Volume 60, Issue 2, August 1976
Index of content:
60(1976); http://dx.doi.org/10.1121/1.381106View Description Hide Description
An integral expression for the spatial coherence function of pure‐tone sound fields is formulated under hypotheses that the component fields of individual paths or modes are locally plane waves and mutually incoherent in the average. The total field is not assumed to be a Gaussian random variable. The resulting coherence is quasi or locally homogeneous, varying slowly with average position as the strength of the component fields change. Statistical estimates for the contributions of these components are taken from existing literature on average transmission loss. The expression for coherence has an explicit dependence on the environmental parameters, the range, and the depths of source and receivers. Numerical results and approximate expressions are developed for channels characteristic of shallow and deep water.
Subject Classification: 20.15; 60.20.
60(1976); http://dx.doi.org/10.1121/1.381107View Description Hide Description
Acoustic parametric convolvers operate by spatially integrating a term produced by the nonlinear interaction of two propagating strain waves. The integrand will have an additional spatially dependent weighting factor and inaccuracies in the convolution output will result if the product term is not uniformly produced and summed. Under appropriate assumptions, the expansion of the acoustic field into an angular spectrum of plane waves is used to show that no erroneous weighting is introduced by diffraction, although diffraction does contribute a uniform loss to the convolver output. Under additional assumptions, the diffraction loss for the convolved output is shown to equal that which occurs in transmitting from one transducer to the other.
Subject Classification: 25.30, 25.35; 35.54; 60.30.
60(1976); http://dx.doi.org/10.1121/1.381108View Description Hide Description
Analysis of the ambient‐noise horizontal directionality data from a recent environmental acoustic exercise indicates that wind‐generated noise at 10 Hz is spatially correlated over large apertures. The noise levels measured by a large aperture, linear towed array were highest near broadside during high‐wind conditions. Spatially correlated wind‐generated noise arriving from the sea surface directly above the array is offered as a possible explanation and is consistent with results of previous wind‐noise experiments. In the limiting case of very high wind speeds (?35 knots), the array response measured for this nearfield effect closely approximated the farfield broadside beam pattern at 10 Hz. At lower wind speed, the array response was wider near broadside. This implies that the spatial correlation of wind‐generated noise at 10 Hz becomes greater as the wind speed increases.
Subject Classification: 28.20, 28.65; 50.80.
60(1976); http://dx.doi.org/10.1121/1.381109View Description Hide Description
Ambient sea noise has been measured at 40 positions in mainly tropical waters near Australia: in the Indian Ocean; the Arafura and Timor Seas; and in the Coral and Tasman Seas(Pacific Ocean). Wind‐speed dependence was observed at all frequencies of measurement from 22 to 5000 Hz. At hydrophones suspended at depths between 8 and 25 m, the wind‐ dependent noise appeared to be independent of bottom depth which varied from 26 to 6700 m. The rate at which wind‐dependent noise increased with increasing wind speed was found to be significantly less than has been observed in the North Atlantic Ocean. Non‐wind‐dependent noise, observed at frequencies below 250 Hz, varied by about 10 dB within a particular sea and by up to 20 dB between seas, and could be broadly related to the relative shipping densities and propagation conditions in the individual seas. In the shallow Arafura and Timor Seas north of Australia, there is evidence that noise of distant shipping is not a significant component of the ambient noise. A model of traffic noise in the Tasman Sea predicts levels in agreement with observed noise levels.
Subject Classification: 30.70.
Use of phase measurements for arrival‐time extraction of overlapping sinusoidal signals in shallow water60(1976); http://dx.doi.org/10.1121/1.381110View Description Hide Description
The use of phase measurement for sinusoidal signal arrival‐time extraction of overlapping signals subjected to multipath conditions in shallow water is considered. The phase‐measurement method is applied to signal data from a research cruise. The method does not give convincing results. For distances greater than 10 000 m the results are negative. The results obtained are strongly dependent on the sound velocity profile. The signal pulse length was a very important parameter in the results. The optimal pulse length was shown to be greater than 10 msec and less than 100 msec regardless of the hydrophone depth for the two depths investigated.
Subject Classification: 30.20; 60.30.
