Volume 44, Issue 1, July 1968
Index of content:
- PROGRAM OF THE SEVENTY‐FIFTH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session A. Engineering Acoustics I: Elastic Obstacles, Diffraction, Transmission and Finite Fields
- Contributed Papers
44(1968); http://dx.doi.org/10.1121/1.1970147View Description Hide Description
Following the theory describing circumferential waves (presented in the preceding paper), an experimental investigation was undertaken to observe and measure these waves. The Rayleigh‐type surface wave has been observed on elastic solid cylinders immersed in water. This wave was excited by an acoustical pulse of 10‐μsec duration projected onto the cylinder at the excitation angle given by sin−1 (c 0/c), where c is the velocity of the circumferential wave and c 0 is the velocity of sound in water, and has been observed to make several circumnavigations of the cylinder. The waves can be detected since they continuously radiate acoustical energy into the water and have been observed on pure aluminum, 304 stainless steel, and various aluminum‐alloy cylinders. Since the surface curvature gives rise to dispersion, cylinders of 3.5‐ to 6‐in. diam were used and the frequency of the incident pulse was varied from 0.15 to 3.5 MHz. This corresponds to a radius‐to‐wavelength ratio of 10 to 400. Experimental and theoretical results will be presented and shown to compare favorably. Inhomogeneities in brass and in aluminum alloys are found to correlate with poor propagation characteristics for some of the cylinders. Examples will be shown.
44(1968); http://dx.doi.org/10.1121/1.1970149View Description Hide Description
The classical, modal formulation of the vibrations and radiation loading of submerged elastic shells [S. Hayek, J. Acoust. Soc. Am. 40, 342–348 (1966)] requires extensive calculations at high frequencies. An alternative formulation in terms of traveling waves familiar from studies of plane‐wave scattering by elastic cylinders and spheres [see bibliography, Doolittle et al., J. Acoust. Soc. Am. 43, 1–14 (1968)] is applied to thin spherical shells excited by radial point forces. The response is expressed as the sum of (1) a primarily structure‐borne field consisting of three modes, two of them rapidly attenuated, characterized, except for membrane effects, by the wave‐numbers of point‐excited flat plates in water; (2) a diffracted field, resembling “creeping waves” on “pressure‐release” spheres. One structure‐borne and two “creeping‐wave” modes suitably represent shell response and surface pressures except in the immediate vicinity of the drive point. [Work sponsored by Office of Naval Research, Structural Mechanics Branch.]
44(1968); http://dx.doi.org/10.1121/1.1970151View Description Hide Description
The formalism of the creeping‐wave theory has been extended to enable calculations of various incident pulse forms. The particle velocity on the cylinder surface and the pressure radiated to a far‐distant observer have been obtained for various pulse shapes. The incident pulse shapes studied are: (1) the delta function, (2) finite modulated step pulse, and (3) finite unmodulated step pulse. The growth curves of particle velocity and pressure have been calculated as a function of a reduced time τ. In addition, the creeping‐wave speed and attenuation angle are given for the case of the delta pulse and the finite modulated step pulse. [Work supported by the Naval Ship Research and Development Center and the Office of Naval Research.]
44(1968); http://dx.doi.org/10.1121/1.4769798View Description Hide Description
The theory of circumferential waves has been extended for large cylinders (ka≫1) in the manner developed by Grace and Goodman [J. Acoust. Soc. Am. 39, 173–174 (1966)]. Numerical evaluations of the secular determinant have been made. The results can be checked easily with the limit ka→∞. Certain difficulties arise in the asymptotic expansions of Bessel functions, which depend on whether the ratio of circumferential velocity to shear velocity is greater than or less than unity. Values of the circumferential wave velocity and attenuation have been computed for steel and aluminum for values of ka pertinent to the experimental data, which will be discussed in the following paper.
Experimental Observation of Three Types of Circumferential Surface Waves on Aluminum Cylinders in Water Using Pulses44(1968); http://dx.doi.org/10.1121/1.1970153View Description Hide Description
Three types of circumferential waves that contribute to the total diffracted acoustic field of an elastic cylinder are isolated by means of a short pulse emitted from a narrow‐beam transducer. These waves are often called “creeping waves”; one called the “Franz type” is primarily related to the geometry of the diffracting body, whereas the other two are more related to the elastic properties of the cylinder. The wave velocities are determined from the pulses observed in the diffracted field of the cylinders whose size parameter ka=323. Schlieren photographs are presented showing the progression of the diffracted wavefronts that result from the boundary waves.
