Volume 34, Issue 12, December 1962
Index of content:
- PROGRAM OF THE SIXTY‐FOURTH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session B. Electroacoustics
- Contributed Papers
34(1962); http://dx.doi.org/10.1121/1.1936955View Description Hide Description
A simple method for the prediction of the stiffness‐controlled voltage response of a hydrophone configured for deep‐sea operation is described. The approach utilizes previous work in this area, enlarged to include the effects of various degrees of decoupling of the acoustic field from the back face of the hydrophone. Two configurations of importance, the spherical and the cylindrical, have been analyzed, and the results presented graphically. For both cases, the equations of equilibrium and the equation of compatibility are solved and evaluated by the use of boundary conditions. These stress equations, thus generated, and the electromechanical constants are then related to yield the voltage/pressure sensitivity equations. The graphs show the effects of thickness to diameter, comparison of the air‐backed to the pressure‐equalized hydrophone, and the various degrees of decoupling of the acoustic field from the back face of the hydrophone. For all cases, both bariumtitanate and lead zirconate titanate are considered.
34(1962); http://dx.doi.org/10.1121/1.1936956View Description Hide Description
Prior to the writing of this paper, it has been difficult to determine an accurate equivalent circuit for certain piezoelectric transducers, particularly for disks 3 in. in diameter and from 0.1 to 0.25 in. thick. It was discovered that the methods currently in use for finding a steady‐state equivalent circuit for these disks, yield results that are highly inaccurate, and, therefore, useless. In all cases, these disks were tested in the thickness‐resonance mode. An equivalent circuit for a transducer has approximately the same impedance as the transducer over a narrow band of frequencies near the mechanical resonance of the transducer. Near the mechanical resonance of the transducer, the impedance locus in the complex plane is circular. The problem can be approached as a 2‐terminal network‐synthesis problem. The equivalent circuit consists of linear, bilateral, passive elements, and is derived from the impedance circle. The mechanical‐resonant frequency can be found from the equivalent circuit. All of the transducers tested had prepolarized bariumtitanate or lead zirconate titanate elements, were pressure‐release mounted, and were operated in the linear region without bias. [This work was sponsored by the Bureau of Ships under Contract No. NObsr‐72627.]
34(1962); http://dx.doi.org/10.1121/1.1936957View Description Hide Description
Beam pattern and directivity index are computed for spherical volume and spherical‐shell distribution of receiving elements. The limiting case of a continuous distribution of elements gives rise to simple closed‐form expressions for the beam patterns. Directivity index is obtained by integration of the beam pattern. The side‐lobe level of the beam pattern for the solid sphere is appreciably lower than that of the corresponding spherical shell. Comparison of directivity index shows an increasingly superior performance for the solid sphere as the acoustic radius ka becomes large. [This presentation represents results of research sponsored by the Office of Naval Research.]
Investigation of Computational Techniques for Prediction of Far‐Field Radiation Pattern of a Transducer from Near‐Field Measurements34(1962); http://dx.doi.org/10.1121/1.1936958View Description Hide Description
The far‐field acoustic‐radiation pattern of a large single‐frequency cylindrical transducer was computed, using various methods, employing data taken in the near field of the transmitter. Beam patterns thus computed are compared to the measured far‐field beam pattern from the same transmitter. Two ground rules were imposed on this study: (1) No assumptions would be made as to any relationship between the pressure and its gradient which cannot be justified theoretically. (2) Any approach which led to the equivalent of the spheroidal wavefunction method would not be considered further, unless the formulation was such that the computation time involved was considerably less. The Kirchhoff‐Helmholtz formula, using both the measured pressure and pressure gradient, gave the best agreement with the measured far‐field directivity. Poorer agreement was shown when the wave equation was approximated by a difference equation, using the measured pressure and its gradient as boundary conditions, and calculating the pressure step by step from the near field to the far field. Other methods employed for calculating the far‐field pressure were found to be inadequate.
Computation of Far‐Field Characteristics of a Transducer from Near‐Field Measurements Made in a Reflective Tank34(1962); http://dx.doi.org/10.1121/1.1936959View Description Hide Description
The directivity pattern and source level of a large, cylindrical transducer were computed by using Kirchhoff's formula with a simple approximation for the normal pressure gradient. The integration was performed numerically over a finite cylinder. Near‐field pressure amplitude and phase measurements were made over the surface of integration by rotating the transducer and measuring the output of a line hydrophone less than 1 wavelength from the active face. The measurements were made on the leading edge of the received pulses—before steady‐state conditions were reached. This was done to simulate this transducer operating in a tank of diameter only about twice that of the transducer. Under such conditions, interference due to tank‐wall reflections would begin when the amplitude of the pulse envelope was as much as 2–3 dB below the steady‐state level. Computed directivity patterns agreed well with equivalent measured ones, but, naturally, computed source levels were 2–3 dB low. A correction term was derived from the first portion of the pulse risetime curves, and was applied to the source‐level values to give good agreement with measured ones. The present results indicate that accurate transducercalibration can be obtained in reflective tanks considerably smaller than those used in previous work.
