Volume 22, Issue 5, September 1950
Index of content:
- PROGRAM OF THE THIRTY‐NINTH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session A: Noise
- Contributed Papers
22(1950); http://dx.doi.org/10.1121/1.1917168View Description Hide Description
The loudness and threshold of an interrupted white noise (of constant sound‐time fraction) was studied over a wide range of interruption frequencies. White noise—alone among auditory signals—has the special property that, when interrupted, no additional audible complicating spectral products are introduced. Both at absolute threshold and at equal‐loudness above threshold, less energy is required for an interrupted white noise than for a non‐interrupted continuous white noise. In many cases, an interrupted noise(sound‐time fraction constant at 0.45) sounds louder than a continuous noise of the same amplitude (but of greater energy). The intensity required at threshold and at equal loudness is minimum for interruption rates in the region of 4–10 per second. The extent of this minimum region increases systematically as the reference loudness level is increased. A conceptual formulation which attempts to account for the results will be presented.
22(1950); http://dx.doi.org/10.1121/1.1917169View Description Hide Description
In a previous paper a method was described for calculating the loudness of pure tones, combinations of tones, and continuous‐spectrum noise from measurements made with contiguous‐band analyzers. This method is supported by recently published results on loudness. The basis of the method is the hypothesis that the loudness of a noise spectrum can be obtained by dividing the spectrum into bands, replacing each band by a pure tone of equivalent r.m.s. amplitude and then determining the loudness of each tone from equal‐loudness contours and a loudnessvs. loudness‐level function. The total loudness is then the sum of the loudnesses of the individual bands. Comparisons of calculated levels with experimental results of subjective loudness judgments are made assuming an analyzer with ten contiguous frequency bands, each 300 mels wide, and an analyzer with five contiguous frequency bands, each about 600 mels wide. The calculations are made using the equal loudness contours of Fletcher and those of Churcher and King. An important feature of this method is that its operations can be performed electronically, thus providing a design basis for a loudness meter. The operating principles of such a meter necessitate a set of band‐pass filters, and a non‐linear network to simulate the ear's response within each band. A meter designed and constructed for this purpose is described.
22(1950); http://dx.doi.org/10.1121/1.1917170View Description Hide Description
On the basis of a number of surveys in commercial offices, criteria for tolerable office background noise levels are proposed. Questions were asked of people working in these offices regarding their ability to telephone, to confer and to perform their work under the existing noise conditions. The speech interference levels of the background noises were determined in each case. The results are presented in a chart, which rates the noise conditions as a function of speech interference levels in each of three general types of offices, (a) executive offices, (b) general clerical offices, and (c) business machines offices.
22(1950); http://dx.doi.org/10.1121/1.1917171View Description Hide Description
Motors and machines are acoustically complex; varied techniques are required for an analysis complete enough to be of practical use. The dynamic sound spectrograph designed by the Bell Laboratories for speech studies, can, with some modification, be used to good effect in determining the precise nature of the transient sounds which up to now have been difficult or impossible to resolve simultaneously in both frequency and time dimensions. Thus, an easy method becomes available for separating electrical from mechanical noise in motors and motor‐to‐load assembled units, without introducing an external drive. Cavity and case resonances are located by observing which frequencies recur, independently of rotor speed. A second step consists in noting the decay pattern following shock excitation of the suspected member. In most machines, frequencies forced by shaft rotation predominate over formant frequencies. Shaft rotation produces, in addition to its fundamental, an elaborate array of harmonics. Comparison of spectrograms and instantaneous pressure traces (oscillograms) is especially valuable in understanding the complete acoustic system. The limitations of the spectrograph in treatment of sounds below 1000 c.p.s. are caused by the linear frequency scale associated with the present system of heterodyne analysis. This can be overcome by the use of an external rejection filter analyzer with output fed to the marking amplifier.
22(1950); http://dx.doi.org/10.1121/1.1917172View Description Hide Description
The problem of reduction of curb or pullaway noise of motor coaches used for city operation is attacked by the use of a magnetic wire recorder. The recordings, which will be demonstrated, are used in jury tests and for corroborative information by playback through band‐pass filters. The noise sources are attacked and reduced in their order of importance and in a form suitable for production and practical operation.