44(1968); http://dx.doi.org/10.1121/1.1970155View Description Hide Description
The scatteredpressure field examined is that produced when a plane wave pulse is incident upon an ideally rigid sphere immersed in a fluid. The classical normal‐mode series expresses the scattered field when a harmonic plane wave is incident upon the sphere. By using this series solution in conjunction with an incident pulse of known spectral content, it is possible, by numerical means, to synthesize exactly the resulting reflected pulse at an arbitrary field point. In the present instance, truncated sinusoidal incident pulses are considered, and scattered pulses are determined both at points close to the sphere and points distant from it. The time‐varying amplitudes of these reflected pulses, thus rigorously calculated, are discussed in terms of a model based upon the existence of circumferential or “creeping” waves propagating along the periphery of the sphere. This model considers the reflected pulse to be composed from a specularly reflected replica of the incident pulse and attenuating pulses that travel along the periphery of the sphere from the boundary of the geometric shadow to the observation point.
44(1968); http://dx.doi.org/10.1121/1.1970158View Description Hide Description
The Fresnel‐Kirchhoff theory for optical diffraction around the edge of a semi‐infinite screen offers a means of estimating the diffraction leakage around the edge of an underwater‐sound shielding baffle. To test the utility of the Fresnel‐Kirchhoff theory for this application, experiments were conducted using a Laboratory tank and a frequency of 93 kHz. The baffle consisisted of a ‐thick piece of closed‐cell, cellular rubber bonded to a ‐thick aluminum plate. To explore the sensitivity of the measureddiffraction to experimental conditions, both straight‐ and curved‐baffle edge geometries were used; various inclinations of the baffle to the sound field were also tried. It was concluded that the Fresnel‐Kirchhoff theory offers an accurate prediction of diffraction leakage around an underwater‐sound baffle, even though the actual baffle construction and geometry may vary somewhat from the ideal model for which the theory was developed.
44(1968); http://dx.doi.org/10.1121/1.1970160View Description Hide Description
A series expansion method for determining the farfield from nearfield measurements of coherent and partially coherent sound fields is presented. Current investigations have generally been based on a Helmholtz integral or Green's‐function approach, although one of the earliest methods for sinusoidal oscillation [J. Pachner, J. Acoust. Soc. Am. 28, 86–89 (1956)] was actually based on a series expansion of the nearfield solution in spherical wave‐functions. By confining the measurements to a spherical surface, Pachner was able to evaluate the expansion coefficients through the orthogonality of the Legendre polynomials. However, in the method presented here, the expansion is truncated to include only the number of terms equal to the number of nearfield measuring points. In this way, the coefficients may be evaluated through a matrix inversion for nearfield measuring points at any location. A comparison of exact and predicted farfields is presented for some simple source distributions. The method has been extended for crosspower nearfield spectral measurements. [Work supported by the Office of Naval Research.]
44(1968); http://dx.doi.org/10.1121/1.1970162View Description Hide Description
Previously, the author [J. Acoust. Soc. Am. 43, 709–715 (1968)] showed that the correct wave equation needed to describe the propagation of sound in an acoustic Luneburg lens is , where p is the acoustic pressure, ρ the density, Co a constant reference speed, and is Luneburg's index of refraction. Here a is the lens's radius and r any radial point within it. A similar equation holds for the electromagnetic Luneberg lens with the dielectric coefficient ε replacing ρ. However, and consequently, it can be only one possible function of position. This is not so for the acoustic case, since (κ the compressibility), and κ can be a function of position. Previously, we investigated the case where ρ was approximately constant, so we could neglect the density‐gradient term in the wave equation. (Strictly speaking, ρ cannot be constant if n is a function of position.) This case corresponds to W. J. Toulis' compliant‐tubing lens, and we found that the theory agreed well with experimental data. In this paper, we examine a different possibility, viz., κ constant and ρ a function of position given by . We then compare the two cases.
44(1968); http://dx.doi.org/10.1121/1.1970164View Description Hide Description
This paper introduces a method for measuring the transmission loss and phase distortion that an acoustic signal undergoes in passing through plates. The physical setup consisted of two hydrophones on either side of a steel plate and an underwater speaker at a “quiet” test facility. Their “white‐noise” signals were recorded simultaneously on an FM tape recorder. The tape recordings, filtered from 100 to 9000 Hz to avoid aliasing, were digitized at 18 000 samples/sec, introduced into a UNIVAC 1107 computer, and the cross‐correlation functions obtained. The cross correlation between the speaker and hydrophone separates the signal coming along various paths according to their time of arrival. The section of the correlogram containing the incident signal is considered, while all other values are set to zero; then the power and phase associated with this section are determined. The section containing the transmitted signal is given similar consideration. The incident power minus the transmitted power gives the transmission loss. Mathematical description of the transmission loss is found to be in good agreement with these experimental results.
