ACOUSTICS OF BUILDINGS, THE. The subject of the acoustics of auditoriums deals with the behaviour of sound in rooms and in halls, and the conditions which determine their fit ness or unsuitability for the hearing of speech and music.
Certain characteristics of sounds emitted in a closed space re quire to be borne in mind. The sounds proceed outwards from the source in spherical waves until they strike the boundaries of the room. At the boundaries they are partly absorbed or trans mitted to an extent depending upon the nature of the surfaces, but the remainder is reflected. Simple echoes may thus arise, and if any large portion of the surface is appropriately curved objection able focussing of sound may occur. However, continuing the his tory of the reflected sound, it travels quickly, so that through numerous successive reflections its energy is very soon more or less uniformly diffused throughout the room, every element of volume being filled with waves proceeding in every direction. Again, since sound is a form of energy, when once produced in a confined space it will continue to exist until it is transmitted by the boundary walls or transformed into some other kind of energy, generally heat. Consequently if the boundaries are hard and non absorbent, little energy is absorbed at each impact and, owing to the very many reflections which occur before the sound dies away, prolonged reverberation is noticed. With more porous wall sur faces the aerial vibrations are damped out by frictional conversion into heat within the pores of the material, and reverberation is re duced. If the source of sound is continuously maintained the loud ness in the room increases rapidly until the rate of emission of sound from the source is balanced by the absorption at the sur faces. In the steady state thus reached the loudness at a given point is generally greater than it would be in open air and, apart from echo and focussing effects, more uniform in distribution. It may be mentioned further that with sustained notes interference phenomena may be observed as the ear is moved from side to side, maxima and minima of sound being perceptible at different points. In the acoustics of buildings, however, these latter effects are of subsidiary importance, partly because the use of two ears tends to obscure them, and partly because in speech—and to a lesser extent in music—the sounds are changing so rapidly that an interference system is never completely set up. Finally, resonant vibrations of columns or volumes of air in the chamber may cause distortion of sounds of appropriate pitch, and sympathetic vibra tion of floors in contact with the source, or of panelling close at hand, will increase the volume of sound emitted.
The chief conditions for good hearing in an auditorium are therefore: (a) that loudness should be adequate; (b) that there should be no perceptible echoes or focussing; (c) that there should be no undue reverberation (i.e., each speech sound should die away quickly enough to be inappreciable by the time the next is uttered) ; (d) that where best music is concerned the hall should be non-resonant and as uniformly reverberant as possible for sounds of all musical pitches, in order to preserve the proper rela tive proportions of the components of a complex sound; and (e) that the boundaries be sufficiently soundproof to exclude extra neous noise.
Loudness is also enhanced by keeping the ceiling of a hall low so that it may act as a reflector to strengthen sounds reaching re mote seats. This device is specially valuable for Council cham bers in which members speak from where they stand, for no other surface can be equally effective for all positions of the speaker. In order that the audience in galleries may benefit from the re flection, the galleries should be under the main ceiling of the hall.
Electrical amplifying equipment, in conjunction with large loud speakers, is now used successfully to increase the volume of sound, particularly in the case of speech. When this apparatus is used the assistance of other devices is not required. Amplification, however, must not be excessive, for excessive loudness gives un due prominence to low pitched sounds. It is necessary, therefore, to amplify only to an extent such that remote listeners can hear with comfort, to place the projectors well above the speaker's head, and so to direct them that the sound is not excessive for hearers near the platform. Loudness is then fairly uniform over the floor space and the majority of the audience has the impres sion of listening to only one source of sound—the speaker him self. Where a hall has a number of high galleries a double ring of projectors may be required above the speaker's head, one in clined downwards to the body of the hall, the other upwards to serve the galleries. Pl. I., fig. r, showing two loud-speakers in position on the north side of the chancel of Liverpool cathedral, is an example of loud-speaker mounting in harmony with sur roundings.
The question of loudness does not often arise in connection with orchestral and choral performances, and auditoriums for these pur poses may be larger than those satisfactory for ordinary speech. It has been estimated by Heyl from considerations of loudness that the optimum volume of a concert hall increases from 50,000 cu.ft. for an orchestra of ten instruments to 800,000cu.ft. for one of 90 instruments. For musical performances a wooden platform undoubtedly acts as a useful sounding board for instruments and performers in contact with it.
For the prevention of echo it is desirable to avoid altogether ceilings of greater height than about 4o feet. Where this restric tion is not acceptable, it is useful to coffer the ceiling deeply so that the echo may be scattered to some extent, and to apply highly absorbent materials in the panels. A less obvious expedient, but frequently an important one, is to apply absorbents to the upper parts of walls, and so to suppress sound which otherwise would be reflected upwards and returned to the floor of the audi torium later as an indirect echo. Frequently back walls should also be treated. In some cases, such as barrel vaulting of great height and pronounced curvature, a canopy of absorbent fabric hung near the focal region may have special value.