60(1976); http://dx.doi.org/10.1121/1.381111View Description Hide Description
The objectives of this report are to review selected published measurements of the attenuation, or energy damping, of low‐strain shear waves in surficial water‐saturated sands and silt clays (mud) which might occur as marine sediments. In various computations, a linear viscoelastic model is favored in which velocity dispersion is negligible, the linear attenuation is proportional to the first power of frequency, and the specific dissipation function, 1/Q, and the logarithmic decrement are independent of frequency. The logarithmic decrement is favored as a measure of energy damping because of research in soil mechanics. The very sparse data indicate that in water‐saturated sands and silt clays, the logarithmic decrements are mostly between 0.1 and 0.6. If approximate values of shear waveenergy losses are required for generalized computations, it is suggested that a value for the logarithmic decrement of 0.30±0.15 be assumed for sands, and 0.2±0.1 for silt clays. Measured logarithmic decrements of compressional waves in sands average about 0.10±0.03; in silt clays about 0.02±0.01. The average values of the ratio of compressional‐to‐shear‐wave logarithmic decrements, using the above average values, would be 0.3 for sands and 0.1 for silt clays.
Subject Classification: 30.20; 40.50, 30.30.
60(1976); http://dx.doi.org/10.1121/1.381112View Description Hide Description
The problem of the transmission of sound from an airborne source into the ocean, through the rough oceansurface, is considered. A high‐frequency approximation is adopted, requiring for its validity that the sound wavelength in air be (much) less than typical radii of curvature and than typical displacements of the oceansurface. This condition is approximately met in many applications. Attention is focused on observation directions with small grazing angles. The result is approximately unaffected by surface roughness when the grazing angle is greater than the rms surface slope, and approaches a limit (proportional to the variance of the surface slope) when the grazing angle is less than the slope. The result is independent of the source height, under the stated approximations, as is also the case for the transmission through a plane surface. The enhancement due to surface roughness, even at moderate wind speed, ranges up to 10 dB.
Subject Classification: 30.25.
60(1976); http://dx.doi.org/10.1121/1.381113View Description Hide Description
We discuss two aspects of the general problem of sound propagation in an ocean characterized by a range‐dependent profile. We first consider a normal‐mode formalism for describing a situation in which the eikonal approximation can be applied in the horizontal plane. The difficulty with this approach is shown to be an ordering in the geometric optics path‐length parameter due to mode coupling. By example we indicate how this difficulty can be overcome. We then use the formalism to discuss the approximations which make up the parabolic equation method. The parabolic equation is derived by assuming the eikonal approximation in the horizontal plane and a degenerate spectrum approximation. The degeneracy approximation is reflected in a stationary phase approximation to an integral representation for the pressure field. We then derive improved equations by relaxing these assumptions. We relax the degeneracy assumption by considering an improved stationary‐phase approximation and we relax the eikonal assumption by considering the ’’supereikonal’’ approximation.
Subject Classification: 30.20, 30.25, 30.50.
60(1976); http://dx.doi.org/10.1121/1.381090View Description Hide Description
A hydrodynamic model for flows in the deep ocean is developed in order to determine the velocity field and sound‐speed distribution for use in acoustic transmission problems. A scaling of the governing equations is constructed that explicitly includes sound speed. A subsequent perturbation expansion yields a set of approximate equations for motions nearly in geostrophic and hydrostatic balance, such as large‐scale, quasisteady currents and Rossby waves. The quasigeostrophic potential vorticity equation or a simpler limiting case of this equation arises from the perturbation scheme to govern higher‐order dynamics of the stream function for these flows. The results of the analysis are used to obtain a significant simplification of the ray equations of geometrical acoustics for moving media. For the particular class of flows considered here, the model equations are applicable if the ocean depth is about 1 km or greater and if the spatial and temporal scales of variation of the motions are of the order of 100 km and 10 days, respectively. A solution for a flow such as the Antilles current is derived. Isospeed curves for this solution are shown in a plane perpendicular to the current, and specific features of the curves are discussed and interpreted.
Subject Classification: 30.25, 30.20; 28.60.
60(1976); http://dx.doi.org/10.1121/1.381091View Description Hide Description
Two types of electromagneticultrasonic transducers were made and tested on mild steel, stainless steel, and aluminum. One of the transducers generates and detects a relatively large amount of longitudinal wave and simultaneously a small amount of radially polarized transverse wave in nonmagnetic metal. The other is for a large amount of radially polarized transverse wave and a small amount of longitudinal wave, also in nonmagnetic metal. Both transducers showed different characteristics for ferromagnetic metal, giving greater efficiency to the radially polarized transverse wave. It was also shown experimentally with an electromagnetic transducer that the mode conversion between longitudinal waves and radially polarized transverse waves occurs fairly efficiently at a vertical reflection.
Subject Classification: 35.60, 35.80; 85.40, 85.48.