- Session B. Psychological Acoustics I
- Invited Paper
44(1968); http://dx.doi.org/10.1121/1.1970166View Description Hide Description
Laws stating that TTS varies as the logarithm of time disagree with experimental results very early or very late in these processes. Better agreement with observations results from regarding TTS as the sum of four components that vary exponentially with time. During constant stimulation, each component increases, approaching an upper limit asymptotically. After stimulation is terminated, each component diminishes toward zero asymptotically. Using data from published TTS studies, the magnitude and time constant of each component was determined. One component grows toward a limiting value dependent on stimulation intensity and test frequency, gaining half the remaining magnitude every 45 min; and decays afterward so that the magnitude is halved every 3 h. Another component grows to about 13 dB in most exposures and, afterward, decays to half‐value every 8 min. Two more components, one positive and the other negative, decay at rapid but different rates to produce the “bounce” often observed during the first 2 min of recovery. Only the first‐mentioned component seems related to permanent noise‐induced hearing loss. The simple electrical analog of the theory makes possible a simple TTS meter for appraising noise hazard.
- Contributed Papers
Relationships between Several Indices of Acoustic Reflex Function and Susceptibility to Temporary Threshold Shift Produced by Different Spacings of Impulsive Noise and by Continuous Noise44(1968); http://dx.doi.org/10.1121/1.1970168View Description Hide Description
Temporary threshold shift (TTS) was measured following exposure, on separate days, to 110 dB thermal noise or to 20 successive 166‐dB‐peak, 58‐μsec‐duration, spark‐gap‐produced impulsive noises, spaced 0.2, 2.0, or 10.0 sec apart. Pearson correlations were calculated between these measures and estimates of acoustic intratympanic reflex thresholds, durations, rise times, and decay times, obtained with a Madsen impedance bridge and averaged over 10 trials at each of three eliciting frequencies on an averaging computer. There were indications of negative correlations of reflex duration and TTS and of positive correlations of reflex threshold and TTS, especially at the 2.0‐sec impulse spacing.
44(1968); http://dx.doi.org/10.1121/1.1970170View Description Hide Description
A group of 18 uniaural chinchillas was exposed for 2 h to a 2‐oct‐wide band of noise having an over‐all SPL of 124 dB and band limits of 750 and 3000 cps. Most animals were untestable (i.e., had thresholds in excess of 110 dB SPL) at one or several test frequencies immediately after exposure. The median shift 1 day later was about 115 dB at 1, 2, 4, and 8 kc/sec, and recovery proceeded slowly for the following 3 mo., reaching a final permanent threshold shift (PTS) on the order of 50 dB. Correlations were calculated between these values of PTS and various tests of temporary threshold shift (TTS) performed on the animals before the high‐intensity exposure. Implications of these results in regard to tests for susceptibility to noise‐induced hearing loss in humans will be discussed. [Research supported by the U. S. Public Health Service.]
44(1968); http://dx.doi.org/10.1121/1.1970172View Description Hide Description
Puzzling results from two previous studies of binaural fatigue, one involving dichotic exposure [W. D. Ward, J. Acoust. Soc. Am. 39, 1262(A) (1966)] and the other binaural phase effects [W. Melnick, J. Acoust. Soc. Am. 42, 179–184 (1967)] indicated a need for repetition. For the dichotic experiment, 24 listeners were exposed for 5 min to either 700 Hz at 115 dB SPL or 2.4–4.8‐Hz noise at 112 dB monaurally and to both dichotically. Dichotic exposure produced a reduction in TTS at 4, 6, and 8 kHz for the ear exposed to the high‐frequency noise similar to that seen in the earlier experiment. In the binaural phase experiment, 15 listeners tracked TTS for 6000 Hz following 2 min of exposure at 4000 Hz, 120 dB SPL in three sessions under three conditions: monaural, binaurally in phase, and binaurally 180° out of phase. In this experiment, contrary to the results reported earlier, no difference was observed for the three exposure conditions.
44(1968); http://dx.doi.org/10.1121/1.1970174View Description Hide Description
The purpose of these experiments was to test Ward's hypothesis that recovery time is a function of the initial loss and independent of test frequency and to extend his findings to higher test frequencies. Twenty‐five subjects were exposed to 110‐dB sinusoidal stimulation at 700, 2700, and 8000 Hz at varying durations so as to produce approximately equal (within‐5‐dB) losses (10–30 dB) at one of 13 test frequencies ranging from 250 to 12 000 Hz. Recovery was followed until thresholds had returned to within 5 dB of the pretest levels. The results indicated that subjects took significantly more time to recover from high‐frequency than from low‐frequency exposures. The analysis is still in progress, but these results suggest that recovery from high‐frequency sinusoidal exposure is slower than recovery from low‐frequency sinusoidal exposure.