22(1950); http://dx.doi.org/10.1121/1.1917173View Description Hide Description
Measurements were made of the sounds of three types of commonly used automobile horns. Over‐all sound levels on the axis at three feet ranged from 108 to 125 db. Loudnesses ranged from 125 to 140 phons. One type of horn gave sound levels of 88 db at 50 feet and 74 db at 300 feet, with corresponding loudness levels of 104 and 81 phons. Fundamental frequencies of all horns ranged between 160 and 380 c.p.s. Two types had harmonic overtones, with large amplitudes below about 2000 c.p.s. and smoothly decreasing amplitudes at higher frequencies. The third type, with the most unpleasant sound, had inharmonic overtones, some of which were greater in amplitude than the fundamental. Sound from a pair of horns at various distances was measured inside a closed automobile The over‐all level was 60 db (loudness, 72 phons) at 50 feet and 50 db (loudness, 53 phons) at 300 feet. Filtering out overtones above 1200 c.p.s. reduced the annoying character of horn sounds markedly and reduced the loudness by four phons at 50 feet and only one phon at 300 feet. Since the typical horn is louder than necessary at close range, use of a low pass acoustic filter on automobile horns appears desirable.
- Session B: Transducers, Recording
22(1950); http://dx.doi.org/10.1121/1.1917174View Description Hide Description
The conventional forms of the electromechanical coupling equations exhibit symmetry for electrostatic coupling and antisymmetry for electromagnetic coupling, or vice versa. Both types of coupling can be represented in symmetrical form, however, if the transduction coefficient for electromagnetic coupling includes a space‐operator k that embodies the sign convention associated with axial vectors. The same kind of conventional electric network representations can then be used for either type of coupling. There is a sound physical basis for the appearance of a space‐operator k in the transduction coefficients for electromagnetic systems in a manner completely symmetrical with the appearance of the time‐phase operator j in electrostatic coupling. Violations of reciprocity in interconnected systems, as discussed by MacMillan and others, can be dealt with analytically in a straightforward way by observing that the operators k and j do not commute, i.e., k j = − j k. Antisymmetrical coupling terms arise in the equations for all‐mechanical systems involving gyroscopes, just as in the electromagnetic case, but symmetry can be restored in these cases also by using the same kind of quaternionlike space‐operator.
22(1950); http://dx.doi.org/10.1121/1.1917175View Description Hide Description
Acoustical calibration of a microphone may be made conveniently either by free‐field calibration or by calibration using a small cavity. Free‐field calibrations are difficult because of the necessity for establishing a space which allows for propagation of a single progressive wave, without any interference. A progressive wave may be set up in a room which will be free from interference at a given point until sound energy arrives by reflection from the nearest wall surface. By sampling the output from a microphone during this period of time, a signal may be obtained which determines the microphone response to a free progressive wave. Periodic repetition of this process gives a series of pulses from the microphone which may be used to obtain a reciprocity calibration.
22(1950); http://dx.doi.org/10.1121/1.1917176View Description Hide Description
A close‐talking, noise‐canceling microphone has been developed which responds to the second order of the pressure gradient and which has only one diaphragm. Since there are four sound pressures involved in a second‐order gradient microphone, it has been deemed necessary in the past to have four surfaces for the four pressures to act upon. This microphone has sound entrances to the two surfaces of a single diaphragm spaced and oriented in such a manner as to produce the second‐order effect, thereby increasing the signal‐to‐noise ratio over that obtained in a first‐order gradient microphone. Mathematical analyses are made of the microphone first as a purely theoretical microphone with infinitesimal spacing of the sound entrances, then as a microphone with dimensions between sound entrances which are practical for use in a microphone of this type.
22(1950); http://dx.doi.org/10.1121/1.1917177View Description Hide Description
A pressure meter of extremely small size for the measurement and recording of absolute pressures was developed. It is in the form of a small tube of 2.5‐mm diameter and 5‐mm length. The tube is closed by a thin iron membrane. The deflection of the membrane causes a change of the reluctance of a magnetic circuit. This reluctance change is transformed into voltages by means of a carrier frequency system. The pressure meter can be built with a maximum sensitivity of about 100 mm Hgpressure per full meter scale deflection. The sensitivity is limited by the attainable deflection of the membrane and the electrical amplification that can be applied which is limited by the “magnetic noise level” and the zero‐point stability. Using thicker membranes pressure meters of this type for higher pressure ranges can be built. The natural frequency of the membrane is higher than the carrier frequency (max. ca. 15 kc). The range of a flat frequency response characteristic is therefore not limited by the natural frequency of the membrane but by the carrier frequency and the damping of the membrane is not very important. With a carrier of 10 kc a flat frequency response characteristic from zero to about 1000 c.p.s. can be obtained. Problems concerning the sensitivity of the pressure meter for temperature variations and for accelerations and the problems of electrical and mechanical qualities and manufacture resulting from the small size of the pressure meter are discussed.
