ACOUSTIC RESEARCH AT THE NATIONAL PHYSICAL LABORATORY*

G. W. C. Kaye, Teddington

The National Physical Laboratory, situated 12 miles southwest of London, at Teddington, is a government institution carrying out investigations and tests of the same kind as the National Bureau of Standards in America.

Work in acoustics was begun at the National Physical Laboratory in 1922. Since that time, the rapid development of technical acoustics in the cinema, radio, and gramophone industries, the advances in architectural acoustics, and also public interest in the problem of combating noise have brought about an ever-increasing number of requests to the laboratory for various acoustic tests and investigations. The acoustic work carried out in the laboratory may be divided into two groups. The first group includes work in architectural acoustics, including the determination of the absorption coefficients of materials used to produce the required reverberation, and measurements of the sound insulation of buildings. The second group is concerned with absolute measurements of sound; this includes measurements of the parameters of various types of acoustic apparatus and measurements of noise. At the beginning of its activity the laboratory had no specially equipped premises, and only in 1933 did it receive the completed part of a new building.

The Acoustic Laboratory

The acoustic laboratory is situated in a quiet locality, away from main roads with heavy traffic. Fig. 1 shows the plan of the laboratory. The completed portion of the building consists of a room for measuring reverberation and two rooms for measuring sound insulation, with adjoining rooms for apparatus and staff. The parts of the building shown on the plan by dotted lines—namely, the upper room for measuring the sound insulation of floors and the highly damped room for basic acoustic measurements—are to be completed in the near future. To obtain the greatest possible acoustic insulation, the walls of the laboratory are made of massive brickwork; the walls, floor, and ceiling are double. The inner walls,

* Journal of the Acoustical Society of America, Jan. 1936. Translated by B. G. Shpakovsky.

the floor and ceiling of each laboratory (35 cm thick) are completely independent of the external ones and rest on isolating pillars fixed on separate foundations. The isolating pillars consist of cork slabs (about 7.5 cm thick). A special hydraulic installation makes it possible to raise the inner room (weighing 150–200 t) and, if necessary, to replace the cork isolating pillars. Ventilation of the room is carried out

![Fig. 1 and Fig. 2]

Fig. 1. Plan of the acoustic laboratory:
a — measuring rooms, b — reverberation room, c — room for measuring sound insulation, d — deadened room, e — air-heating room, f — storage-battery room, g — electrical room, h — storeroom.

Plan

Section

5 0 5 10 m

Proposed new extension

Fig. 2. Reverberation room.
The photograph shows: a loudspeaker, a microphone, and a test panel made of sound-absorbing material.

by forcing in air, while a constant temperature in the working rooms is maintained by passing warm air between the double walls. All electrical power is brought to the working rooms by means of flexible wires passing through openings between the double walls. If necessary, the entire room, or any part of it, can be electrically isolated by disconnecting the power from external sources.

Sound absorption coefficients

The laboratory has long been engaged in measurements of the sound absorption coefficient. Various methods have been mastered for determining absorption coefficients both at normal incidence (the standing-wave method)² and under the conditions of a reverberant room. The absorption coefficients determined by this latter method,

are precisely those required in calculations of the acoustic properties of rooms; however, even the first method yields substantial data sufficient for preliminary acquaintance and comparison of various materials. The coefficients of sound absorption were determined by the reverberation method over several years in a temporary room, before the construction of the new reverberation chamber. The considerable advantages of the latter made it possible to increase the accuracy of the measurements and to extend them over a large range of frequencies. The new reverberation chamber, shown in Fig. 2, has a volume of \(270 \text{ m}^3\), and is asymmetrical both in plan and in section.