60(1976); http://dx.doi.org/10.1121/1.381092View Description Hide Description
Poles of the infinite plane‐wave reflection coefficient are used to show a correspondence between Rayleigh and Lamb modes of vibration. It is demonstrated that a Rayleigh vibrational mode is a special type of Lamb mode of vibration. Further, it is shown that it should be expected that one vibrational mode for a thick plate should be similar to the theoretically predicted vibrational mode of an infinite half space, a Rayleigh mode. Thus, it is consistent to use a thick plate as an approximation to an infinite half space and expect results predicted by Rayleigh‐wave analysis.
Subject Classification: 35.54; 40.24.
Experiments involving sonification of liquid helium at 4.2 K in the cathode–anode interspace of a photodiode tube60(1976); http://dx.doi.org/10.1121/1.381093View Description Hide Description
An experimental study of the photo‐injection of electrons into liquid helium at 4.2 K during sonification shows a dramatic increase in the observed photocurrent. The increase in photocurrent results from a sonically induced increase of the electron bubble mobility and a reduction in mean thermalization time due to an increase in local density fluctuations in the vicinity of the photocathode. The increase in mobility of the electron bubbles in the sound field has the same functional dependence as static pressurization, suggesting an effective rectification mechanism. Studies in which the sound field is abruptly turned off show that the half‐life of potential cavitation nuclei within the microbubble field is 7.6 sec. These sonically induced changes in the electronmicrobubblephotocurrent provide a unique method for the study of the formation and population dynamics of cavitation nuclei.
Subject Classification: 35.47, 35.32.
60(1976); http://dx.doi.org/10.1121/1.381094View Description Hide Description
Acoustical coupling of a turbulence‐excitedpanel and a supersonically moving fluid is investigated. Response of a panel with a single rib discontinuity is considered. Radiation from the panel into the fluid induced by a uniform line force is investigated and the line admittance of the panel in the presence of supersonic flow is obtained. Numerical results are presented for the line admittance and for the ribbed and unribbed infinite panel response. A relation is given which upon integration will yield the radiation damping of a finite panel (radiation into the supersonic flow).
Subject Classification: 40.35; 28.65; 20.60.
60(1976); http://dx.doi.org/10.1121/1.381095View Description Hide Description
The ability to detect gaps in continuous sinusoids was studied in both single‐ and two‐channel tasks. From single‐channel tasks, psychometric functions were obtained. Pulse and gap functions were found to be essentially parallel, with gap detection being 3–4 dB poorer than pulse detection. In two‐channel tasks, two‐channel detection of gap signal events produced a lower level of performance than two‐channel detection of pulse signal events. However, in an unusual condition where a gap stimulus produced what sounded like a pulse to the observer, performance in two‐channel gap detection and two‐channel pulse detection was quite similar. This latter result suggests (1) gap stimuli enter the system as efficiently as pulse stimuli, and (2) it is the system’s inferior ability to process perceptual gaps which leads to the lower performance normally observed in two‐channel gap detection. In a fourth two‐channel task which required the observer to detect gap stimuli in one channel and pulse stimuli in the other, no detrimental effect on performance could be detected—a result illustrating the flexibility of the two‐channel processing mechanism. Implications of the results for the INR model of two‐channel processing are discussed.
Subject Classification: 65.75, 65.58, 65.68.
60(1976); http://dx.doi.org/10.1121/1.381096View Description Hide Description
2AFC scores were obtained on 50 untrained subjects who had to respond to a 3% increment or decrement of the fundamental frequency of two successive signals. The first signal in each trial was a reference, containing six harmonics of 200 Hz. Main variables were the number of harmonics of the second signal in a trial (either 3, 2, or 1) and the presence or absence of continuous background noise. Low scores were observed in the absence of the noise (typically 65% for the 3‐ and 2‐component signals, and virtually 50% for the 1‐component signals), whereas in the presence of the noise much higher scores were observed (typically 90% and 80%, respectively). The results indicate that background noise has great impact on the perceptibility of a ’’low pitch’’ related to the fundamental frequency, even more so than the number of components presented (within the range of three, two, or one components): at a low S/N ratio, even a single harmonic may give rise to the perception of a subharmonic low pitch. Subsequent experiments with three trained subjects indicated that this potential of a single pure tone to evoke a subharmonic pitch deteriorates for harmonic numbers of about n=7 to 11. The theoretical implications of these subharmonic pitches of a pure tone to pitch perception in general are discussed.
Subject Classification: 65.54, 65.75, 65.35.
