Volume 33, Issue 8, August 1961
Index of content:
33(1961); http://dx.doi.org/10.1121/1.1908935View Description Hide Description
The method of free response refers to the following listening situation. Against a background of noise, a weak signal is presented several times in a long (2‐min) observation interval. The temporal intervals between the presentations of the tones are randomly distributed; consequently, the listener does not know when a tone will occur, and he does not know how many tones will be presented. From one series of observation intervals to the next, the listener is instructed to adopt various criteria and to press the single response‐key each time he “hears a tone.” The problem consists in the determination of a procedure that allows the total number of yes responses to be partitioned meaningfully between “hits” and “false alarms.” A model is developed in which the measurable quantity, rate of response, is related to the “hit rate” and to the “false‐alarm rate.” Although the criterion adopted by the listener cannot be directly evaluated, the use of a wide range of criteria makes it possible to estimate the detectability ds of the signal. Two experiments are described, and the results support the model.
33(1961); http://dx.doi.org/10.1121/1.1908878View Description Hide Description
Listeners were required to detect an auditory signal against a background of “white noise.” The effects (1) of giving trial‐by‐trial information as to whether or not a signal was delivered, and (2) of giving the subject an opportunity to hear the signal before the test sequence began, were studied at two levels of signal energy. The results were analyzed within the context of the theory of signal detectability. Subjects who were given an opportunity to hear the signal before the test sequence began maintained a stable level of performance throughout the experimental session. On the other hand, subjects who were given no opportunity to hear the signal performed near chance level at the beginning of the session but showed gradual improvement as trials progressed. The effect of trial‐by‐trial feedback was surprisingly small in all groups. Near the end of the session, the signal was demonstrated to all subjects and the differences between the groups vanished.
Normalized Representation of Noise‐Band Masking and Its Application to the Prediction of Speech Intelligibility33(1961); http://dx.doi.org/10.1121/1.1908880View Description Hide Description
Measurements are made of the masking of pure tones by various bands of noise. Curious masking spreads beyond upper frequency limits of noise are observed and these are expressed uniquely by means of the relative masking and the incremental bandwidth. Then the normalized representation of noise‐band masking in some restricted ranges is achieved and applied for the prediction of speech intelligibility. Predicted articulation scores agree fairly well with measured ones.
33(1961); http://dx.doi.org/10.1121/1.1908882View Description Hide Description
Pitch shifts with changes in sound level were investigated for two types of stimuli, one whose pitch was hypothesized to be analyzed on the basis of “place” and the other whose pitch was presumed to be related to “periodicity.” Eight subjects made alternate monaural pitch matches with four stimuli; pure‐tone, dc pulses (both with energy at the fundamental frequency‐place pitch), and two varieties of ac pulses (with little energy at the fundamental frequency‐periodicity pitch). The fundamental in all cases was approximately 100 cps. Generally, pitch shifts were small, although there were marked differences in the extent and in the variability of the pitch matches as a function of the type of stimulus. Partly because of large differences between subjects, statistically significant shifts were found for only two of the stimuli, one “place” and one “periodicity” stimulus. These shifts were similar in direction, with the pitch shifting downward as level is first increased, then upward with a further increase. Although there appears to be little in the data to support duality of pitchanalysis, it is possible that all stimuli were perceived on the basis of periodicity. However, it seems difficult to handle pitch shifts with changes in level within the confines of “periodicity” analysis.
33(1961); http://dx.doi.org/10.1121/1.1908884View Description Hide Description
Widely divergent amounts of perstimulatory adaptation have been reported previously, possibly due to differences in technique of measurement. One method often used determines perstimulatory adaptation from a series of simultaneous binauralloudness balances between a continuous stimulus in the adapting ear and a stimulus intermittently presented to the test ear. The present study attempts to evaluate the effect of the characteristics of intermittency of the test stimulus upon the measured adaptation in the adapting ear. With a 4000‐cps adapting tone presented at 75‐db sensation level, 16 combinations of on‐ and off‐duration of the test tone were investigated using 11 listeners. For all experimental conditions the adapting curves showed the same general shape, with a rapid initial decline, followed by a more gradual decline reaching asymptote after 5 to 6 min. As the on‐time of the test stimulus increased, less adaptation was seen, except for off‐times of 30 sec or greater where on‐time no longer influenced adaptation. This is interpreted as indicating that greater amounts of adaptation took place in the test ear as on‐time was lengthened; but for the stimuli used, 30 sec was sufficient for recovery to occur. In general, however, variation in on‐time produced greater changes in measured adaptation than did similar changes in off‐time of the test stimulus.