34(1962); http://dx.doi.org/10.1121/1.1936960View Description Hide Description
The technique of pulse compression by the use of a linear frequency‐modulated signal in conjunction with a delay network as detector is well known. If the duration of the signal is T and the instantaneous frequency varies from f 0 to f 1, the slope of the delay network must be −T/(f 1−f 0). Since the Doppler effect keeps the product T(f 1−f 0) constant, the quotient varies as the square of the Doppler shift, and so this signal is only usable if the Doppler shift is small. This paper describes a modified chirp signal and its matched filter detector that are not subject to this limitation. It is shown that the signal x(t) defined by the equation possesses the property that the ratio between any 2 adjacent intervals enclosing m cycles of the signal is a constant. It is further shown that a delay line with taps spaced in a geometrical progression is a matched filter for detecting this signal. The matched filters for Doppler‐shifted signals may then be formed by extending the progression at each end and connecting together the same number of adjacent taps.
- Session C. Psychoacoustics I
34(1962); http://dx.doi.org/10.1121/1.1936961View Description Hide Description
A series of experiments designed to assess the limit of a human observer's level of sensitivity is reported. A mathematical theory of the detection process is introduced which explicitly considers the sequential effects observed in psychophysical data. This variable‐threshold model for signal detection is an axiom system based on a stimulus‐sampling theory of the learning process. The theory generates predictions for all aspects of an observer's response protocol, e.g., mean response probabilities, associated variances, sequential predictions such as autocorrelations, runs, etc. It thereby permits a detailed treatment of individual trial‐by‐trial data. Experimental data are presented for both yes‐no and forced‐choice auditory‐detection situations, with and without information feedback to the observer. The model is compared to signal‐detection theory [Tanner and Swets (1954); Green (1960)] and to another threshold model [Luce (1961)].
Application of an Energy‐Detector Model with Internal Noise to the Detection of Signals in the Presence of Pedestals at Varying Noise Levels34(1962); http://dx.doi.org/10.1121/1.1936962View Description Hide Description
An energy‐detector model for monaural auditory signal detection which incorporates noise sources assumed to arise within the observer in addition to external noise is described. Its detection performance for 2‐interval forced‐choice experiments was calculated assuming constant internal noise and noise proportional to the square of the mean detector output. To test its applicability, data from human observers were obtained at noise levels of 60 and 80 dB SPL, for a single ratio of signal energy‐to‐noise spectral density in combination with various pedestals (sinusoids of equal amplitude present in both stimulus intervals). The pedestals were 90° out of phase with the signal. Results indicated that there was no significant amount of internal noise which is constant for different noise levels, and that the detectability calculated for the model under the assumption of added noise proportional to the square of the mean output of the detector conformed to the data reasonably well. Additional experiments with in‐phase pedestals at noise levels 20 and 50 dB apart confirmed the absence of significant constant internal noise, but indicated the need of further modifications of the model to fit the rising portion of the curve, some of which are discussed.
34(1962); http://dx.doi.org/10.1121/1.1936963View Description Hide Description
The masking produced by 2 continuous sinusoids located equidistant in frequency from the signal, a pulsed sinusoid of , was studied. The 2 sinusoidal maskers were of equal energy. Their separation in frequency Δf was varied. The main parameter of the experiment was the frequency of the pulsed signal. The results can conveniently be summarized in one graph. Along the abscissa, we plot the frequency separation of the maskers (Δf). Along the ordinate, we plot the intensity of the pulsed signal (Is ) needed for some constant level of detectability. As the frequency separation of the maskers increases, the curve descends, indicating the pulsed tone is becoming easier to hear (Is ∝1/Δf). For moderate frequency separations (Δf<100 cps), the results are independent of center frequency. For example, essentially the same signal intensity in SPL is needed at 250, 1000, or 4000 cps to overcome the masking produced by 2 continuous tones when these maskers are located 20 cps above and below the frequency of the signal. At much wider frequency separations, the results are dependent on center frequency, and a critical band of a kind may be measured. The width of the “critical band” inferred from this experiment is about the same size as the half‐response points of the curve relating the amplitude of the traveling‐wave envelope to frequency; i.e., the “critical band” is nearly 600 cps wide at 1000 cps. [This work was supported in part by the U. S. Army Signal Corps, the U. S. Air Force ESD Contract AF19(604)‐7459, the Office of Naval Research, and in part by the National Science Foundation (Grant G‐21807).]