Echoes from high ceilings are largely eliminated when electrical amplifying equipment is used for increasing loudness, for the speech projectors are usually placed at least 25ft. above floor level.
The general direction of likely echoes and the desirable situa tions for absorbents may usually be inferred from an inspection of sections of the chamber. For instance, on the assumption that sound obeys the laws of reflection of light, P1. I., fig. 3, upper, is an analysis of sound reflection within the vertical longitudinal sec tion of a council chamber. Floor reflections are omitted. On the scale adopted, it shows portions of a sound wave of ter travelling a total distance of 6oft. from the black dot which represents a speaker on the floor of the chamber. For convenience arrows have been added to indicate the tracks of a number of the waves. In interpreting such drawings, however, it must always be borne in mind that, while for large surfaces sound tends to obey the ordinary laws of optical reflection, it usually spreads considerably beyond the limits applicable to light owing to its relatively greater wave-length.
Experimental methods of analysis have been developed in which observation is made of the progress of a sound pulse within a model or within model outlines of suitable sections of the build ing. This was first carried out by W. C. Sabine in America, using the technique of sound-pulse photography which had been devel oped by various physicists, particularly Toepler, Dvofik, Mach, Boys and Foley. The sound pulse is produced by means of an electric spark in a model section having open sides. As this pulse spreads in the model it is illuminated instantaneously by light from a distant electric spark. A silhouette of the model is thus cast upon a screen or photographic plate and, in addition, owing to refraction of light by the sound wave, the position of the sound pulse within the section is also shown. Sabine's own work contains many beautiful photographs relating to theatre acoustics. Fig. 3, middle, is a sound-pulse photograph taken in England at the Na tional Physical Laboratory, relating to a section of a council cham ber closely similar to that analyzed geometrically in fig. 3, upper, the floor being absent. General correspondence with fig. 3, upper, is marked, but the spreading of sound beyond optical limits is clearly brought out. Pl. I., fig. 2, due to A. H. Davis and N. Fleming of the Laboratory, is a series selected to show the earlier and later progress of a sound pulse in a very similar section. The first of the series shows the outgoing wave proceeding equally in all directions from the source, and later photographs show the subsequent history of the wave after reflection from the boun daries. In particular, the third photograph shows that the main ceiling reflection reaches hearers in the right-hand gallery within less than i/i8th of a second after the passage of the original sound and is thus useful to them, while hearers in the left-hand gallery—which is in an alcove—are largely denied such enhance ment of the sound.
In another method of study which is simpler in technique, use is made of the approximate similarity between sound waves and water waves. A model section of the building is laid flat in a tank of shallow water. Ripples are produced at a point corresponding to a speaker's position by dipping a small object in the water, and the reflection of ripples at the boundary illustrates the reflection of sound in the actual building. To facilitate study and photog raphy of the ripples, the tank has a glass bottom and light from an arc lamp passes upwards through the glass bottom and casts a shadow of the model and of the waves upon a screen mounted above. Pl. I., fig. 3, lower, is a ripple photograph for the audi torium section already referred to in connection with the geometri cal and the sound-pulse analyses.
When echoes are noticed in existing buildings they may be traced in suitable cases by ear. For the survey the hiss from an arc light placed at the focus of a reflector is a useful source of sound, first recommended by F. R. Watson. The reflections of the arc hiss may be traced by ear, but the beam of light accompany ing the beam of sound reduces the labour of locating the place from which reflection is taking place.
It is frequently thought that defective acoustics in a hall can be cured by stretching wires to and fro near the ceiling. Whatever may be the value of such wires—and it is very doubtful whether they have any value at all—they do not reduce reverberation. There are instances where miles of wire have been employed, with no useful effect.
Neither is the introduction of electrical amplifying equipment a universal cure for excessive reverberation. Used with discrimina tion—say, in a case of reverberation due to a very high ceiling— it has a value which arises from the directive action of the loud speakers and from the fact that, in projecting the sound down wards upon the audience-covered floor, the loud-speakers are directing it towards what is usually the most absorbent area in the auditorium. Apart from this directive action, however, the equipment does nothing to hasten the decay of sound which per sists sufficiently to cause confusion. Indeed, in a highly reverber ant hall this persistent sound affects the microphone and tends to cause the apparatus to emit a continuous note. However, the reverberant condition of a hall may be measured, controlled, and even predicted in advance of construction and arranged to con form with the condition generally approved.
