of distance only along the x coordinate. Socond, it is assumed that there are no losses of dispersion (scattering of the wave by turbulence or temperature gradients, et.) in the air, so that. the wave does not lose energy as it is propagated. Dissipative and dis- persive cases are not treated analytically in this book, but are discussed briefly in Chaps. 10 and 11 Steady-state Solution. In nearly all the studice that we make in this ‘text we are concerned with the steady state. As is well known from the theory of Fourier series, a steady-state wave can be represented by a Jinear summation of sine-wave functions, each of which is of the form VO = V2 orl cos (st + 6 (2.330, For example, if ys is sound pressure, we write 2 = Spt) = Y V2 [pel 08 (ot + 0.) (2.380 where w, = 2rf,;f, = frequency of vibration of the »th component of the wave; 9, is the phase angle of it; and /2 |¢| (or |p") is the peak magnitude of the component. Because the wave is propagated without change of shape, we need consider, in the steady state, only those solutions to the wave equation for which the time dependence ateach point in space Js sinusoidal and which have the same angular frequencies w, as the source Borrowing from electsical-cieuit theory, we represent & sinusoidal funtion with a frequency w hy the real part of a complex exponential function, Thus, ata fixed point in space 2, we have the sound pressure, YG) = VIRe lolze"] (2.340) PUz,t) = V2 Re [plz )e (2.340)26 INE WAVE EQUATION AND SOLUTIONS — (Chap. 2 where p(x) is a complex function (fe, it hes a real and an imaginary part) that gives the dependence of p on z. The product of +/2 times the magnitude of p(z) is the peak value of the sinusoidal sound pressure function at z, ‘The phuse angle of p(z) is the phase shift measured from some reference position, Generally we omit writing Re although it always must be remembered that the real part must be taken when using the final expression for the sound pressure, In the steady state, there- fore, we may replace fs and fs of Eq. (2.82) by a sum of functions each having a particular angular driving frequency Ww, s0 that PD = Y pales) = Y VERE lipyremore + promenjeee] (2.35) ‘The part of Eq. (2.85) within the brackets is the same as that within the brackets of Eq. (2.34). ‘The factor +/2 is introduced s0 that later p.* and p_' may represent complex rms functions averaged in the time dimension. ‘The + and ~ subscripts indicate the forward and backward traveling waves respectively. tis apparent that the first term of Hq. (2.85) represents an outward- traveling wave whose rms magnitude |p,*| does not change with time { or position z. A similar statement may be made for the second term, which is the backward-traveling wave. It is customary in texts on acousties to define a wave number k, ee kat. Blo (2.36) Also, let us drop Re and the subscript » for convenience. Any one term of Eq. (2.35), with these changes, becomes PEED = V2 [Dla)er] = V2 (Dyer + peer) (2.37) Similarly, the solution to Eq. (2.22a), assuming steady-state conditions, ula) = VF (get + u_eitatey (2.38) tis understood that the real part of Eqs. (2.37) and (2.38) will be used in the final answer. The complex magnitudes of p, and p-or wy and ware determined from the boundary conditions ‘The complex rms pressure and particle velocity are found directly from Bqs. (2.37) and (2.38) by eanceling -/Ze™ from the right-hand sides, When the remaining fonetion is converted into magnitude and phase angle, the magnitude ts the quantity that would be indicated by a rms sound pressure meter. Note, however, that when we take the real part of P(e.) 0F w(t), the quantity +/2 e must bein the equation if the proper values for the instantaneous pressure and particle velocity are to be obtained Pat lV] SOLUTIONS OF THE WAVE EQUATION 27 Brample 22. Assume tha for the steady state, at x point x = 0, the sound pres rare in a onedimensional outward-raveing wave has the recurrent form shown by the dotted curve in the sketch below. This wave form i ven by the real part of the equation P10.) = VF (er 4 2 (e) What are the partite velocity and the particle displacement «function of time st z= 5m? () What ate the rms valies of these two quantities? (c) Are the ‘ins values dependent upon 2? ° 3005 ‘oor Tas seconds, Solution. a. We have for the solution of the wave equation giving both = and ¢ (sco Eq. (2.371 nia = Va Prom Ba 2s) wee hat = ch omen wie) = jh wat = 3 wed And from Hq, (280) we have PB (4. sansa 4 2B grimuerae Sow a8" ies Aba = Sm, 2/2 = 5/348 = 00145 see ata tat) = 23 (aesoneosian 4 detmaresieny sod tt = 2 (igen sto etme) ‘aking the apart ofthe two preci equation size) = 2 co (ots — ut) +200 cts 2.9 ean = 23 phgnin ou 0.) + 52 sn ns —270)] Note hat enh rm nthe parte dplcemen i on of in pase withthe velocity and Use the wave shape is different. As might be expected, differen tistion ‘nphasioes the28 THE WAVE EQUATION AND SOLUTIONS —(Chap.2 ‘Thoee equations are plotted below: q q 2. The rms magnitude of a sine wave is equal to ite peak value divided by VB. ‘This may be verified by squaring the cine wave and Gnding the average value over one cyele and then taking the square root of the rat. If two sine waves of diferent square root of the sums of the squares of the individual rms magoitudes, 99 that iby VEER ~ 0011 m/tc0 tn de VGR) + Ga) <8 x10 Thoms values w and é, are independent of x fora plane progressive sound wave, 24. Solution of Wave Equation for Air in a Rigidly Closed Tube. For this example of wave propagation, we shall consider a hollow cylindrical Devin ston Holiow yd tube fig emiston 7 = t t tele re, 23. Rigidly terminated tube with rigid side walls. ‘The velocity at = value of 72 ny com al 9 /ce has tube, closed at one end by a rigid wall and at the other end by a flat vibrating piston (soe Fig. 2.3). ‘The angular frequency of vibration of PatiVj SOLUTIONS OF TH WAVE EQUATION 29 the piston is w, and its rms velocity is ue. We shall assume that the diameter of the tube is sufficiently small so that the waves travel down the tube with plane wave fronts. In order that this be true, the ratio of the wavelength of the sound wave to the diameter of the tube must be ‘greater than about 6. Fo. 24, Porton of the tuhe showing the diection and magnitude of movement ofthe ‘ie particles sv fnction of LAL position , the particle volocity and displace- ent ere maxituin, At position B, they are zero. Particle Velocity. ‘The form of solution we shall select is Eq. (2.38). Set tis equal to the rims velocity of the