34(1962); http://dx.doi.org/10.1121/1.1936964View Description Hide Description
The discrimination of the rate of onset of a white noise was examined for exponential onsets. Upon a given trial, either one of two onsets was presented to the listener. His task was to vote whether a given onset was “fast” or “slow.” On successive trials, the shorter onset time was held constant. The longer onset was varied to achieve 75% correct discrimination. The average ratio of the longer to the shorter onset was determined for each of 5 listeners. In terms of the interval between 10%–90% of asymptotic level, the average ratio varied from about 4.5 to about 1.5 as the shorter onset time varied from 1 to 500 msec. Over the range of onsets from 5–100 msec, the ratio was about 2.0. That is, an increment of 100% in the onset of a white noise is required over a wide range of onsets for 75% correct discrimination.
34(1962); http://dx.doi.org/10.1121/1.1936965View Description Hide Description
A high‐pass filtered noise will increase the difference limen (DL) for frequency for a simultaneously presented low‐frequency, unmasked sinusoid. Since only the basal end of the basilar membrane should be interfered with by such filtered noise, this would seem to indicate the participation of the basal end in the perception of a low‐frequency sinusoid, contrary to a traditional place theory. These data are from 20 normal‐hearing listeners making either unilateral, alternate‐pitch matches or constant‐method judgments for a 150‐cps sinusoid at a 5‐ to 50‐dB sensation level (SL). Two interfering noises, one periodic and one aperiodic (either unilateral or contralateral to the sinusoid), were presented at the maximum level possible without masking the sinusoid at its threshold. Both noises contained essentially no energy below 400 cps. The frequency DL's were increased approximately 20% by all the noises, regardless of ear of presentation. The increase would thus appear to be attributable to a general, bilateral phenomenon, such as distraction, rather than the interference with the basal end of the ear receiving the sinusoid. [This work is based on a Ph.D. thesis completed at the State University of Iowa under the direction of Dr. Arnold M. Small, Jr., and was supported in part by the National Science Foundation.]
34(1962); http://dx.doi.org/10.1121/1.1936967View Description Hide Description
This paper supplements work of Garner, Eisenberg, and Fodor in studying thresholds of pure tones varying in duration among subjects with normal and impaired hearing. A rise‐fall time of 1 msec was used for all stimulus durations and frequencies. The change in absolute threshold per log change in tone duration differed among frequencies for normal listeners, with lower frequencies producing larger slopes. For listeners with impaired hearing, the point at which threshold ceased to change with increasing tone duration (i.e., the “critical duration”) occurred at shorter tone durations than among normal listeners. The amount of “threshold shift” for a 3‐msec tone (the difference between the threshold for the 3‐ and the 1000‐msec tones) was inversely correlated with the subject's threshold for a 1000‐msec tone of the same frequency; correlation coefficients of −0.74 (500 cps), −0.74 (1000 cps), and −0.79 (4000 cps) were obtained when all listeners (normal and impaired hearing) were included. Results are considered in relation to stimulus characteristics and critical‐band theory.
34(1962); http://dx.doi.org/10.1121/1.1936968View Description Hide Description
Laboratory studies were undertaken: (1) to validate, if possible, equal noisiness contours for bands of noise and to extend the range of these contours to higher sound frequencies; (2) to determine the effects of the presence of intense pure‐tone components in wide‐band noise upon judgments of loudness and noisiness; and (3) to equate the effects of rise time, duration, and decay time of a sound upon its judged noisiness. The data indicate that: (1) Presently published equal‐noisiness contours have the proper shape, or nearly so, up to about 6000 cps. An extension of the contours is proposed for use with sounds having higher frequency components. (2) The presence of strong pure‐tone components in a background of broad‐band noise has a much more dramatic effect on judged noisiness than judged loudness. (3) It appears, from a preliminary study, that the trading relation between over‐all duration ( to 20 sec) and peak sound‐pressure level, keeping rise‐ and decay‐time ratios constant is fairly well represented by the “equal energy” rule—i.e., halving the total duration reduces judged noisiness to the same degree as a 3‐dB reduction in sound‐pressure level. The relation of these findings to procedures for the calculation of perceived noise level is discussed. [This work was jointly supported by the National Aeronautics and Space Administration and by the Federal Aviation Agency, under Contract NASr‐58.]
