Volume 35, Issue 3, March 1963
Index of content:
35(1963); http://dx.doi.org/10.1121/1.1918446View Description Hide Description
Amplitude fluctuations of surface‐reflected signals from a source at 50‐ft depth were recorded by two receivers at 2600 fathoms. Some features of the observations are attributed to ocean‐wave focusing effects. It is shown that the signal‐amplitude distribution may be described by using a Rayleigh distribution modified by adding a single, focused contribution of random phase. An observed delay in the envelope of the signals recorded by the two receivers that is greater than the wavefront delay is explained as resulting from the sweeping action of an acoustic caustic produced by the ocean waves. Auto‐ and cross‐correlation functions were computed from the data.
35(1963); http://dx.doi.org/10.1121/1.1918448View Description Hide Description
A study was made of the accuracy to be expected from long‐range SOFAR measurements and of some of the factors which limit that accuracy. Depth charges were fired along a track across the South Atlantic and into the Indian Ocean. The signals were received on Fernando de Noronha and Ascension Island. Distances ranged from 75 to 4700 nautical miles.
The results show that it should be possible to obtain an accuracy of about 50 m in 1000 miles from a 20‐shot sample. Off‐axis shooting seriously affects the accuracy. Large topographic features, which do not extend into the sound channel axis itself, cut off the front of the signal, without impairing arrival‐time accuracy, while continental land masses block the signal entirely. Attenuation of the peak‐signal intensity with increasing range along an unimpeded path is consistent with cylindrical spreading.
35(1963); http://dx.doi.org/10.1121/1.1918450View Description Hide Description
The solution to the approximate integral form of the reflection in the “far field” has been developed, as well as a graphical, or numerical, method for the case of a rigid, finite plane rotated about an edge, and for a wedge shape. The Kirchhoff approximation is used. The reflection may be conveniently presented in diagrams in the complex plane. Extension of the method has been made to the case for the magnitude of the reflection at any orientation of a reflector. Physically measurable quantities are required in the evaluation of this sum. Although the evaluation is not exact, rapid, or inclusive of all reflection mechanisms, it does make possible the derivation from shapes which are not analytically describable.
35(1963); http://dx.doi.org/10.1121/1.1918452View Description Hide Description
A variety of structures in the form of spheres can provide sharp focusing and directivity gains comparable to those from a circular piston with the same diameter. With an isotropic refractive sphere, spherical aberration does not impair focusing except for angular resolutions of the order of a few degrees and less. Luneberg lenses with compliant tubes or refractive liquid packets eliminate spherical aberration, but the focal distance may change as a result of dispersion in the resultant index of refraction. The outer shell of a Luneberg lens alone, and without the inner core, provides an angular resolution which is comparable to that for the full lens, but with a decrease in the directivity gain which is not as great as the reduction in the amount of refractive matter. Primary feeds with considerable directivity are essential for all types of lenses except those with the focus near the outer surface of the lens.
Some Effects on Listener Accuracy of Competing Messages Varied Systematically in Number, Rate, and Level35(1963); http://dx.doi.org/10.1121/1.1918454View Description Hide Description
A criterion speaker read intelligibility‐test sentences against a background of competing voices in fixed combinations with respect to number, level, and rate as controlled by a time compression‐expansion device. The essential finding was that all main factors when raised in number or amount were associated with decrements in intelligibility scores. Statistical analysis yielded a significant three‐way interaction; hence trends in individual main effects must be interpreted conservatively.
35(1963); http://dx.doi.org/10.1121/1.1918456View Description Hide Description
This paper is concerned with the effect of noise cross correlation on the detection of a signal in noise by means of a correlation detector. It also considers the effect of noise cross correlation on statistical errors arising in the approximate measurement of signal cross correlation by the sample mean of the output of a multiplier‐averager. In the detection problem, the probability of detection and sample size are investigated under the influence of noise cross correlation. In a specific application, the effect of noise cross correlation is determined in the case of a spherical, isotropic noise field. Detailed results are presented when the inputs have the characteristics of RC‐filtered white noise and when the detector includes a finite‐time perfect averager. In the measurement problem, the influence of noise cross correlation on the accuracy of correlation measurement is studied from a probability standpoint. Here, too, the required increase in sample size is investigated.
35(1963); http://dx.doi.org/10.1121/1.1918458View Description Hide Description
A theory of the clarinet is developed based on the experimental observation that for weak tones the reed and air‐column vibrations are nearly sinusoidal. The clarinet is assumed to be a cylindrical air column open at one end and closed at the other by a diaphragm containing a slit of variable width, corresponding to the aperture between reed and mouthpiece. A velocity potential appropriate to a tube with wall friction is assumed. The impedance of the slit as a function of opening and pressure across it is evaluated experimentally and checked against theory. The volume flow through the slit is calculated from the velocity potential and equated to the flow calculated from the slit impedance. The flow and impedance both depend on the pressure, which in turn is again calculated from the velocity potential. Expressions for the operating frequency and threshold blowing pressure are obtained by assuming the flow to consist of a small alternating component superimposed on a steady component. The frequency is found to be below the resonance of the system considered as a tube closed at the reed end, the shift varying nearly linearly with the slit opening and depending on the reed damping. The threshold blowing pressure is found to be proportional to the opening and to the reed stiffness. Operating frequencies and pressures for the artificially blown clarinet were measured experimentally and found to be in very good agreement with values calculated from the theory.