vibrating piston at x = 0, and set Lequal to the length of the tube, The boundary conditions are Atz = 0, u(0,) = V2 we, so that aye Ata = thu = 0, s0 that ef uel = 0 239) Remember that sny = Se Hence = xahsin Al (2.40) and wet ; wn ean) whieh gives ws lad) = VB te LS ny sin k(t — 2} un wk 49) Note that the «/% and the time exponential have been left out of Eq (243) so that both ve and w are complex rms quantities averaged ove time, Refer to Fig, 24. If the length Land the frequency are held constant30 THE WAVE EQUATION AND SOLUTIONS —{Chap. 2 the particle velocity will vary from a value of zero at x = 1 to a maxi mum at |~ z= 2/4, that is, at J — x equal to one-fourth wavelength. In the entire length of the tube the particle velocity varies according to a sine function. Between the end of the tube and the d/4 point, the oscillatory motions are in phase. In other words, there is no pro- gressive phase shift with z. This type of wave is called a standing wave because, in the equation, z and cf do not occur as a difference or & sum in the argument of the exponential funetion. Hence the wave is not, propagated In the region between 1 — z= /4 and lz =/2, the particle velocity still has the same phase except that its amplitude decreases sinusoidally. At I— x =/2, the particle velocity is zero, In the region between | — x = 2/2 and 1 ~ x = the particle velocity varies with 2 according to a sine function, but the particles move 180° out of phase with those between 0 and 2/2. This is seen from Eq. (2.43), wherein the sines of arguments greater than © are negative. If we fix our position at some particular value of x and if Lis held con- stant, then, as we vary frequency, both the numerator and denominator of Eq. (2.43) will vary. When A is some multiple of x, the particle velocity will become very large, except at x= 0 or at points where E( ~ x) isa multiple of x, that is, at points where — z equals multiples of X/2. ‘Then for kl = nz (2.44) Equation (2.43) would indicate an infinite rms velocity under this condi- tion. In reality, the presence of some dissipation in the tube, which was neglected in the derivation of the wave equation, will keep the particle velocity finite, though large. ‘The rms particle velocity 1 will he zero at those parts of the tube where k( — 2) = nefandnisan integer or zero. That is, (2.45) In other words, there will be planes of zero particle velocity at points along the length of the tube whenever 1 is greater than X/2. Some examples of the particle velocity for 1 slightly greater than various multiples of 4/2 are shown in Fig. 2.5. Two things in particular are apparent from inspection of these graphs. First, the quantity n deter mines the approximate number of half wavelengths that exist between the two ends of the tube. Secondly, for a fixed wu, the maximum velocity 1 For the type of source we have assumed and no diaspation, this ease breaks down for Hl = ne Pat IV] SOLUTIONS OF THE WAVE EQUATION 31 of the wave in the tube will depend on which part of the sine wave falls atx = 0. For example, if 1 — nd/2 = /4, the maximum amplitude in the tube will be the same as that at the piston. If] — nh/2 is very near zero, the maximum velocity in the tube will become very large. Let us choose a frequency such that n = 2 as shown, Two factors determine the amplitude of the sine function in the tube. First, atz = 0 the sine eurve must pass through the point we. Second, atx = I the sine curve must pass through zero. It is obvious that one and only one sine wave meeting these conditions can be drawn so that the amplitude is determined, Similarly, we could have chosen a frequency such that a2) — FH aut x0 a Fic. 25. Variation ofthe particle velocity (x for =0, 482 function ofthe distance along the tue of Fig 2 for thee frequencies, ce, forthree wavelengths. ALx = 0, therms particle velocity the particle velocity is zero. The period T= 1/f n « 2, but where the length of the tube is slightly less than two half wave- lengths, Tf this ease had been asked for, the sine wave would have started off with a positive instead of a negative slope at x = 0. Sound Pressure. ‘The sound pressure in the tube may be found from the velocity with the aid of the equation of motion (Eq. (2.4a)], whieh, in the steady state, becomes p= ~iepfude (2.46) ‘The constant of integration in Eq, (2.46), resulting from the integration of Eq, (24a), must be independent of 2, because we integrated with32 TH WAVE EQUATION AND SOLUTIONS — [Chep. 2 respect to x. ‘The constant then represents an increment to the ambient. pressure of the entire medium through which the wave is passing. Such An increment does not exist in our tube, so that in Eq, (2.46) we have set the constant of integration equal to zero. Integration of Eq. (2.46), after we have replaced u by its value from Eq. (2.42), yields pe) = “Ine VE na HL 2 ean) P= ~ineuy 2H 2) 4s) Note that the /% and the time exponential have been left out of Eq. 2.48) so that both p and tz are complex rms quantities averaged in time. ‘The rms pressure p will be zero at those points of the tube where i(U = 2) = nx + 7/2, where n is an integer or zero, (2.49) The pressure will equal zero at, one or more planes in the tube whenever 1 s greater than )/4. Some examples are shown in Fig. 2.6. Here again, quantity n fs equal to approximate number of half wavelengths in tube Refer once more to Fig. 2.5 which is drawn for ¢= 0. ‘The instan- aneous particle velocity is at its maximum (as a function of time). By comparison, in Fig. 2.6.at ¢ = 0, the instantaneous sound pressure is zero. At a later time = 7/4 = 14f, the instantaneous particle velocity has vecome zero and the instantancous sound pressure has reached its m: num. Equations (2.42) and (2.47) say that whenever k(0 ~ 2) is a small number the sound pressure lage by one-fourth period behind the yarticle velocity. At some other places in the tube, for example when 1 ~ 2) lies between d/4 and 2/2, the sound pressure leads the particle relocity by one-fourth period. ‘To see the relation between p and u more clearly, refer to Figs. 2.5 and 2.6, for the case of n = 2. In Fig. 2.5, the particle motion is to the right whenever w is positive and to the left when it is negative. Hence, at the 2/2 point the particles on either side are moving toward each other, 80 hat one-fourth period later the sound pressure will have built up to a naximum, as can be seen from Fig. 2.6, At (I — z) = 4/2, the particles wre moving apart, so that the pressure is dropping to below barometric ws ean be