Discrimination and labeling of noise–buzz sequences with varying noise‐lead times: An example of categorical perception60(1976); http://dx.doi.org/10.1121/1.381097View Description Hide Description
The onset of a noise [0.9–2.1 kHz, 55 dB SPL (A weighted)] preceded that of a buzz [100 Hz, 0.5–3.0 kHz, 70 db SPL (A weighted), 500 msec] by −10 to +80 msec and both terminated simultaneously. Eight adults discriminated among noise‐lead times in an oddity task. In separate sessions, they labeled singly presented stimuli with either of the two responses: ’’no noise’’ or ’’noise.’’ The results are highly similar to those reported for the categorical perception of synthetic plosive consonants differing in voice‐onset time. On the average, discrimination was best across a noise‐lead‐time boundary of about 16 msec, where labeling also shifted abruptly. These results and those of categorical perception, generally, are interpreted in terms of Weber’s law as applied to a single component within a stimulus complex. It is concluded that categorical perception of sounds is not unique to speech and suggested that it may be a general property of sensory behavior.
Subject Classification: 65.75; 70.30.
Susceptibility to auditory fatigue: Comparison of changes in cochlear nerve potentials in the guinea pig and chinchilla60(1976); http://dx.doi.org/10.1121/1.381098View Description Hide Description
The susceptibilities of the guinea pig and chinchilla to acoustic fatigue are compared. For this comparison they were exposed to a 6‐kHz tone. The exposures were within the range of 60–110 dB SPL. The N 1 potential of the cochlear nerve and the cochlear potential were recorded before and after exposure. Both species suffered greatest N 1 changes a 1/2 octave above the exposure tone. No appreciable changes in cochlear potential were observed. In agreement with previous studies for equal SPL’s, the chinchilla was found to be more susceptible to fatigue than the guinea pig. However, when the exposures are based on maximum cochlear‐potential output, these animals suffer similar losses in N 1. This suggests that impedance differences in their ears accounts for the susceptibility difference. Nerve fatigue is described and found related to hair cell distortion.
Subject Classification: 65.66, 65.40, 65.42.
60(1976); http://dx.doi.org/10.1121/1.381099View Description Hide Description
Temporal resolution of a silent interval (gap) imbedded in a three‐tonal‐pulse train was investigated. Two frequencies were alternated in a three‐event array where the gap could appear between the first and second or second and third auditory event. Frequency and gap position effects were systematically explored using a temporal two‐interval forced‐choice paradigm and a modified tracking procedure. Gap detection was found to vary as a function of antecedent event duration and signal‐comparison stimulus similarity. Indications of a filter process similar in bandwidth to the empirical critical band was found in fully correlated signal‐comparison stimulus conditions. Obtained results are consistent with fixed‐filter signal‐envelope detection models found in the simultaneous and forward‐masking literature.
Subject Classification: 65.75, 65.68.
60(1976); http://dx.doi.org/10.1121/1.381100View Description Hide Description
Masking period patterns, i.e., the SPL of just audible triggered sequences of short high‐frequency test tone bursts as a function of their temporal spacing throughout the period of a masker, were measured using the following maskers: low‐frequency two‐tone and three‐tone complexes, Gaussian‐shaped impluses and differentiated Gaussian impulses with both condensation and rarefaction peaks. The similarity between masking period patterns and neurophysiologically measured period histograms was confirmed with respect to the temporal structure. However, distinct differences appear in the amplitude structure of both patterns, indicating a strong nonlinearity involved in the transformation of the masker’s time function into the masking period pattern. Starting a few decibels above threshold in quiet, the peaks of the patterns are elevated by 10 dB for each 5‐dB increment of masker level. This fact points towards a square law for high‐frequency test signals and low‐frequency maskers. The steepness of slope of the masking period pattern seems to reach a maximal value of 10 dB/msec. Relations between neurophysiological period histograms and corresponding masking period patterns as well as differences between the patterns produced by condensation and rarefaction peak maskers, respectively, are discussed.
Subject Classification: 65.58, 65.35, 65.68.
60(1976); http://dx.doi.org/10.1121/1.381101View Description Hide Description
A measure was obtained of the amounts of variability in fundamental voice frequency F 0 in simple sentences produced by several adult male speakers. A comparison was made of intra‐ and interspeaker variability, and a detailed analysis was performed to determine the nature of this variability. Great variability was found between speakers, but more important was the finding that the intraspeaker variability was essentially as great as the interspeaker. This indicates that even for a single speaker’s utterances, considerable ’’noise’’ is present which can obscure the prosodic ’’signals.’’ The F 0 variability involves seemingly random interleavings that affect the relative F 0 values within an utterance, as well as, simple shifts in average F 0 that affect the absolute F 0 of the contour as a whole.
Subject Classification: 70.40, 70.70.