35(1963); http://dx.doi.org/10.1121/1.1918460View Description Hide Description
The tones were produced by the pipe organ in the Smith Auditorium at Brigham Young University. They were picked up by a microphone placed at various positions in the auditorium and recorded on magnetic tape. The tape was taken into the laboratory where analyses of the tones were made. The structure of full organ tones becomes very complicated. For example, when the three keys for the major chord were depressed, there were 229 partials whose frequency and level were measured.
After the analysis of the tone was made, a synthetic tone was constructed by our synthesizer. Judgment tests were made by juries to see if synthetic tones could be distinguished from real tones. These juries were unable to distinguish between them. The paper discusses musical warmth. It shows that it is definitely related to the level variation of the partials. These variations are due to the frequencies of several partials being close together, causing beats.
Methods and apparatus were developed so that a tone could be warmed to any extent without using the large number partials produced in the pipe organ. A method of rating the warmth of organ tones is proposed and will be used in our future work on all musical tones.
35(1963); http://dx.doi.org/10.1121/1.1918462View Description Hide Description
The paper applies elementary circuit ideas to bowed‐string instruments and their component parts. Parameters are defined and calculations based on simple circuit diagrams for the main resonance and the air resonance; curves describe their combined performance. The relative importance of circuit resistances—wood loss, radiation, and wall‐surface loss—is discussed. Wall‐surface loss is an important component of air decrement. No material improvement is to be expected from change in decrement or enclosure volume.
A theory for the wolfnote is given in terms of the beating of two equally forced oscillations, together with a criterion for its occurrence and a method for its elimination.
The paper analyzes principles of dimensional scaling between members of the violin family and shows why the cello and viola are more susceptible to wolftone than the violin.
A study of impedance requirements in wood shows that flexural similarity depends on the parameter c/ρ (compressional velocity over density); high values are in general favorable in the top plate. In the violin, cross‐grain losses probably exceed those along the grain.
35(1963); http://dx.doi.org/10.1121/1.1918463View Description Hide Description
A method of automatic pitch extraction is describe, and examples are shown of results obtained. The speech signals are processed by a real‐time Fourier analyzer whose output is converted to digital form and recorded on tape for processing by an IBM 700 computer. The logic of the computer program, written in FORTRAN language, is described together with a discussion of the accuracy of the over all pitch extraction system.
35(1963); http://dx.doi.org/10.1121/1.1918465View Description Hide Description
Pitch perturbations were computed by measuring the differences between the durations of adjacent fundamental periods from recorded acoustic waveforms. Sound‐synchronized high‐speed motion pictures of the vocal cords were also taken with a laryngealmirror. It was found that pitch perturbations reflect variations in the waveshape of the glottal area wave, as well as variations in glottal periodicity. Perturbations having an absolute value ⩾0.5 msec are apparently induced by transients in the air‐pressure drop across the glottis, which may in turn be caused by changes in vocal‐tract configurations. Speakers with longer fundamental periods tend to have larger perturbations. The pitch perturbations of 23 speakers with pathologic larynges were measured. Certain of the speakers who had pathologic growths on their vocal cords had larger pitch perturbations than did normal speakers with the same median fundamental periods. The “perturbation factor,” which was defined as the percent that perturbations ⩾0.5 msec occurred, was computed for each speaker. The magnitude of the perturbation factor may serve to detect certain types of pathologic laryngeal conditions.
35(1963); http://dx.doi.org/10.1121/1.1918467View Description Hide Description
A pattern‐matching procedure for automatic recognition of talkers was used to study the effects of variations in patterns upon recognition performance. Several utterances of common words, excerpted from context, were spoken by ten talkers and converted to time‐frequency‐energy patterns. Some of each talker's utterances were used to form reference patterns and the remaining utterances served as test patterns. The recognition procedure consisted of cross‐correlating the test patterns with the reference patterns and selecting the talker corresponding to the reference pattern with the highest correlation as the talker of the test utterance. The same recognition procedure was used with patterns reduced to two dimensions. The recognition score for three‐dimensional patterns was 89%. Reducing the original patterns to time‐energy patterns resulted in a lower recognition score; however, when only spectral information was retained, recognition results were the same as those for three‐dimensional patterns. No errors were made in recognition based on a small sample of patterns consisting of pooled spectra of several different voiced sounds.
35(1963); http://dx.doi.org/10.1121/1.1918469View Description Hide Description
The time dependence of the intensity of the light in the Fresnel region diffracted by a progressive wave is measured directly. The measurements are compared to the theory of a periodic moving grating. Suggestions are made for measuring the harmonic content of the wave and the fundamental wavelength of the grating. Examples are given of the available periodic intensity fluctuations, and a method for determining relative pressures is suggested. This paper gives the details of work that was reported at the Fourth International Congress on Acoustics (K. L. Zankel and H. M. Colbert, Fourth International Congress on Acoustics, Copenhagen, 21–28 August 1962).