seen from Fig. 2.6 Figures 2.5 and 2.6 also reveal that, wherever along the tube the magni- ude of the velocity is zero, the magnitude of the pressure is a maximun und viee versa. Hence, for maximum pressure, Eq. (2.45) applies. Specific Acoustic Impedance. It still remains for us to solve for the ipevifie acoustic impedance Z, at any plane x in the tube. Taking the Part IV] SOLUTIONS OF THE WAVE EQUATION 33 ratio of Eq. (2.48) to Bq. (2.43) yields = —iowe cot Ml! = 5X, mks rayls (2.50) where X, is the reactance. Where we have set 1- (21) ‘That is, U is the astance between any plane x in Fig. 2.3 and the end -———— wa ——___ ry] oT ft aan ME rel Y : omer | POEUIET, 0 sl Fa, 2.6, Variation of the sound pressure p(s) as «function of the distance along the tube for three frequencies, i.e, for three wavelengths, Atz = 0, the rms particle Velocity is we and at x =f, itis zero, ‘The period 7 equals 1/f of the tube at L The —j indicates that at low frequencies where cot kl! & 1/k’ the particle velocity leads the pressure in time by 90° and the reactance X,is negative. Atall frequencies the impedance is reactive and either leads or lags the pressure by exaetly 90° depending, respec tively, on whether X, is negative or positive. ‘The reactance X, variuM THE WAVE EQUATION AND SOLUTIONS — (Chop. 2 us shown in Hig. 2.7, If the value of kl’ is small, we may approximate the cotangent by the first two terms of a series 1a cot ht = yp — (2.52) ‘This approximation is valid whenever the product of frequency times the distance from the rigid end of the tube to the point of measurement is very small. If the second term is very small, then it may be neglected with respect to the frst, 2 ft fo ne vty i : "hI ' 1 | ! ! ! 1 | ! ! } | b I | | +e t 0 i i wl A 1 I | fo ry eM 0 Fro, 27. Tho specific acoustic reactance (Prne/tem) along the tube of Fig. 2.3 for a particular frequency, sc, » particular wavelength where 3(4/2) isa litle less than the fue length For this ease, the numberof zeros is 3, and the number of poles is Let us see how small the ratio of the distance to the wavelength \ must be if the second term of Eq. (2.52) is to be 3 per cent or less of the first term. That is, let us solve for I'/A from 2al" d 3) By 5 0.03 (2.53) which gives us ve : 5 = 0.05 (2.54) In other words, if cot ki’ is to be replaced within an accuraey of 3 per cent by the first term of its series expansion, I’ must be less than one-twentieth wavelength in magnitude, Assuming I" < 4/20, Eq. (2.50) becomes 1 1 pot _ re ae 2m ikem —i5h = a0 7 pat = jo, HB rays (2.55) Heneo, the specific acoustic impedance of a short length of tube can be Part IV} SOLUTIONS OF THE WAVE EQUATION 35 represented as a ‘capacitance” elle specific acoustic compliance, of mes; nitude C, = U'/poc*. Note also that C, = '/7Ps, because of Eq. (2.19). The acoustic impedance is of the same type, except that an area factor sppears so that u-2-— pet Bu ~ jo(V Joo) ~ jul, where V = I'S is the volume and S is the area of cross section of the tube. C, is called the acoustic compliance and equals V/pec?. Note also that C4 = V/yPo, from Eq. (2.19). ‘mks acoustic ohms (2.56) Example 2.8. A cylindrical tube isto be used in an acoustic device as an impedance clement. (a) ‘The impedance desired is thet of a compliance. What length should it have to yield w reactance of 1.4 X 10% mks ray xt an angular frequency of 1000 radiana/sec? (6) What is the relative magnitude of the fir. and second terms of Bq. (252) for this case? Solution. ‘The reactance of auch a tbe is P @ en 14x10 he Henee, = 0.1m, o e renee, the second term is about 3 per eent ofthe fist term. 2.6, Freely Traveling Plane Wave. Sound Pressure. If the rigid termination of Fig. 2.3 is replaced by a perfectly absorbing termination, a backward-traveling wave will not occur. Hence, Eq. (2.37) becomes Plat) = VE pyemern 7) where p. is the complex rms magnitude of the wave. This equation also applies to a plane wave traveling in free space. Particle Velocity. From Eq, (2.4a) in the steady state, we have 1 op un-- 12 58) Japa 3 58) Hence, V2 Pe pcre = PEL u(t) = rr a pot (2.59) ‘The particle velocity and the sound pressure are in phase. This is mathe- matical proof of the statement made in connection with the qualitative discussion of the wave propagated from a vibrating wall in Chap. 1 and Fig 11 Specific Acoustic Impedance. The specific acoustic impedance is 2, =F = pe mks rayls (2.60)36 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 ‘This equation says that in a plane freely traveling wave the specific acoustic impedance is purely resistive and is equal to the product of the average density of the gas and the speed of sound, This particular ‘quantity is generally called the characteristic impedance of the gas because its magnitude depends on the properties of the gas alone. It is a quantity that is analogous to the surge impedance of an infinite electrical line. For air at 22°C and a barometric pressure of 10° newtons/m?, its mag- nitude is 407 mks rayls. 2.6. Freely Traveling Spherical Wave. Sound Pressure. A solution to the spherical wave equation (2.24) is vi ec + acm) a 261) where A, is the magnitude of the rms sound pressure in the outgoing wave at unit distance from the center of the ephere and A_ is the same for the reflected wave. If there are no reffecting surfaces in the medium, only the first term of this equation is needed, i.e., Pert) Pen) = VEAL pu 0) Varticle Velocity. With the aid of Eq. (2.48), solve for the particle velocity in the r direction, ure) VE oo( 43h) VEL (14 be 203) Specific Acoustic Impedance. ‘The specific acoustic impedance is found from Eq. (2.62) divided by Eq. (2.63), kr peckr BR Sore 7 (90° = tan mks (2.64) oe ~ i kr rayls (2.64) Plots of the magnitude and phase angle of the impedance as a function of kr are given in Figs. 2.8.and 2.9. The real and imaginary parts, R, and X,, are plotted in Fig. 2.10. Por large values of kr, that is, for large distances or for high frequencies, this equation becomes, approximately, Z.