For purposes of measurement the rate of decay of reverberation in a hall is expressed in terms of the time which elapses after the finish of a sustained note before the general intensity of sound in the room falls to inaudibility at one millionth of its initial value. This interval is called the reverberation period of the hall. It is a convenient practical period because a source of sound such as an organ pipe ordinarily gives rise to an intensity in a room of the order of a million times the minimum audible intensity. More over, it is a definite acoustical constant for a given condition of the room, for it has been found that the observed duration of rever beration is almost independent of the position of the source and of the observer and, for a given amount of absorbent in the room, is largely independent of its distribution.
In order to be able to calculate the reverberant condition of an auditorium in advance of construction, it is necessary to know the factors which determine it. The theoretical basis of the subject, due largely to W. C. Sabine and G. Jaeger, is satisfactory and yields formulae which have been confirmed repeatedly by experi ment. It is found that if V is the volume of a room, and if the surfaces, of total area S, absorb on an average a fraction "a" of the sound energy at each incidence, then the standard reverbera tion period T in which the reverberant sound decays one million f old is given by T=o•o5 V/aS in foot second units or T=o.164 V/aS in metre second units.
In the formulae the mean absorption coefficient "a" of the surfaces of the room is calculable from the absorption coefficients of the component parts. Thus if there are surfaces S2, 53, etc., having respectively absorption coefficients ai, a2, etc., then aS=a1 Si+a2 S2-1-a3 S3+ From the formulae it is seen that reverberation may be decreased by reducing the volume of the room and by increasing the absorb ing power of its surfaces.
Experiments made in a number of acceptable auditoriums in various countries by W. C. Sabine, F. R. Watson, S. Lifshitz and others, show that reverberation tends to have a preferred duration for good acoustics. For halls of moderate size, up to, say, 40,000 cu.ft., which are to be used both for speech and music, it is gen erally agreed that a standard period of about one second repre sents the optimum condition, the audience of course being present. For a hall five times as large, i.e., 200,000 cu.ft., the preferred period is apparently about II sec., and for very large halls of about i,000,000 cu.f t. a period of 2 sec. is indicated. Presumably in very large halls the persistence of sound is tolerated for the sake of increased loudness, and speakers are expected to enunciate more slowly to accommodate themselves to the conditions.
Excessive reverberation is the more serious for speech, but in sufficient reverberation is unacceptable for music. In fact, accept able halls which are used for music alone appear to have rever beration periods some 25% greater than the values given above for halls of corresponding size used for mixed purposes. More over, for concert halls it has been found desirable to leave the region of the stage bare so as to give the feeling of easy response desired by the artistes and, if absorbent is required by hearers to deaden reverberation, it should be placed away from the stage so that both performers and hearers may regard the acoustics as satisfactory. Incidentally, any absorbing material introduced to reduce reverberation has also a value in damping out resonances in the air of the room and thus minimizing distortion that might arise from this cause.
Generally speaking, churches and cathedrals are characterized by long reverberation owing to great size and to the very slight absorbing power of the masonry walls. The condition may well be responsible for the development of the characteristic features of church choral music and intoned liturgy. Ordinary sermons, how ever, cannot be satisfactorily heard, and, in churches where read ing and speaking are dominant factors of the service, the acoustic condition should obviously tend to conform to the optimum rever berant period already discussed.
In connection with the materials that may be used to control reverberation, the following table presents some useful average values for the absorbing power of unit areas of various types of absorbing surface. The values relate to the note (c", vibra tions per second) near the middle of the musical scale, with which most reverberation measurements have been made. Almost inva riably the absorbing power is less at lower frequencies, but differ ences at higher frequencies are not so consistent and, in general are not so marked.
It must of course be borne in mind that the absorption co efficient of a material will be dependent to some extent upon the manner in which the material is attached to the walls and that, owing to diffraction phenomena, somewhat abnormal values may be exhibited by isolated samples of which the dimensions are not great compared with the wave-length of sound.