22(1950); http://dx.doi.org/10.1121/1.1917178View Description Hide Description
Recent interest in the use of transient techniques to test loudspeakers springs from the inadequacy of the steady‐state pressure response to specify completely the behavior of a loudspeaker. Tone bursts have shown some promise as a useful type of electrical signal to excite loudspeakers for transient testing. It appears, however, that Heavyside's step function or Carson's unit impulse function (the delta‐function) may reveal more directly the significant characteristics of a loudspeaker. The response of loudspeakers to each of these latter functions has been examined and typical radiation spectra have been determined experimentally. The results of these analyses seem to yield a substantial amount of useful information about the performance of loudspeakers.
22(1950); http://dx.doi.org/10.1121/1.1917179View Description Hide Description
The term personal radio receiver is used to designate a complete radio receiver with self‐contained power supply, and of such physical dimensions that it can be easily carried by hand or in the pocket. The performance and compactness of personal radio receivers are limited by the efficiency with which electrical power is converted into sound power by the loudspeaker. Since the electrical power output is limited in the personal receiver, the efficiency of the loudspeaker is an important factor. A number of loudspeaker systems have been investigated, both theoretically and experimentally, as follows: direct radiator, combination direct radiator and phase inverter, horn, and combination horn and phase inverter. An efficiency of 25 percent has been obtained with the combination horn and phase inverter. This loudspeaker system has been incorporated in a complete four‐tube radio receiver, having a content of 25 cubic inches.
22(1950); http://dx.doi.org/10.1121/1.1917180View Description Hide Description
The motion of a reproducer stylus in a phonograph groove has been subjected to a mathematical analysis which includes the joint effects of tracing distortion and elastic deformation of the groove wall. The lateral motion of the stylus at fundamental frequency is obtained as a product of three factors: (1) A stylus‐groove response function which increases slowly to a major peak occurring at resonance between the compliance of the groove walls and the effective mass of the stylus, and then falls off rapidly; (2) a translation loss which is small at low frequencies but becomes complete at a cut‐off wave‐length determined by the stylus radius, the tracking force and the elastic properties of the record material; and (3) a scanning loss which arises when the dimensions of the contact surface between stylus and groove wall become significant in comparison with the recorded wave‐length. It is found to be still possible to calculate the harmonic and intermodulation distortion products on the basis of rigid‐wall tracing distortion theory provided the amplitudes of the predicted distortion terms are multiplied by both the stylus‐groove response function and the scanning loss function. Non‐linear effects of deformation turn out to be negligible in comparison with tracing distortion. Measurements made with a series of recordings programmed to exhibit these effects confirm the analysis in all important respects.
22(1950); http://dx.doi.org/10.1121/1.1917181View Description Hide Description
Pertinent aspects of the theory of square‐wave testing will be discussed, and data will be presented showing how certain facts concerning the over‐all response‐frequency characteristic and equalization may be determined, as well as resonances in various parts of the system. Given an adequate recording system, a short cut of “square waves” on the outside of each record provides the user with a rapid means of determining whether or not his playback system is properly equalized and otherwise adequate for reproducing the recorded material.
- Session C: Hearing
22(1950); http://dx.doi.org/10.1121/1.1917182View Description Hide Description
Electrophysiological responses from the auditory systems of cats show phenomena of refractoriness, that is, reduced responses to the second of a pair of successive stimuli. Thus, when two acoustic clicks are delivered in close succession (intervals from 0.2 to 500 msec.), the response to the second click varies with (a) the time interval between the clicks, and (b) the intensities of the clicks. At the round window, the microphonic components of the click response shows no refractoriness. The neural components exhibit marked recovery cycles; for moderate stimulus intensities, these may last as long as 100 msec. At the auditory cortex, the recovery cycles may be even longer and may exhibit several marked phases of sub‐ and supernormality. The cortical recovery cycles are affected by the spontaneous electrical activity of the cortex, by the level of anesthesia, and by the temperature of the animal. While some interaction between the two ears may be shown at the level of the auditory nerve, this interaction appears strikingly at the cortex.