% poe mks rayls (2.65) The impedance here is nearly purely resistive and approximately equal to the characteristic impedance for a plane freely traveling wave. In other words, the specific acoustic impedance a large distance from a spherical source in free space is nearly equal to that in a tube in which no reflections pcour from the end opposite the source. Pat IV} SOLUTIONS OF va = 1 ee 14 ete oe | See tncretmceste, CHEV | gt REE a - £ a gL | od | okt oe | ry t coe * ‘ood “oa 05 10 5 10 20 (erat feat r/R) Fo, 28. Plot of the magnitude of the specific acousticsimpedance ratio !Z/ow in a spherical feeely traveling wave as a funetion of kr, where is the wave number equal tow/ror 2e/sand ris the distance from the center of the spherical source. [21] isthe ‘magnitude of ratio of pressure to particle velocity in a spherical free traveling weve, sand pr ie the characteristic impedance ef wit so Ao * LUI iz 7} + itt 4 « a . t : 5° ani ol | lt" ercraegeetis NN £40) particle velocity 3 Bisiecaate Ey : {ty | i | Pa 9 - ih — + . ee oo * “ood “ox 08° "10, 5° 10 = Uirytorfee(200) Fie, 29. Mot ofthe phate angle, n degre ofthe pee acousticmpedance rt Zulosc in a spherical wave a3 8 function of kr, where kis the wave number equal to wfeor 2r/a, und ris the distance from the centr of the spherical source.38 THE WAVE EQUATION AND SOLUTIONS — [Chap.2 Equations (2.62) and (2.63) are significant hecause they reveal the difference between the responses of a microphone sensitive Lo pressure and a microphone sensitive to particle velocity as the microphones are brought close to a small spherical source of sound at low frequencies. AS ris made smaller, the output of the pressure-responsive microphone will double for cach halving of the distance between the microphone 29) 1 0s| 03] 02 oa 005} 03] 0.02] Po 0.005 0.008] 002] Irapedance con ‘aalogy 0.0005} 0.0008] 0.0002] 00014 = ire oor * G05 “02 03 ‘10 510 20 be Fio, 2.10. Teal and imaginary parts of the normalized spociie acoustic impestance Z.Joue of the air loul ons plating sphere of radius r loeated in free spac. Fe {quency is plotted on a norinalized seale where kr = 2xjr/e = 2er/d.” Note als that the ordinate is equal to Zu/owS, where Zw i the mechaninl impedance, nil to ZaS/no, whore Zs the acoustic impeance. The quantity 8 is the aren for which the impedance is being determined, and py ig the characteristic impesanc: of the edie center of the spherical source. Expressed in deeibols, the output increases Gdb for cach halving of distance, For the velocity-responsive micro- phone, the output variation is not so simple. Only at sufficiently large distances (kr* >> 1) does the output increase 6 dh for each halving of distance. For shorter distances the second term inside the parentheses on the right-hand side of Eq. (2.63) becomes large, and the magnitude of w increases at a rate exceeding +6 db for each halving of distance. For Pat VJ SOLUTIONS OF THE WAVE EQUATION very short distances (Kr? < 1), the rate of increase of w approaches : limit of +12 db for each halving of distance. It is for this reason tha the voice of a radio crooner sounds “bassy”” when he sings very neat tot velocity-sensitive microphone which was designed to have ite best response when located a large distance from the source of sound ‘Another significant thing is to be learned from Eq, (2.64). At low fre quencies it is very difficult to radiate sound energy from a small loud speaker. A small loudspeaker may be likened to a pulsating balloon o! some small radius r. The specific acoustic impedance Z, of the air pre sented to each square centimeter of the balloon is given by Eq. (2.64) and Fig. 2.10. At low frequencies, the impedance becomes nearly purely reactive, and the resistance becomes very, very small, Hence, the powe! radiated by a small loudspeaker becomes very small. At high fre. quencies, kr > 2, the impedance Z, becomes nearly purely resistive and has its maximum value of pc, s0 that the power radiated for a given value of p, reaches its maximum. ‘The important steady-state relations derived in this chapter are sum. marized in Table 2.1. ‘TABLE 21. General and Steady-state Relations for Small-signal Sound Propa. gation in Gases ame Genel equation | Steuyatate euatin Wave equations 718 Joy ya 22) X= Bar apr) _ 1 apr) oF 7a ae Equation of motion oa Bm Mfume Jude grad p= Jao Diplacement Incremental density Incremental temperature40 THE WAVE EQUATION AND SO TIONS — [Chap. 2 parr V Energy Density and Intensity 2.7, Bnergy Density. Energy density is an important concept in acoustics because, in dealing with sound in enclosures, it is necessary to study the flow of energy from a source to all parts of the room. The energy density, ie., watt-seconds per unit volume, is greater near the source than farther away and is the variable that appears in the equations describing the acoustical conditions. On the other hand, the ear and most sound-level meters respond to rms sound pressure. We need to ascertain, therefore, the relation between energy density and sound pressure in sound fields. ‘The energy density associated with the small “box” of gas at any par~ ticular instant is the sum of the kinetic and potential energies per unit volume of the air particles in the box. ‘The kinetic energy density due to the excess pressure of the sound wave Drx is 1 Mut _1 Dax = 5 ye = yee (2.66) where w is the average instantaneous velocity of the air particles in the ox, pe is the average density, and 3f/V is the mass per unit volume. ‘The potential energy density due to the sound wave Dre may be found from the gas law. For very small changes in the volume of the box, we may write [see Eqs. (2.8)] (2.67) (2.9) and substitute the resulting expression for dz, the potential energy density becomes = pap lw Pa 25P, Des (2.08) When the sound pressure p is equal to zero, the potential energy due to the sound wave must be zero. The arbitrary constant of integration is therefore also equal to zero. ‘The total energy density due to the sound wave D = Dax + Des, oF Dw = 3 (on + 5) (2.69) ‘This equation is true at any instant at a given point in space 2.8, Energy Density in Plane Waves. Energy Density in a Plane Pree-progressive Wave. From Eqs. (2.57) and (2.59) we have seen that the pressure and particle velocity in a plane free-progressive (outgoing) Part VI ENERGY D PNSITY AND INTE 41 wave are equal to (zt) = Re V2 pret = /2 Ips cos [k(ct — 2) + 6] (2.70) V2 Me psc = V2 Pol Ee Feel cos h(ct ~ 2) + 6] (2.71) u(z,t) = Re where py = |pyle. ‘The instantaneous energy density for such a wave in the steady state is, from Lage. (2.68) and (2.19), equal to D(z) = |psl* DG pan) cost Ia(ct — 2) + 8] = Dpal? 1+ cos 2at ~ 2/e + Of) mc 2 WAL (1 + cos 2u(t — 2/0 + 6/4) @72) “This equation says that for a plane fre-progressive wave, at all times, the Kinetic and potential energy densities are equal at a given point in space but that they vary with position or with time sinusoidally from zero to twice their average value. ‘The situation here is different from that fora