The nonparallelism of the walls and the inclination of the ceiling reduce the possibility of the formation of standing waves in the room and create a more uniform and homogeneous distribution of sound, which underlies all reverberation formulas. The double, tightly closing entrance doors are made entirely of steel; the inner door is \(9.1 \text{ cm}\) thick and weighs \(2^{1}/_{2}\) t, while the outer one is \(6.25 \text{ cm}\) thick and weighs more than \(1.5\) t. The hatches located in the ceiling of the room are made double and are operated by electric motors. The floor consists of painted concrete, and the walls and ceiling of painted cement plaster. In the empty room the reverberation time at 500 Hz is about 13 sec. In one of the walls there is a shallow niche, closed by steel doors \(2 \text{ cm}\) thick. In this niche about \(10 \text{ m}^2\) of material can be placed in order to determine its absorption coefficient. If necessary, the material under investigation may be spread out on the floor. The adopted method for determining the absorption coefficient is based on Eyring’s formula and consists in determining the decay time of the reverberant sound in the empty room (or when the material in the niche is closed by the steel doors) and in the room with the material in it. The rate of decay is measured by an objective method. As the sound source, a loudspeaker with a howling tone is used; in order to ensure a uniform distribution of sound energy throughout the room. The reverberant sound is picked up by a microphone, the generated e.m.f. being fed to an amplifier with a variable degree of amplification, located in the adjoining measuring room. Switching off the loudspeaker automatically actuates an electric chronoscope, which is stopped by means of a thyratron relay when the magnitude of the e.m.f. at the terminals of the microphone amplifier falls to a preassigned value. If the amplification is increased by 10 db, this causes a corresponding lengthening of the decay time of the reverberant sound before the chronoscope is automatically stopped, and the resulting difference between the two chronoscope time readings gives the time during which the intensity of the reverberant sound decreases by 10 db. Further increase of the amplification in steps of 10 db makes it possible to determine the decay time of the reverberant sound as its intensity changes within various limits, up to 60 db and more. Usually the measurements are made at various frequencies, varying by octa-

taves, from 125 to 8000 Hz, with the measurements for each frequency being made at four different positions of the microphone.

Fig. 3 presents an example of practically uniform sound decay within 60 db for an empty room and for a room containing 10 m² of the material under investigation. This figure is typical for measurements made with sound of frequency 500 Hz and higher. At lower frequencies the decay of the sound becomes less and less uniform.

Fig. 3. Decay of sound intensity in a reverberation room. Frequency 500 ± 50 Hz. Absorption coefficient of the material 0.85.

As a result of constant demands from industry received by the laboratory during recent years, a large number of different materials have been investigated, such as: acoustic plaster, acoustic tiles, wood-fiber boards, acoustic felt, slag wool, glass wool, and absorption-coefficient values from 0.05 to 0.9 have been obtained. In 1933 the laboratory, having only temporary equipment, took part in an international comparison of absorption coefficients organized by the American Acoustical Society[^3]. The material tested at that time was recently re-examined in the new reverberation room, and the previous measurement results were confirmed. It is intended to pose for investigation the question of the dependence of the absorption coefficient on the size, shape, and distribution of the material.

Sound Insulation

Walls. The development of a method for measuring the sound insulation of walls, partitions, etc., against sounds arising in air was begun in 1924 in a space consisting of two rooms separated from each other by a double wall with an opening of 1.5 × 1.2 m², closed by the specimen under investigation. The rooms were strongly damped, and in one of them a sound beam was produced, directed at a certain angle toward the opening. The sound-transmission coefficient of the specimen under investigation was determined on the basis of measurements of the sound intensity that had passed into the second room through the open opening and through the opening closed by the specimen under investigation. In this way it was established for a large number of materials, beginning with paper and ending with an 11-cm brick wall, the fundamental proposition according to which the reduction of sound when passing through a single-

...ordinary homogeneous partition is fully determined by its weight (per 1 m²) ^{4,5}.

Much research has also been carried out on composite partitions. Properly designed double partitions may prove to be more sound-insulating than an ordinary partition having the same total weight. However, an irrationally chosen distance between the two parts forming one system of a double partition may make it less sound-insulating than an ordinary partition of the very same weight ^6, as is clear from the example in Fig. 4.

Fig. 4. Influence of the distance between the frames of a double window on the magnitude of the sound insulation. The curves show the minimum sound insulation.

New rooms for measuring sound insulation, shown in Fig. 5, make it possible to carry out measurements on specimens of larger size and with greater insulating properties. They are located on different foundations independent of one another and have the same irregular shape and general construction as the reverberation room already described, with the sole exception that the entrance door is made double, of wood 7.5 cm thick. The opening separating these two rooms can be closed by the test specimen, up to 3 × 2.4 m² in size. In one room there is a hoisting crane for moving heavy specimens. At present the measurements are carried out in the reverberating room, i.e. not muffled, and the sound falls on the specimen at the most varied angles; but a device is provided that makes it possible to muffle the room in those cases when directed sound must be used for the measurement. A loudspeaker located in one of the rooms gives a “howling” tone. Measurements of the average sound level in each room with the opening open and with the opening closed by the test specimen make it possible to determine the value of the sound insulation for the given partition. The average sound level is determined by means of a microphone placed in various positions. The movement

Acoustic Works at the National Physical Laboratory

The microphone is operated from an adjacent measuring room, where the generator feeding the loudspeaker and the microphone amplifier are located. Measurements are usually made at several frequencies in the range from 100 to 4000 Hz.