33(1961); http://dx.doi.org/10.1121/1.1908886View Description Hide Description
The phenomenon of “contralateral remote masking” is described: the elevation of threshold sensitivity to a low‐frequency tone in one ear produced by a high‐frequency band of noise in the other. Evidence is presented indicating that this effect is mainly attenuation induced by reflex activation of the stapedius muscle, although probably some central masking is also involved. The strength of the activation (1) increases linearly with sound pressure level of the arousal noise beginning at about 85 db SPL, (2) decreases linearly with frequency level of the noise, and (3) gradually decreases with time (i.e., adapts), reaching an asymptote after about 3 min. Individual differences in reflex activity could not be explained in terms of differences in resting thresholds. The relation of this phenomenon to ipsilateral remote masking and loudness adaptation is studied, and its effect on temporary threshold shift and on loudness judgements at high intensities is discussed.
33(1961); http://dx.doi.org/10.1121/1.1908888View Description Hide Description
This paper examines some sequential effects found in the behavior of observers in the signal‐detection situation. The Tanner‐Swets‐Green model treats the subject as a stable decision maker operating on information from a noisy but unbiased transducer. It is suggested that this model may be profitably replaced by one in which a bias is introduced by a simple finite‐state machine which makes the subjects' behavior a function not only of the present stimuli but also of past stimuli and responses.
33(1961); http://dx.doi.org/10.1121/1.1908890View Description Hide Description
A correlator detector is considered which processes the outputs of two identical collinear arrays of uniformly spaced elements. When the input signal is sinusoidal, the mean system output is bounded by the product of the space factors of the arrays. Complex amplitude factors are introduced following each element, and it is shown how to choose them in order to optimize the main lobe‐width‐side‐lobe‐level relationship of the space‐factor product or envelope. In addition, it is proved that the use of amplitude factors for improving envelope resolution gives rise to a signal‐noise degradation relative to the corresponding uniform amplitude system. Various numerical results are given, including the fact that the optimum system provides an envelope main‐lobe‐width reduction of approximately thirty percent when twenty or fewer elements appear in each array.
33(1961); http://dx.doi.org/10.1121/1.1908892View Description Hide Description
“Quasi stereophony” is defined as the reproduction over two or more loudspeakers (or binauralearphones) of different sound signals derived from a single audio signal. The purpose of quasi‐stereophony is to create (from a single audio signal) an illusion of spatially distributed sound sources. Quasi‐stereophonic reproduction does not permit correct localization but does share with true stereophony the properties of “depth” and “ambience” which are important attributes of stereophony (for the casual listener perhaps even more important than correct localization).
This paper describes a new filtering method for producing quasi‐stereophony. In contrast to earlier proposals, the new filtering method leaves the amplitude spectrum of the sound intact. The same kind of filter has also been used for generating “colorless” artificial reverberation. Experimental results indicate that both quasi‐stereophony and artificial reverberation can be achieved without spectral distortion.
33(1961); http://dx.doi.org/10.1121/1.1908894View Description Hide Description
Measurements are presented of the noise produced by a 1.5 in. diameter air jet, with an exit Mach number of 0.66, impinging perpendicularly on a plane, rigid plate. The over‐all sound power output increased rapidly, as the nozzle‐to‐plate separation distance was decreased. The over‐all sound powergenerated, when the plate was 2 diam from the nozzle, was 10 db greater than that produced with the plate removed. For a 2‐diam plate separation the over‐all sound‐pressure levels (SPL's) (measured at a radius of 24 nozzle diameters from the center of the jet exit in the horizontal plane through the jet centerline) were 15 to 18 db greater than those produced at corresponding positions with the plate removed, while for a 20‐diam separation, the increase varied between 2 and 7 db. The spectrum of the noise changed as follows as the separation distance was increased: (a) the peak frequency decreased, (b) the pronounced peak changed to a broad one, and (c) the magnitude of the peak decreased.
33(1961); http://dx.doi.org/10.1121/1.1908896View Description Hide Description
An amplitude‐modulated single‐sideband waveform can be separated into two parts, viz., an amplitude or “envelope” term, and a frequency‐modulation term. Some existing proposals for the use of the frequency‐modulation part in speech compression schemes have been investigated, and although these schemes were found to be fallacious, some results having possible application to speechformant tracking were obtained.
33(1961); http://dx.doi.org/10.1121/1.1908898View Description Hide Description
A thin rectangular liquid jet impinges on the apex of a rigid wedge and, under suitable circumstances, sets itself into any of a number of modes or “stages” of steady‐state transverse oscillation; any mode has associated with it a pattern of vortex production. Excerpts from motion pictures show sequences of jet configurations corresponding to the different modes of oscillation. In a photographic history depicting the buildings of oscillations in an initially quiescent jet, particular interest is attached to the fact that oscillations appear beforevortices have developed. Observations from these photographs and results from an earlier paper are compared with predictions of recent theories of edge tone production.