pendulum where the kinetic energy and the potential energy vary in opposite phase, &.c, one ie a maximum when the other iz a minimum Here, energy is being transported away from the source. Conversely, the pendulum is a conservative system. ‘When averaged over either a length of time equal tol = 1'/2 = 1/2f or a distance in space 2 ~ /2~ o/2f, we find the average energy density to be equal to Ipal? 2 13) al wattce/m 273) D, where |ps| is the magnitude of the rms value (in time) of the sound pres- sure measured at any point in the sound wave. Note also that pect = 7Ps as stated before. Inspection of Eq. (2.60) shows that we may let Pe pot where 1, is the rms value (in time) of the velocity at any point in the wave, Then, uy (274) Dara = |1sl¥00 (275) ‘Equations (2.73) and (2.75) give the relations among rms sound pressure, particle velocity, and energy density Energy Density in a Plane Standing Wave. From Eqs. (2.42) and (2.47) we have that piety = VE need EM cow — 2) aaa42 THE WAVE EQUATION AND SOLUTIONS — (Chap.2 where @ is the phase angle of us. 608 (ot + 0) sin k= 2 UE) = VE ug) SRE Sin BE 2) em In this case, the kinetic and potential energy are 90° out of time phase ‘The situation is analogous to that for a pendulum becuse in both eases the systems are conservative. ‘The instantaneous energy density for such a wave in the steady state is, from Eqs. (2.69) and (2.19), equal to 08 2(ut + @) cos 2(l — 2) T= cos 2E D(z) = |udlioe 2.78) ‘When averaged over either a length of time f equal to 7/2 or a distance in space equal to 4/2, we find the average energy densily to be equal to Dog = ltl, wateane/mt ex where us| is the magnitude of the rms velocity of the piston at 2 = 0. This equation shows that, for « constant value of jus, the average energy density varies from |uy/?p9/2 to infinity depending on the value of H = 2el/r A better way of representing the average energy density is in terms of the rms pressure. If, by definition, we let the rms value of the pressure be related to the mms velocity uy at x = O by the formula pi = one (280) we have (zt) = V2 [pil sin (wt + 6) cos k(t — 2) (281) Then Eq. (2.79) becomes = iat Dow = PO (282) Here, {pi| is the magnitude of the rms value (in time) of the mazimum value (in space) of the sound pressure. If we measure the rms value of the sound pressure in space by moving a microphone backward and for- ward over a wavelength and averaging the varying output in a rms rectifier, then [Poul = [psl/-V2 and Dig = Weal’ yattace/mt (283) where |pow| is the magnitude of the rms value of the sound pressure averaged in both space and time, Note that Eq. (2.83) is identical to By. (273) Part V1 ENERGY DENSITY AND INTENSI 3 Example 24. Calculate the average energy density in a plane free-progressive Sinusoidal sound wave with a maximum particle displacement of 0.01 em at a quency of 100 eps. Soluion, From Eq, (2.80) we find that the ras yrtile velovity 1 tga wine 80 2e x 100 x OL > one = 00 X 0.01 9.645 ma /sce VP x10 ‘The average energy density is given by Ka. (275), Dag = (0.0485)* X LB = 234 X 10-9 wattseo 2.9. Energy Density in a Spherical Free-progressive Wave. ‘The energy density in a spherical free-progressive wave can be shown to be equal to 81) where [p, is the magnitude of the rms value (in time) of the sound pres- sure at a point a distance r from the center of the spherical souree. If the produet of the distance r and the frequency is large (2k? >> 1), the average energy density is the same as for plane free-traveling oF standing wave, as can be seen from Eqs. (2.73) and (2.83). Near the sonree, however, the eneray density hecomes very large. ‘This occurs because the impedance {sce Eq. (2.64)] becomes largely reactive and the stored energy becomes high. 2.10. Sound Energy Flow—Intensity, Later in this text we make frequent reference to the flow of sound energy through an acoustie system, Because of the law of the conservation of energy, the total acoustie energy starting from a source must be completely accounted for in the system. At any part of an acoustic system, we should be able to state the amount of energy flowing through that part per unit time, and it should equal the power emanating from the source minus any intervening losses. In Part IT we defined intensity as the average time rate at which energy is flowing through unit area of the acoustie medium. In the mks system, the units of intensity are watts per square meter. The intensity is actually the product of the sound pressure times the in-phase component of the particle velocity. General Equation for Intensity. We can find the average intensity J 1 given direction at a given point in the medium by performing the operation T= Reptgcos ¢ (285) SL. B Kinder and A. TR. Frey, “Fundamentals of Acoustss,” pp, 167-168, John Wiley & Sons, Ine, 1980. 1 The average power supplied by an clrtrical generator to a circuit equals the voltage times the in-phase component of the current. "Tht power ean be shown to quel Re (E*1), where B and 7 are the complex ema voltage and curtont, respectively.a THE WAVE EQUATION AND SOLUTIONS — [Chap.2 where p* is the complex conjugatef of the rms sound pressure p, qis the complex rms particle velocity in the direction the wave is traveling, and @ is the angle between the direction of travel and the direction in which the intensity is being determined. ‘The symbol Re indicates that the real part of the product is to he taken. Intensily in a Plane Free-progressive Wave. For a plane free-progressive sound wave the intensity equals = Re pot Pt co T= Re pom Be é 2.80) Another way of looking at the question of intensity for & plane progres- sive wave is to say that all the energy contained in a column of gas equal in Jength to em must pass through unit area in 1 see. Hence, the inten- sity is T= Dang 005 4 287) So, regardless of whether the intensity is determined from (2.86) or (2.87), wwe get for & plane free-progressive wave that Pal oo g = 1 = BE cose Intensity im a Plane Standing Wave. In a plane standing wave the pressure and particle velocity are 90° out of phase in time [see Fqs. (2.76) and (2.77)] so that the real part of the product ptu is zero. Hence, for a plane standing wave, ius love cos (2.88) T=0 (289) Physically, tbis means that as much sound energy returns to the source as travels away from it. Intensity in a Spherical Free-progressive Wave. For a spherical progres cive wave, we got the pressure p from Eq. (2.62). By definition, let pe a em (2.00) Then, Plot) = V2 pact eo) ‘The quantity p, is equal to the complex rms pressure at any point a dis- tance 7 from the center of the source. Hence, the particle velocity u(r) at any point a distance r is Em (14 jbo asm HIE pis represented by [ple then pe is [ple Similarly, if p i represente hy e+ jy, then 9" i represented by px ~ Je ur. Ps Vi ENERGY DENSITY AND INTENSITY 45 or the complex rms particle vel we B15] Substitution of the sound pressure at, p,, and Eq, (2.98) into Bq, (2.85) vields ) a I= Rept iF Be cos & eo) ‘owe here, as before, ¢ is the angle between the direction of travel of the wave and the direction in which the intensity is being determined. ‘Wo can derive these results in a different way. Equation (2.08) states that, for kr large, p and u fora spherical wave are nearly in time phase and P(r) = pucu(r) a8 shown by Eq. (2.05). Hence, for kr large, we see fron. Eq, (2.88) that in a spherical wave for large distances I = |ur|?pec 008 6. ‘The total power at any radius r is equal to W = er! = 4r"|pl?