The efficiency of a sound-absorbent material is usually due to the porosity of its surfaces, and the acoustic plasters referred to in the table have an aerated texture arrived at during mixing either by mechanical frothing or by the addition of a gas generating ingredient. They give a pleasing finish for ceilings and for parts of wall surfaces where roughness is not objectionable. The acous tic tiles available are similar in appearance. Fibrous boards have also been developed for sound absorption; various soft materi als such as clean hair felt, jute felt, eelgrass, fireproofed wood wool, asbestos or slag wool may be applied in panels, behind a screen of canvas, net or rep. Plate I., fig. 4 is a photograph of the lecture hall of the Royal Institute of British Architects, which has been treated in this manner over the whole of the back wall and frieze, and in the spandrills of the central dome, the covering canvas being in special shades to match the decoration. The screen chosen should be as light as possible so that it does not reflect sound, but transmits it to the absorbent beneath, and is best unpainted. Paint yields a more sanitary surface, but when applied with a brush, the absorbing power of the panel is reduced by as much as 15 to so% for one coat, with a further, but smaller, reduction for a second coat. Distemper sprayed over the screen is a usual form of decoration, but the membrane should not be previously sized. If undesirable reverberation in an auditorium is considered in advance of construction, space in the hall may be suppressed. To prevent the entrance of sound to a room, it is necessary to isolate against air-borne sound and structure-borne vibration.
Air-borne sound may be excluded by having walls and partitions sufficiently massive and rigid, and avoiding openings. Usually it is found that a rigid brick wall presents adequate obstruction to the entrance of aerial sounds. In very special cases double walls are erected, preferably on separate foundations, with an air space between them. Fillers of felt or slag wool, etc., do not add to the insulation in this case, but, by bridging the air space, slightly facilitate the conduction of vibration. Transmission of sound through partitions is mainly due to flexural vibration of the parti tion under the action of the incident sound, so that increasing the weight and the rigidity of a partition both improve the insulation. Generally speaking, the fundamental resonant frequency of ma sonry constructions—brick walls, hollow tile, lath and plaster—is quite low and, in consequence, for incident sounds of ordinary pitch it is the mass of these partitions which is the most important factor. In certain experiments by P. E. Sabine, the following average figures for resistance to aerial sound of various frequencies were found for this class of partition irrespective of the actual type of construction employed : *This corresponds roughly to a 4k-inch wall of medium brickwork.
To interpret these figures, it should be realized that the usual intensity of speech is about one million times the minimum audible intensity, and one-millionth of the intensity at which sound be comes painful. The total reduction for two completely separated walls would be equal to the product of their individual values, but, in practice, resonance of the air space and incomplete isolation of foundation result in lower figures. Some quantitative informa tion has recently become available concerning various types of stud partitions but there is not yet unanimity among observers as to the absolute magnitude of the transmitter sound.
It is important that doors and windows be in the position most conducive to quietness, and that they shut completely so as to exclude the relatively considerable amount of sound that can enter through cracks. In door construction rigidity, as well as mass, is important. Frequently, for effective insulation, doors must be double, and in extreme cases even triple, separate frames are employed. In windows heavy plate-glass is best and small well braced or leaded panes are rather better than large panes. Double windows are most effective when they are in separate insulated frames. Particular attention must also be paid to the ventilating arrangements, which may readily facilitate transmission of sound from room to room. It appears most satisfactory for the rooms to be supplied with ventilation ducts which communicate sepa rately with a distant supply chamber, and it is advantageous to line the interior of the ducts and of the chamber with felt-like absorbent.
To isolate a room from structure-borne sounds, such as those arising from motors, lifts, hot-water pipes and attached machinery generally, it is necessary to break up the continuity of the solid conducting path. To this end layers of felt-like material are some times interposed in the path of the sound. It is essential, how ever, that where such insulation is employed no bolts, tie-pieces or other rigid members pass from structure to structure through the layer, unless thoroughly insulated by bushing and by using insulating washers under metal ones. The simplest place to intro duce absorbent is frequently at the source. Motors and machinery may be insulated from the floor by substantial layers of materials such as cork, felt or rubber, or the special combinations of cork and felt, etc., obtainable commercially for the purpose. Alter natively, machinery may be mounted upon springy supports which are themselves insulated from the structure by suitable materials. Lift shafts should be insulated from the main structure and should be surrounded with corridors or stairs. Continuous metal pipes such as air pipes, hot water pipes, etc., may carry sound to con siderable distances ; they should, if possible, have insulating sec tions and should not pass through important partition walls.
Felt-like insulation may also be introduced into the actual struc ture of a building. For instance, a wall may be insulated from the floor upon which it stands by means of a layer of felt, and a further layer may be used to insulate it from the ceiling. In steel constructions insulating material can be employed as a bedding for foundations, columns, girders, cross beams, etc., to counteract the transmission of sound which would otherwise occur so readily.
Finally, whether sounds originate within a room or are trans mitted from outside, the loudness within depends upon the quan tity of sound absorbent the room contains. The provision of soft carpets, of hangings and draperies, of upholstered furniture, and of specially absorbent plasters or materials for ceilings and walls, all tend to improvement, although in themselves they are not suf ficient to obviate the necessity for the various special devices previously set out.