The rooms described, intended for testing the sound insulation of partitions, can be quickly adapted for investigating the attenuation of sound in ventilation ducts and silencers and for experiments on noise suppression in machines.

Floors. Owing to the absence of the upper operating room (Fig. 1), a temporary installation was constructed for determining the comparative insulating properties of various floors with respect to impact sounds (for example, sounds caused by footsteps). A floor specimen (measuring \(2.4 \times 1.5\ \mathrm{m}^2\)) was set into a heavy reinforced-concrete frame installed above an opening in a reinforced-concrete floor 15 cm thick. The frame rested on a rubber strip, which insulated it from the parts of the concrete floor surrounding the opening. A special impact machine with 4 hammers, driven by a motor, produces a sequence of impacts corresponding in force to footsteps. Two types of hammers were used: hard (ceramot) and soft (rubber).

Fig. 5. Chambers for determining sound insulation with a window into which the specimen under investigation is inserted

The loudness of the noise produced below, under the floor being investigated, was determined subjectively by a group of researchers. A large number of complex floors were tested. Work in this direction is being carried out jointly with the Architectural Research Committee by means of a similar temporary installation, making it possible to investigate floor specimens of still larger size (\(5.4 \times 4.5\ \mathrm{m}^2\)).

The work carried out in the laboratory setting on the investigation of the sound-insulating properties of floors and walls is often supplemented by visits to buildings under construction. An extensive program of acoustic investigations of buildings is being carried out on assignment from the Ministry of Internal Affairs. The demands for corresponding investigations that have come in recent years from private firms are so numerous that it has proved necessary to continue operating the old installations alongside the newly built ones.

Architectural Acoustics

The laboratory sets itself the task of creating good acoustics in halls both already built, but possessing defects with respect to acoustics, and newly under construction.

An apparatus was constructed for photographing the process of propagation of sound impulses in a model of a building and was applied to reveal the causes of undesirable echo and for demonstration purposes ⁷·⁸. Research in this direction is being conducted at the request of the government, public organizations, and private firms, both local and foreign.

Absolute Measurements of Sound

Absolute measurement of sound is absolutely necessary for the calibration of microphones and their subsequent use in all kinds of sound measurements. The methods used in the National Physical Laboratory are ultimately based on the use of Rayleigh’s disk, which makes it possible to measure the velocity of particles of the medium. Condenser microphones, calibrated in absolute units, serve as laboratory standards with which working microphones are compared. For the calibration of microphone standards, thin, round, glass disks of 1 cm diameter, silvered on one side, are used. The disks are suspended on thin quartz threads, whose torsional modulus has been measured beforehand, and the velocity of the particles of the medium can be calculated from the measurement of the angle of deflection.

Determination of the sensitivity of the microphone standard, i.e. the ratio of the electromotive force induced in the microphone to the magnitude of the sound pressure in a free wave, is carried out in a highly damped room, shown in Fig. 6. The room is approximately a cube with a side of 3 m (external dimension) and is lined inside with a layer of sea grass 15 cm thick, over which is suspended a layer of cotton-wool wadding with a total thickness of 30 cm. The floor, damped in exactly the same way, is moreover covered with a removable metal grating. As the sound source, a loudspeaker installed in one corner of the room is used. The particle velocity at a number of points located on the axis of the loudspeaker is determined from the deflection of Rayleigh’s disk, read from outside with the aid of a light spot moving along a transparent scale. The corresponding pressures in the sound wave are calculated by means of known formulas. Then, in place of the disk, a microphone is installed, and the electromotive force obtained at its terminals is measured for each of the same series of points. These measurements are made for various frequencies, from 10,000 to 300 Hz. At still lower frequencies the reflected wave (from the damped walls of the room) begins to make itself felt, and the simple formulas valid only in a free (traveling) wave are no longer applicable.

At these frequencies, indirect methods are used to determine the sensitivity of the microphone.