/pee. Hence, for a spherical wave, Ww T=, forg=0 (2.95) By the law of conservation of energy, W is independent of r if there are uio losses in the gas so that the intensity varies inversely as the square of the distance r. From Eq. (2.90) we see also that: the square of the rms magnitude of the sound pressure at any point varies inversely with the square of the distance r. Hence, because the intensity / at any point varies similarly, it is directly proportional to the square of the sound pressure at that. point. ‘This result agrees with that shown in Eq. (2.94) Example 2.5. A spherical sound source is radiating sinusoidally into free space | watt of acoustic power at 1000 eps. Caleulate (a) intensity in the direction the wave js traveling; (6) sound pressure; (e) particle velocity; (d) phase angle between (0) and (¢); (2) energy density; and (f) sound preseure level at @ point 30 em from the ‘eonter of the source, (Assume 22°F and 0.751 m Hg.) Solution. a. ‘The intensity maybe found fo 2.95). >, The rins sound pressure eomes from Ig. (2:94). Ind = Vase = VORB XAT = 18.97 nowtone/m* ‘ The rms particle volocity is given by Bg. (2.98), be = (Or x 1000/344.8)(08) = 5.46 be VTE A 1897 VT + HR tel = Fe VUES 1897 VT BB aaz4 mine46 THE WAVE EQUATION AND SOLUTIONS — (Chap. 4. The phase angle @ between pe and ue may be found from Faq. (264) 0 = 90" — tant kr = 90" — 796" = 10.4" The energy density is given by Hq. (2.84) BEC + ah) = pata ( Pow pact AI * aunt) ~ TEx TO + +2162 X 10-4 wattsee/n* J. The sound pressure love is found from Hq. (1.18). oe 18.97 SPL = 20 log 5 OTT = 119.5 db re 2 X 10-4 nowton/m (re 2 X 10-4 mierobar) ‘This sound pressure level is about 15 db higher than the highest level tht is meas tured at 25 1t above a full aymphiony orchestra. In other words, 1 watl of acoustic power ervatcs a very high sound pressure level atl ft from the souree CHAPTER 3 ELECTRO-MECHANO-ACOUSTICAL CIRCUITS parr VI Mechanical Circuits SS 3.1, Introduction. ‘The subject of electro-mechano-acoustics (some. times called dynamical analogies) is the application of electrical-cireuit theory to the solution of mechanical and acoustical problems. In clas- sical mechanics, vibrational phenomena are represented entirely by differential equations. ‘This situation existed also early in the history of telephony and sadiv. As telephone and radio communication developed, it became obvious that a schematic representation of the elements and their interconnections was valuable. These schematic diagrams made it Possible for engineers to visualize the performance of a circuit without laboriously solving its equations. The performance of radio and tele- vision systems can be studied from a single sheet of paper when sich schematic diagrams are used. Such a study would have been hopelessly 4ifficult if only the equations of the system were available. There is another important advantage of a schematic diagram besides its usefulness in visualizing the system. Often one has a piece of equip- ment for which he desires the differential equations. ‘The schematic iagram may then be drawn from visual inspection of the equipment. Following this, the differential equations may be formed directly from the schematic diagrams. Most engineers are trained to follow this pro- cedure rather than to attempt to formulate the differential equations directly. Schematic diagrams have their simplest applications in circuits that contain Jumped elements, é¢., where the only independent variable is time. In distributed systems, which are common in acoustics, there may be as many as three space variables and a time variable. Here, a schematic diagram becomes more complicated to visualize than the differential equations, and the classical theory comes into its own again. There are many problems in acousties, however, in which the elements are Jumped and the schematic diagram may be used to good advantage. a48 ELECTRO-MECHANO-ACOUSTICAL CIRCUITS (Chap. 3 Four principal requirements are fulfilled by the methods used in this text to establish schematic representations for acoustic and mechanical devices. They are’ 1. The methods must permit the formation of schematic diagrams from visual inspection of devices 2. They must be capable of such manipulation as will make possible the combination of electrical, mechanical, and acoustical elements into one schematic diagram. 3. They must preserve the identity of each element in combined cir- cuits 60 that one can recognize immediately a force, voltage, mass, inductance, and so on. 4. They must use the familiar symbols and the rules of manipulation for electrical circuits Several methods that have been devised fulfill one or two of the above four requirements, but not all four. A purpose of this chapter is to present a new method for handling combined electrical, mechanical, and acoustic systems. It incorporates the good features of previous theories ‘and also fulfils the above four requirements. ‘The symbols used conform with those of earlier texts wherever possible.!