34(1962); http://dx.doi.org/10.1121/1.1936969View Description Hide Description
Carefully presented instructions are a basic requirement of successful psychoacoustical scaling studies. The term in the instructions which refers to a subjective experience uniquely defines the type of scale constructed. This paper describes the degree to which a psychoacoustical scale would be modified by changes in instructions. The results of 2 studies are reported. The first discusses the differences in subjects' responses to instructions based on 3 different key words—loudness, noisiness, and annoyance. The second is concerned with the effect on subjects' responses of mixed instructions (i.e., where more than one key word appears in a single set of instructions). The results indicate that different instructions “set” subjects to give different emphasis to various physical aspects of sound, and that instructions based on a combination of 2 key words yield ratings significantly different from ratings given to the same sounds when the key words are used individually. A recommendation is made that instructions be selected which set subjects to emphasize sound dimensions relevant to particular noise problems.
- Session D. Underwater Acoustics: Oceanography, Sound Velocity, and Absorption
34(1962); http://dx.doi.org/10.1121/1.1936970View Description Hide Description
Interpretation of the principle features of sound velocity in terms of the basic oceanographic parameters—temperature, salinity, and depth—is undertaken using the functional relationship between the velocity of sound and the temperature, salinity, and pressure of the sea water through which the sound is propagating as expressed by Wayne D. Wilson of the U. S. Naval Ordnance Laboratory. The effect of independent variation of these parameters on the sound‐velocity profiles is treated mathematically for a uniform and a stratified ocean. The slope, slope reversals, kinks, and minimums in the sound‐velocity profiles are related to environmental conditions. A comprehensive discussion of the causes of the long‐ and short‐term changes in the features of the profiles is included. The use of sound‐velocity data to infer primary oceanographic parameters is demonstrated using the temperature‐salinity relationships.
34(1962); http://dx.doi.org/10.1121/1.1936971View Description Hide Description
In situ measurements using two Bureau of Standards velocimeters have been obtained on several deep dives. Preliminary results indicate an agreement with previous in situ measurements aboard Bathyscaph TRIESTE to 6000‐m depth. Sound speeds are less than Kuwahara below 6000 m. The difference becomes more negative with depth, and reaches a value of −2 m/sec at 9000 m. This is different from the pressure effect predicted by the Wilson equations. Dives will conclude early in August, and a final precision calibration will be made. It is expected that the analysis will be complete by the time of presentation.
34(1962); http://dx.doi.org/10.1121/1.1936972View Description Hide Description
The dynamic behavior of a gas or liquid is determined by the conservation equations for mass, momentum, and energy, together with the equation of state. These equations with certain restrictions have been investigated in some detail. The 3 cases that have been considered in solving these equations for the physical parameters of the medium are: (1) no energy source put into the medium, (2) a deterministic energy source put into the medium, and (3) an undeterministic energy source in the form of Gaussian white noise put into the medium. In the 1st case, an expression for the temperature distribution as a function of boundary conditions is obtained. In the 2nd and 3rd cases, a statistical solution is obtained. The 2nd case leads to an expression for the expected value of the temperature fluctuations, and the 3rd case gives expressions for the physical parameters of the medium in terms of the Wiener statistics.
34(1962); http://dx.doi.org/10.1121/1.1936973View Description Hide Description
The attenuation of acoustic energy in the ocean at a frequency of 40 kc was measured by means of a bottom‐mounted array of 7 hydrophones in 1800 ft of water. The source of short pulses was a ship‐mounted projector. Measurements were made by determining the peak amplitude of the received pulses as a function of range. Spherical transducers were used to minimize the effect of patterns on the measurements. These measurements agreed with the unpublished values obtained by APL, Washington. In addition, the ambient sea noise as a function of sea state was measured. The values obtained, when corrected for the attenuation over the depth of 1800 ft, agree with published values within from sea state 3 down to below sea state 1, where system noise became predominant.
34(1962); http://dx.doi.org/10.1121/1.1936974View Description Hide Description
The salinity and temperature dependence (the latter being significant mainly at lower frequencies) of the quantity A in Lieberman's equation for sound absorption in the ocean bare been obtained by the use of data contained in publications by Kurtze and Tamm and by Wilson and Leonard. Using the values for pure water absorption βf 2 and for the power relaxation frequency f 0 used by Del Grosso, the predicted absorptions for various frequencies f agree with the careful measurements of Murphy, Garrison, and Potter made in Dabob Bay in 1958.
- Session E. Architectural Acoustics: Auditorium Acoustics
34(1962); http://dx.doi.org/10.1121/1.1936975View Description Hide Description
Philharmonic Hall, part of the new Lincoln Center for the Performing Arts in New York City, seats 2644 and has a cubic volume of 865 000 cu ft. It is designed exclusively for symphonic concerts, although during its early years it will serve a variety of purposes. This paper describes the acoustical design, the activities of “tuning week,” 28 May to 2 June 1962, and the experiences during the first month of its operation following its dedication on 23 September 1962. Data taken during “tuning week” indicate that its midfrequency reverberation time will be approximately 2.0 sec, and the ratio of reverberation at low frequencies to that at midfrequencies approximately 1.25.