35(1963); http://dx.doi.org/10.1121/1.1918471View Description Hide Description
Current methods for ultrasonicinspection of metals and other solid materials are limited by their inability to determine the geometry of concealed defects if these defects are comparable in size, or smaller than the width of the ultrasonic search beam. Whenever such relatively small flaws are involved, ultrasonic pulse echo testing, if conducted at a single frequency, will yield the flaw location only, but no information on the flaw geometry. It is therefore very difficult to interpret the test results. The difficulty can be overcome by introducing the novel ultrasonic test method described in this paper. In this procedure, ultrasonic signals are utilized which contain a broad band of frequencies, and in analogy to optics, can therefore be considered as “white” ultrasonic pulses. The form and spectral energy distribution, or “color,” of such ultrasonic pulses is influenced by the geometry of a defect from which they are reflected. Hence, an analysis of the defect echo yields information on the defect configuration. The successful application of the ultrasonic pulse analysis method for differentiating between flaws of various configurations is illustrated By a number of test examples.
35(1963); http://dx.doi.org/10.1121/1.1918473View Description Hide Description
Existing theories for predicting the acoustic radiation damping of structures are reviewed. A uniform pressure approximate theory for predicting the acoustic radiation damping of a single flexible panel, forming part of an otherwise rigid plane baffle, formulated by D. J. Mead [“The Effect of a Damping Compound on Jet Efflux Excited Vibrations,” Aircraft. Eng. 32, 374 (1960)], is developed into an exact theory, taking into account the nonuniform pressure distribution over a vibrating surface; a general expression is also developed to include any form of panel boundary support. The equations developed have been evaluated numerically by computer, for ease in predicting the acoustic damping factor for a wide range of panel sizes. The acoustic damping of a single panel forming part of a flat baffle and also of a cylindrical baffle has been measured experimentally by three methods.
35(1963); http://dx.doi.org/10.1121/1.1918475View Description Hide Description
Rods exhibit some very complex vibrational characteristics. The ends behave like cavity resonators in a narrow frequency band and vibrate with amplitudes as much as 80 times the amplitude of the incident wave. At low frequencies, bending vibrations of rods are described by the complex Young's modulus, and the Bernoulli theory of bending leads to the same result as the exact theory. At high frequencies, bending vibrations are described by the shear modulus, and the loss factor approaches or becomes equal to that of shear waves.
Plates exhibit a behavior similar to bending vibrations in rods. At low frequencies, Young's modulus and its loss factor describe the vibrational behavior of a thin plate. As the frequency increases, the modulus approaches the shear modulus, and the loss factor approaches the loss factor for shear waves after fluctuating in a periodic manner between the loss associated with Young's modulus and that of the shear modulus. The measurement, of the loss factor by bandwidth and decay methods agree, provided a narrow bandpass filter is used so that the low‐frequency modes, which are always excited by a point force, are suppressed.
35(1963); http://dx.doi.org/10.1121/1.1918477View Description Hide Description
The transient response of damped nonlinear isolation mountings to pulselike foundation displacements has been determined theoretically. The response of both “stiffening” and “softening” springs has been examined, and large departures of spring stiffness from linearity have been considered. The effective duration τ of the displacement pulses has been varied through a broad range of values. Shock spectra for three displacement pulses of widely different duration have been determined in the case of a linear spring. It is found that only when τ is small can both the displacement and acceleration of the mounted item be reduced significantly, in which case the displacement of the item relative to its foundation has attained a large constant value. The shock spectra obtained when the item is resiliently mounted and unmounted are shown to differ in level by a factor that is given—in two well‐defined regions—by the ratio of the maximum displacements or the maximum accelerations above and below the isolation mounting. In such cases it is possible to predict the difference in level of the shock spectra by a single figure of merit.
35(1963); http://dx.doi.org/10.1121/1.1918479View Description Hide Description
The response of two‐ and three‐element isolation mountings to sinusoidal, steplike, and pulselike input displacements has been determined theoretically. In contrast to the two‐element or simple mounting, the transmissibility of the three‐element mounting at resonance can be made small without reducing the rate at which transmissibility decreases at high frequencies (12 dB per octave). The transient response of the mountings has been described in terms of the dependence upon a parameter γ of the suitably normalized maximum values of displacement (SDR), relative displacement (RDR), and acceleration (SAR) of the mounted item. The parameter γ is the ratio of the half‐period of natural vibration of the mounting system to a time τ, which represents either the step rise time or the duration of the pulse. Shock spectra have been computed for input transients defined by three widely different values of γ. In two well‐defined regions, the resiliently mounted and unmounted shock spectra are shown to differ in level by a factor that is given by the SDR or the SAR. At the higher values of γ, it is apparent that the three‐element mounting can reduce the displacement of the mounted item in the manner of a heavily damped simple mounting, while reducing the acceleration of the mounted item more effectively.
- LETTERS TO THE EDITOR