44(1968); http://dx.doi.org/10.1121/1.1970176View Description Hide Description
An intensive study of the attenuation properties of numerous ear protectors and the statistical treatment of the data obtained, explains and/or documents many previously suspected or noted irregularities in measurement procedures. Detailed data on subjective measurement, directionality, linearity, and asymmetry of attenuation of ear protectors were collated. These findings will be presented along with a description of the nonparametric statistics used to ascertain central tendencies, sources and magnitudes of variability, and degrees of correlation of the data. In the study, nine ear protectors, 10 trained subjects, and two artificial heads were utilized. The USA Standards Institute's method (i.e., free‐field threshold shift) was used to obtain subjective measurements. The artificial heads were employed for the acquisition of detailed information on the noise‐exclusion properties of each ear protector.
44(1968); http://dx.doi.org/10.1121/1.1970178View Description Hide Description
Subjects were exposed, on different test days, to 166‐dB (peak, normal incidence) impulses 34, 58, 72, or 96 μsec in duration spaced 1 sec apart. For each pulse duration, the subjects were first exposed to one pulse, then the number of pulses was doubled on successive days until the temporary threshold shift (TTS) following exposure exceeded 30 dB. Intercorrelations of numbers of impulses required to reach criterion TTS at each duration were obtained; they were highest when durations were similar. At the largest pulse duration, a median of only four impluses was required to achieve criterion; some individuals exceeded criterion at one impulse, while others required hundreds. No evidence was seen of a 4‐kHz notch; maximum shift was at high frequencies (10–15 kHz) and relatively independent of impulse duration. There is reason to believe if one exceeds allowable TTS in the speech‐range frequencies with this kind of impulse noise, there is a chance of producing permanent high‐frequency loss.
44(1968); http://dx.doi.org/10.1121/1.1970180View Description Hide Description
Cross‐modality tests, matching the apparent intensity of a 100‐Hz vibration to the noisiness of bands of noise, have been conducted to measure the effects of background noise upon judged noisiness. The growth function for noisiness behaves like a modified power function of the form , where ψ is noisiness, I is the intensity of the stimulus, I 0 is the threshold intensity, and k and n are constants that depend upon frequency. A calculation scheme has been developed that reduces the sound‐pressure level of each band by an amount dependent upon the signal‐noise/background‐noise ratio. For signal‐noise/back‐ground noise band‐level ratios of greater than 65 dB, the band correction is equal to 0. For realistic background spectra and signal‐noise/background‐noise over‐all level ratios of 40 dB, the total correction is approximately 3 PNdB. [Research supported by the Federal Aviation Administration.]
44(1968); http://dx.doi.org/10.1121/1.1970182View Description Hide Description
Judgment tests have been conducted to measure the growth of noisiness as a function of sound‐pressure level (SPL) for tones and bands of noise over a range of frequencies and reference SPL's. Adjustment‐test results at 1000 Hz showed a mean value of 11.5 dB per doubling or noisiness except for the lowest reference level (50 dB SPL) where the mean was 16.7 dB. Magnitude‐estimation results yielded values between 20 and 27 dB per doubling of noisiness. The specific value used for the growth of noisiness did not significantly affect the calculation of relative PNL values. Equal noisiness contours were obtained for: (1) pure tones in a free field, (2) bands of noise in a free field, (3) bands of noise in a diffuse field, (4) 1‐sec and 4‐sec stimulus durations, and (5) for loudness and noisiness instructions. No significant differences were found for the latter two conditions. [Research supported by the Federal Aviation Administration.]
- Session C. Speech Communication I: Anatomy and Physiology of the Vocal Tract
44(1968); http://dx.doi.org/10.1121/1.1970184View Description Hide Description
Pulsed ultrasound has been used to monitor and record the motion of the lateral pharyngeal wall during speech. A time‐motion (T.M.) method of display was recorded on the screen of a storage oscilloscope. An audio signal was simultaneously displayed and permanent records were made by photographing the storage oscilloscope screen. Tests were conducted to insure that the recorded signals were produced by the motion of the lateral pharyngeal wall. Distance calibration was made using a known displacement in a water bath. The effect of different ultrasonic frequencies was investigated by making measurements with both 2.25‐ and 5‐MHz ultrasound beams. The results of all the measurements are reported in the following paper. [This research was supported by the National Institutes of Health.]