22(1950); http://dx.doi.org/10.1121/1.1917183View Description Hide Description
The investigation of hearing by the use of a fatiguing tone and an immediately subsequent short‐duration test tone, first used with good success by de Maré and later adapted for special problems by Lüscher and Zwislocki, Gardner, and Munson and Gardner opens a wide field for study of suprathreshold events in the ear. Among the many variables are the frequency, intensity, and duration of each of the two tones, the frequency relations between the two tones, and the recovery interval, if any, between them. Most of these variables have not been studied independently. The present paper explores fatigue as a function of (1) duration and intensity of the fatiguing tone, and (2) duration of recovery interval. Both variables will be considered with respect to the frequency of the fatiguing tone.
22(1950); http://dx.doi.org/10.1121/1.1917184View Description Hide Description
“Tone‐pips” with a basic frequency of 2000 c.p.s. were produced by passing rectangular pulses of duration through two sound‐effects filter sections in cascade. The high pass and the low pass cut‐offs (18 db per octave) of each unit were set to 2000 c.p.s. The transducer was an Atlas PM 25 loudspeaker. The tone‐pips appear on the oscilloscope as brief trains of waves that begin gradually and reach maximum amplitude at the third wave. The amplitude then falls off only a little less rapidly. A single pip sounds like a metallic click. When the pulses were repeated at 123 per second at sensation levels from 5 to 40 db all of our subjects reported hearing a “buzz.” Other descriptive terms used were “metallic,” “high pitched,” “continuously interrupted,” and “rough.” No listener, even when directly questioned and even when a pure tone of 123 per second had been sounded a few seconds previously for comparison, ever reported any trace of a low pitched component corresponding to 123 per second. Inasmuch as the frequency of nerve impulses in each nerve fiber is 123 per second under these conditions, the absence of a sensation of low pitch is contrary to the frequency theory of pitch perception and favors the place theory even for very low tones.
A Consideration of the Intensity‐Loudness Function and Its Bearing upon the Judgment of “Tonal Range” and “Volume Level.”22(1950); http://dx.doi.org/10.1121/1.1917185View Description Hide Description
Acoustical intensity not only affects the loudness of sounds, but also profoundly influences the listener's perception of certain ranges of frequencies. The data of Fletcher and Munson demonstrate that the effective frequency response of the ear varies with signal intensity‐level. On the basis of this variation, it is possible to account for certain anomalies appearing in Eisenberg and Chinn's study of listeners' preference for frequency ranges and intensity levels in the reproduction of speech and music. It is also possible to explain the disparity between their results and those of Olson's investigation of preference for frequency ranges. For example, it can be shown that when frequency is plotted against loudness, raising the intensity level from 50 to 70 db will add more than one whole octave downward to the effective frequency response of the ear at 50 millisones loudness.
22(1950); http://dx.doi.org/10.1121/1.1917186View Description Hide Description
In an earlier paper data were reported indicating there was no difference between minimum audible field and minimum audible pressure at low frequencies. No satisfactory explanation was available at that time as to why these results differed from previously published data except the statement that body noises, breathing, etc., had been minimized. Further work has now produced a satisfactory explanation. Furthermore, if a well‐sealed volume of 6 to 7 cm3 is maintained invariant over the eardrum and the mechanical isolation of the signal source progresses from good to bad, the difference between MAF and MAP varies from 0 to about 16 db. Data have been obtained for deviation of MAP from MAF as the volume and mechanical isolation are kept constant, the seal being broken by measurable amounts. Finally, to study the effect for the smallest possible volume, the plugged ear was fed acoustically from a high impedance source that was well‐isolated mechanically. With this arrangement the results indicate that for the 2‐cm3 volume the difference between MAP and MAF can be made less than 4 db by carefully controlling the subject. This control involves the position of the teeth, tongue, arms, etc.
22(1950); http://dx.doi.org/10.1121/1.1917187View Description Hide Description
The middle ear of rabbits was photographed during exposure to high intensity sounds. Extreme amplitudes of the eardrum were observed and measured. Rupture of the eardrum occurred when the signal was increased further (motion picture film). The protective devices of the ear against intense signals are discussed. The reflex of the middle earmuscles is demonstrated. Tetanic contractions of both muscles dampen the movements of the ossicles. Another mechanism which most likely acts as protection for the vulnerable inner ear is the change of axis of the stapes at high intensity signals. The stapes movements were studied in fresh temporal bones. A small mirror was attached to the vestibular side of the footplate. Acoustic stimulation was carried out through the outer ear canal. The responses are asymmetrical. The negative amplitude is greater than the positive or sometimes negative only. Signals increasing in intensity produce a gradual change of the axis of vibrations. It was also observed that the stapes may suddenly change the axis of vibration as described previously by v. Békésy. The change of axis is interpreted as a protective device. Other features of the ear which act as protection against overstimulation are discussed.