Another method for determining the sensitivity of a standard microphone, i.e., the ratio of the electromotive force arising in the microphone to the magnitude of the sound pressure applied to the microphone membrane, is based on measurements using a Rayleigh disk in a standing wave. The Rayleigh disk is suspended on the axis of a cylindrical tube with a cross-sectional diameter approximately equal to the membrane of the microphone, which closes one end of the tube. Standing waves are established in the tube, excited by a loudspeaker placed opposite the open end of the tube. The sound frequency is selected so that the Rayleigh disk is located at a velocity antinode; the angle of deflection of the disk measured in this case makes it possible to calculate the magnitude of the sound pressure acting on the membrane of the standard microphone. By this method, calibration can be carried out in the frequency range from 3500 to 62.5 Hz. The upper frequency limit is constrained by the possibility of transverse oscillations arising in the tube, while the lower limit is constrained by the inconvenience of using excessively long tubes. Therefore, for frequencies below 62.5 Hz, another method was developed for calibration using a special piston emitter (pistonphone). The latter consists of a small piston set into oscillation by means of the moving coil of an electrodynamic reproducer. The piston excites oscillations in a small chamber, one side of which is closed by the membrane of the standard microphone to be calibrated. A steel needle attached to the reverse side of the piston rests against the horizontal arm of a lever bent at a right angle and suspended on a torsion thread. On its vertical arm a small mirror is fixed, throwing a light spot onto a scale. The deflections of the light spot along the scale, occurring during oscillations of the piston, make it possible to measure the amplitude of the piston oscillations and to calculate the corresponding values of the sound pressure in the chamber, and therefore also near the membrane of the standard. With the aid of the method described, the standard microphone was calibrated in the range—

Fig. 6. Sound-treated room for acoustic measurements

zone from 10 to 400 Hz. The calibration curves of the standard are shown in Fig. 7. It is evident from the figure that both calibration methods give identical results at frequencies lying below 400 Hz. Therefore, for a microphone of the accepted size (about 7.5 cm in diameter), the results of calibration in a standing wave (which determines the so-called pressure sensitivity) at low frequencies, where the method of calibration in a freely propagating wave (which determines the so-called free-sound-field sensitivity) cannot be directly applied, are quite sufficient for judging the sensitivity in the sound field even at low frequencies. In the frequency range from 62.5 to 400 Hz, where measurements were made by means of the two calibration methods, the agreement of the calibration results obtained appears highly satisfactory. In the case of “velocity” microphones, which respond essentially to pressure gradients, as well as microphones that respond simultaneously both to sound pressure and to the pressure gradient, another method must be applied for calibration and for determining the sensitivity in the sound field at low frequencies. The calibration method used in this case consisted in producing standing waves in a wooden tube about 6 m long and 60 cm in cross-section, in the closed end of which a standard microphone was mounted, determining the pressure magnitude at this point. The sensitivity in the sound field of the microphone being calibrated was determined by calculation on the basis of the readings it gave when placed at various points of the standing wave with a known distribution of pressures and velocities.

Fig. 7. Calibration curves of a standard microphone.

Loudspeakers and Other Acoustic Instruments

Measurements of the acoustic power emitted by loudspeakers are made in the open air on the roof of the building of the Physical Laboratory. The loudspeaker is installed on a platform at the edge of the roof and is directed with its horn toward the open space. The measuring microphone, placed at a height-

cating forward on a long lever, makes it possible to measure pressures in a sound wave at distances up to 2.5 m. The rotating lever makes it possible to take polar diagrams of sound radiation. Properly arranged reflectors deflect the reflected sound away from the microphone. The loudspeaker is excited by means of an audio-frequency generator, and the microphone readings are recorded automatically; measures are taken to eliminate (by means of special electrical circuits) extraneous noises and the influence of wind. Distortions of the form of the sound curve introduced by the loudspeaker are established by harmonic analysis of the sound recorded from a loudspeaker fed with sinusoidal alternating current. Similar measurements are also carried out with telephones. In this case the telephone is mounted together with an “artificial ear,” calculated and constructed so as to create the same acoustic load on the telephone diaphragm as the average human ear, and pressure measurements are made in the cavity of the artificial ear.

In the laboratory, frequency characteristics are being taken of gramophone adapters, acoustic amplifiers (prostheses) for the deaf, and also the rate of decay of sound intensity (the decrement of damping) of tuning forks used by otolaryngologists in the investigation of hearing. In addition, investigations are being carried out in other directions as well; for example, the velocity of propagation of sound in tubes is being studied.