“* 3.2. Physical and Mathematical Meanings of Circuit Blements. ‘The circuit elements we shall use in forming a schematie diagram are those of clectrical-circuit theory. These elements and their mathematical mean- ing are tabulated in Table 3.1 and should be learned at this time. There are generators of two types. ‘There are four types of circuit elements: resistance, capacitance, inductance, and transformation. There are three generic quantities: (a) the drop across the circuit element; (0) the flow through the circuit element; and (¢) the magnitude of the circuit element. Attention should be paid to the fact that the quantity ais not restricted to voltage ¢, nor b to electrical current i. In some problems a will repre- sent force f, or velocity wu, or pressure p, or volume velocity U. In those cases b will represent, respectively, velocity u, or force f, or volume °B, Gehlshoj, “lectromechanical and Hectroscoustieal Anslogies,” Academy of ‘Technical Sciences, Copenhagen, 1947 *F. A. Firestone, A Now Anelogy between Mechanieal and Hlectrcal Systems, J. Acoust. Soe. Amer, 4: 249-267 (1988); The Mobility Method of Computing the Vibration’ of Linear’ Mechanical and Acoustical Syatems: Mechanical-clectrical Analogies, J. Appl. Phyr., 8: 373-387 (1938). *H. F, Olson, “Dynamical Analogies” D, Van Nostrand Company, Inc, New York, 1988. “WP. Mason, Electrical and Mechanical Anslogies, Bell Syslem Tech, J., 20: 405-414 (1941). A. Bloch, Electro-mechanical Anslogies and Their Use for the Analysis of Mechani- cal and Bleetrosmechanieal Systems, J. Ina. Elec. Eng., 92: 157-169 (1945), {Among the four circuit elements, the first three are two-poles, This list is exhaus- tive, The transformation element isa four-pole, There are other loses fonr-poes Which one might have chosen in addition, e., the ideal gyrator, Part. VI) MECHANICAL CIRCUITS 49 velocity U, or pressure p. Similarly, the quantity ¢ might be any appropriate quantity such as mass, compliance, inductance, resistance, etc, The physical meaning of the circuit elements ¢ depends on the way in which the quantities a and b are chosen, with the restriction that ab thas the dimension of power in all eases, ‘The complete array of alterna- tives is shown in Table 3.2. ‘TABLE 3.1. Mathematical and Physical Significance of Symbols ] T ‘Mooring somtt | tame : | Constant-drop | The quantity a is independent of tor. ‘The arrow points to the positive terminal ofthe generator |Constant-ow |The quantity b is independent of | generator what is connected to the gen- erator. ‘The arrow points in the ireation of positive fow Resistancetype | a sement Sent fend f ba fom | ee — ee clement gst $ type element, be a 138 aa ‘An important idea to fix in your mind is that the mathematical opera~ tions associated with a given symbol are invariant. If the element is of the inductance type, for example, the drop @ across it is equal to the time derivative of the flow b through it multiplied by its sizec. Note that this rale is not always followed in electrical-cireuit theory because there con- ductance and resistance are often indiscriminately written beside the symbol for a resistance-type element, The invariant operations to be associated with each symbol are shown in columns 3 and 4 of Table 3.1.| Part VI) MECHANICAL CIRCUITS 51 CHANO-ACOU 3.3, Mechanical Circuits. Mechanical-circuit elements need not “kK always be represented by electrical symbols. Since one frequently draws " a mechanical circuit directly from inspection of the mechanical device, bis. ‘more obvious forms of mechanical elements are sometimes useful, at least until the student is thoroughly familiar with the analogous cireuit. We shall accordingly devise a set of ‘‘mechanieal” elements to be used as an introduction to the elements of Table 3.1 em Ma leney TABLE 3.3. Conversion from Mobility-type Analogy to Impedance-type Analogy, or Vice Versa Acoustical | ats = [MECHANICAL NULOGES {AOOUSTIAL MALDGIES Impedance analogy P u ovis pe [_trowtnce ype | Meity ype | tnpetonn Spe ca ee (a i Ec a ry ToS we PS BS att! on Ra Je= Me ‘and Acoustical Circuits Impedance ‘analogy f T ‘Mechanical Mesh aus (nobiy Boe) ech to was empress) {us 2 poe In electrical circuits, a voltage measurement is made by attaching the Jeads from a voltmeter across the two terminals of the element. Voltage is a quantity that we can measure without breaking into the circuit, ‘To measure electric current, however, we must break into the circuit because this quantity acts Ghrough the clement. In mechanical devices, on the other hand, we ean measure the velocity (or the displacement) without disturbing the machine by using « capacitive or inertially operated vibra~ tion pickup to determine the quantity at any point on the machine. It is not velocity but force that is analogous to electric current. Force can- i not be measured unless one breaks into the device. Values for a, & and ¢ in Electrical, Mechani ese TABLE 3.2 : ans yas oie ° ’ 1 PreferrRO-M CHANO-ACOUSTICAL CIRCUITS (Chap. It becomes apparent then that if a mechanical element is stetly analogous to an electrical element it must have a velocity difference appeating between (or acrou) is wo terminals ands free acing Cough it Analogounly als, the produet of ie rma foree/ in newlane ard the inphase eompovent of the nn velocity w in meter por second Je the oF ———r—C— rrr torreeponds to a voltage and a force to a ourant, the mobilgnine tmalogy. tis also known as the “averse” asalogy ‘Many text teachin addition a "dieet” analogsy Tt the opposite of the mobility analogy in that force is made to correspond to voltage and rh ———F_———— Impedance type analogy To fain the student with bath consepte ‘all examples will be given here both in mobility-type and impedance-type analogies. Mechanical Impedance Zy, and Mechanical Mobily tw. ‘The mechan- Oe —r———i— tt mechanical device, "We commonly use the symbol Ze or mechani impedance, where the subscript M stands for “mechanical.” "The unit tre nenton-acconda per meter, or mka mechaneal hina "The mechanical mcilty isthe inverse of the mechanical impedance, 1Linthe complex ata volrtytatce nea gees eaat a ooh devieesWeommonly ure the symbol fs for mechanical mating Te Unite are meters por seco per newton, or mia rascal 14. cal mohms.