Noise Measurements

For several years the problem of noise—i.e., the measurement of noises, the investigation and analysis of noises of various origins, and measures for combating noises—has constituted a considerable part of all the work of the Acoustics Laboratory^10–15^. Loudness measurements were made both by subjective methods, with the aid of instruments of the Barkhausen noise-meter type, and by objective methods. In doing so, the advantages and possibilities of each type of instrument were studied. The laboratory developed its own design of an apparatus for the objective measurement of the loudness level of noise of medium intensity. This apparatus, shown in Fig. 8, consists of a condenser microphone, an amplifier, a rectifier, and an indicator. The degree of amplification given by the amplifier can be changed in steps of 1 db. Specially calculated electrical circuits serve to correct the frequency characteristic of the microphone, and also to reduce the amplification at low and high frequencies, in order to bring the indicator readings into accord with the sensitivity of the ear according to equal-loudness curves at a loudness level of about 70 db. The constructed instrument imitates the properties of the ear also in that it responds correctly to intermittent sounds, both to individual impulses and to a sequence of impulses of various degrees of repetition. A description of this instrument will be published shortly.

Beginning in 1929, considerable work was carried out on assignment from the Aeronautical Research Committee on the study of airplane noise[^16]. Measurements were made both on airplanes in flight and on noise produced by propellers and engine exhaust gases under experimental conditions. In connection with the improvement of exhaust-noise mufflers, a study was made on a model of a muffling exhaust pipe, considered as an acoustic filter.

Fig. 8. Portable acoustimeter (noise meter) for objective measurement of sound intensity and loudness level

The laboratory has carried out studies of transport noise for the railway company and frequently provides consultation on problems of noise control in equipping premises for testing airplane engines, printing houses, transformer substations, noisy offices, etc.

Recently, on assignment from a committee under the Ministry of Transport, the laboratory began carrying out a large program of work on the study of noise produced by mechanical transport[^17]. Objective measurements were made of the loudness of noise produced by various motorcycles, automobiles, and trucks under different operating conditions, namely: a) in a stationary position with the engine running and b) when moving at various speeds and accelerations, both on level ground and uphill.

As a result of our investigations, the committee appealed to the Minister of Transport with a proposal to prohibit both the sale and the use on highways of motor transport producing noise exceeding a certain established limit. For private automobiles this limit is 90 db at a distance of 5.4 m from the center of an automobile moving at a speed of 50 km per hour, or 95 db at a distance of 5.4 m from the exhaust pipe of a stationary automobile with the engine operating at full power.

The laboratory was also invited to carry out, for the committee of the Ministry of Transport, a study of horns and other signaling devices used in motor transport. The committee intends to take measures to establish a definite characteristic for the signals used, with the aim of reducing those unpleasant auditory sensations that may arise from a signal.

References

1. G. W. C. Kaye, Nature 124, 202, 1934.
2. A. H. Davis and E. I. Evans, Proc. Roy. Soc., A 89, 127, 1930.
3. P. E. Sabine, J. Acous. Soc. Am., 6, 239, 1935.
4. A. H. Davis and T. S. Litter, Phil. Mag., 3, 177, 1927.
5. A. H. Davis and T. S. Litter, Phil. Mag., 7, 1050, 1929.
6. I. E. R. Constable, Phil. Mag., 18, 321, 1934.
7. A. H. Davis and X. Fleming, J. Sci. Inst. 3, 393, 1926.
8. N. Fleming, Proc. Seventh Int. Cong. Phot., 318, 1928.
9. G. W. C. Kaye and G. C. Sherratt, Proc. Roy. Soc., A 141, 123, 1933.
10. G. W. C. Kaye, Proc. Roy. Inst., 26, 435, 1931.
11. G. W. C. Kaye, Eng. 134, 314, 432, 1932.
12. G. W. C. Kaye, Anti Noise League Exhibition Handbook, 1935.
13. G. W. C. Kaye, Science Progress, January 1936.
14. A. H. Davis, Nature, 125, 48, 1930.
15. A. H. Davis, Eng., 138, 663, 1934.
16. A. H. Davis, Aeronautical Research Committee Reports and Memoranda, No. 1542, 1933.
17. Ministry of Transport, H. M. Stationery Office, 1935.