+ —— Je selon a omelet opto that force. ‘The units the kilogram, AUN ght ase LL ———_N convection ip needed to et it in motion” Howevty, the Je. Me force acting on t mase and tho renllant acederalon {ArH reckoned with reapet to ube earth neta rae) co thet Inreality the second terminal of masse the earth he mechanival symbol used to represen mast is shower i Fig. 31 ‘The upper end of the mase moves with a velocity wth respect fo te ground. ‘The }shaped configuration represents, the "aesnd” terminal Of the mass end has aero velocity. "The force cam be measured by a tuilable device inorted between the point I and the next elomen cr reverator connecting toi Maan Mw obeys Newton second aw that au) Sy = Mu @y {The word “mohm” stands for mobility ohm. The vty are meters pee second Pact VE) MECHANICAL CIRCUITS 53 ‘whore /() is the instantaneous force in newtons, Mw is the mass in kilo- grams, and 1(0) is the instantaneous velocity in meters per second, In the steady state [see Eqs. (2.38) to (2.35)], with an angular frequency «sequal Lo 2r times the frequeney of vibration, we have the special ease of Newton's second law, joM we 82) where j = \/=i as usual and f and w are rms complex quantities, The mobility-type analogous symbol that, ; we use as a replacement for the mechanical I symbol in our circuits is a capacitance type. [>| re Wisshown in Fig. 320, Themathematical fu eae Py Say 4 ‘operation invariant for this symbol is found from Table 3.1. In the steady state we have am or = L658) ty toe tpt jue oN (a) 0) ‘This equation is seen to satisfy the physical Fra. 3.2. (a) Moblity-ype law given in Bq, (3.2). Note the similarity $4 (impedance type ym- in appearance of the mechanical and analo- coussvmbolsin Figs. 8.1 and 3.2a, Inelectrical eireuits the time integral of the current through a capacitor is charge, The analogous quantity here is the time integral of force, which is momentum, ‘The impedance-type analogous symbol for a mass is an inductance. It is shown in Fig. 3.25, ‘The invariant operation for steady state is 4a = jucb orf = juM yn. Ialso satisfies Bq. (82). Note, however, that in this analogy one side of the mass element is not neces- 1) Ju, sarily grounded; this often leads to confusion. In elec- trical circuits the time integral of the voltage across an inductance is fluxctums. The analogous quantity here is momentum, Mechanical Compliance Cy. A physics} structure is 2ll Yur said to be a mechanical compliance Cw if, when itis acted ie. 83. Me. on by a force, its displaced in direct proportion to the chasiealsymbol force. The unit is the meter per newton. Compliant for a mishan~ elements usually have two apparent terminals ‘The mechanical symbol used to represent a mechanical compliance isa spring. It is shown in Fig. 3.3. ‘The upper end of the element moves with a velocity u: and the lower end with a velocity us “The force required to produce the difference hetween the velocities u; and tu: may be measured by breaking into the machine at either point } or point 2. Just as the same vrrent would be measured at either end of an element in an electrical cirewit, 20 the same force will be found here at either end of the compliant element. Cw54. RO-MECHANO-ACOUSS AL CIRCUITS [Chap.3 Mechanical compliance Cy, obeys the following physical ia, ant foa or 0 Z fue ea) where Cy is the mechanical compliance in meters per newton and u(t) is the instantaneous velocity in meters per second equal to us — u, the difference in velocity of the two ends, In the steady state, with an angular frequency w equal to 2x times the frequency of vibration, we have, i jae 5) where f and u are taken to be rms complex quantities. ‘The mobility-type analogous symbol used as 2 replacement for the mechanical symbol in our circuits is an inductance. Tt is shown in Fig, 3.4a. ‘The invariant mathematical operation f © that this symbol represents is given in Table 3.1. In the steady state we have w= jeoluf G6) _ ae gee Ls Ly In electrical circuits the time integral of the voltage acrose an inductance is flus-tusns esaty-toe —Ingetne-tye ‘The analogous quantity here is the time in- () (0) ‘tegral of velocity, which is displacement, Fro. 34. (2) Mobility-type ‘This equation satisfies the physical law ni) npotaeetype given in Eq, (QS). Note the wimaiy in fei fr athe Sompir cance the mechanical and analogous symbols in Pig 8.8 and 3. ‘The impedance-type analogous symbol for a mechanical compliance is a capacitance. It is shown in Fig. 3.45. ‘The invariant operation for steady state isa = b/jue, or f = u/jwCy. It also satisfies Bq. (3.5). In electrical circuits the time integral of the current through capacitor is the charge. The analogous quantity here is the displacement, Mechanical Resistance Ry, and Mechanical Responsiveness tw. A physical structure is stid to be a mechanical resistance Ry if, when it is acted on by a force, it moves with a velocity direetly proportional to the force. ‘The unit is the mks mechanical ohm. ‘We also define here a quantity ry, the mechanical responsiveness, that, is the reciprocal of tw. ‘The unit of responsiveness is the mks mechanical mohm, ‘The above representation for mechanical resistance is usually limited to viscous resistance. Frictional resistance is excluded because, for it, the ratio of force to velocity is not a constant. Both terminals of resistive ements can usually be located by visual inspection, Pact Vi) MECHANICAL CIRCUTTS 55 ‘The mechanical element used to represent viscous resistance is the fluid dashpot shown schematically in Fig. 3.5, ‘The upper end of the element. moves with a velocity w and the lower with a velocity us. ‘The force required to produce the difference between the two velocities u; and ts may be measured by breaking into the machine at either point 1 oF point 2 Mechanical resistance Ry obeys the following physical law, foram de where f is the force in newtons, 1 is the difference between the velocities tu, and uz of the two ends, Hy is the mechanical resistance in mechanical ohms, ie, newtons/(meter per second), and ry is the mechanical responsiveness in mks mechanical mohms, i.e., meters per second per newton. ‘The mobility-type analogous symbol used to replace the mechanical symbol in our circuits is a resistance, It is shown in Fig. 3.6a, ‘The ody tye gta al be (a) 0) al be Fa. 35, Mechanical Fic.3.6. (@) Mobility-typeand Fra. 3.7. Mechenical symbol for mechanical @) impedaneetype symbols for eymboi for a constant (iseous) resistance. mechanical resistance. velocity generator, invariant mathematical operation that this symbol represents is given in Table3.1. Incither the steady or transient state we have we raf G8) In the steady state w and f are taken to be rms complex quantities. ‘This equation satisfies the physical law given in Ea. (3.7) ‘The impedance-type analogous symbol for a mechanical resistance is shown in Fig, 3.00. It also satisfies Bq. (3.7). ‘Mechanical Generators. ‘The mechanical generators considered will be cone of two types, eonstant-velocity or constant-force. A constant. relocity generator is represented as a very strong motor attached to a shuttle mechanism in the manner shown in Fig. 8.7. ‘The opposite ends of the generator have velocities uy and us. One of these velocities either 1 oF ua, is determined by factors external to the generator. The differ-