33(1961); http://dx.doi.org/10.1121/1.1908900View Description Hide Description
The geometrical or high‐frequency approximation to solutions of the two‐dimensional wave equation in an inhomogeneous medium is considered. A new ordinary differential equation for a quantity which is inversely proportional to the geometrical high‐frequency field intensity (the square of the field amplitude) is derived. This equation, along with the standard ray and phase equations form a system form which a complete wave solution in the high‐frequency asymptotic limit can be calculated numerically, e.g., through the use of a differential analyzer. The examples of a homogeneous medium and a plane stratified inhomogeneous medium are discussed, and the results of the preceding analysis are verified in these two special cases.
33(1961); http://dx.doi.org/10.1121/1.1908902View Description Hide Description
The results of measurement of the acoustic impedance of some viscoelastic fluids in circular tubes are presented. The fluids studied include water, petroleum oil, glycerol, silicone fluid, milling yellow solution, and agar solution, the last three showing elasticeffects. One of these elasticeffects is the changing of the acoustic reactance from the inertance type to the compliance type. Measured results are presented for frequencies from 3 cps to 300 cps and for tubes having radii in the range 0.0172 cm to 0.354 cm. Methods are presented for determination of the complex coefficient of shear viscosity from the acoustic impedanceproperties and examples are given.
33(1961); http://dx.doi.org/10.1121/1.1908904View Description Hide Description
In the calculations concerning the vibrations of thin elastic shells, several approximations called “shell‐theories” have been introduced in an attempt to simplify the mathematical formalism. In this paper one of these approximation methods, known as the Epstein method, is used to obtain the reflected field produced by a plane wave impinging on an infinite plate. The results are given to first order in kh, where k is the wave number and 2h is the thickness of the plate. A comparison with the exact results shows exact agreement to first order.
33(1961); http://dx.doi.org/10.1121/1.1908906View Description Hide Description
Structural fatigue caused by combined sinusoidal and random vibration is of concern in flight proofing of electronic equipment, since such excitation is required by many current ballisticmissile equipment specifications. How to predict the fatigue under this vibratory excitation is discussed with particular reference to the relationship of combined vibration to a single sinusoidal substitute.
Equations and curves for obtaining the sinusoidal substitute are presented and their application illustrated on a single resonator system.
33(1961); http://dx.doi.org/10.1121/1.1908908View Description Hide Description
The vibration analysis of shallow spherical shells is extended to (a) frequencies of the order of magnitude of the first thickness‐shear mode in an infinite plate and (b) moderately thick shells. A tenth‐order system of three uncoupled differential equations is derived, which govern the nonsymmetric dynamic deformation of a shallow spherical shell subjected to arbitrary time‐dependent surface loads, and separable solutions are obtained in terms of Bessel functions. As an example, a frequency equation is deduced for the determination of natural frequencies higher than those accurately predicted by the classical theory for free vibration of a shallow spherical cap with a clamped edge.
33(1961); http://dx.doi.org/10.1121/1.1908910View Description Hide Description
A method is presented for computing acceleration, deflection, and velocity response of a damped, single‐degree‐of‐freedom isolator, subject to a random disturbance between finite frequency limits. A closed‐form solution is obtained for the case where the power spectral density is constant within any given bandwidth. The results are plotted in such a manner that the responses to a variable power spectral density function can readily be computed.
33(1961); http://dx.doi.org/10.1121/1.1908912View Description Hide Description
Sound‐speed measurements and corresponding oceanographic data were obtained aboard the bathyscaph TRIESTE to a depth of 5760 m. Subsequent measurements were made in shallow northern waters by lowering the velocimeters and Nansen bottles from a surface ship. Near‐surface measurements agree with the laboratory data of either Del Grosso or Wilson. One set of the deeper measurements aboard the bathyscaph indicates a general agreement with Wilson, but the measured values were less than values computed by Wilson's equations and this difference increased with depth. The acceleration of gravity was measured at a depth of 2286 m and the increase with depth was in rough agreement with the equation in H. V. Sverdrup's The Oceans (Prentice‐Hall, Inc., Englewood Cliffs, New Jersey, 1949).
33(1961); http://dx.doi.org/10.1121/1.1908915View Description Hide Description
The problem of propagation of low‐frequency sound in shallow water over several layers and a basement is approached through an image theory in which the first thin layers are lumped together with the water layer. This effective structure is assumed on the basis of expected good transmission through a thin layer (less than one wavelength thick having acoustic properties not very different from those of its bounding media. The required reflection coefficient at the lower boundary of the lumped layer, expressed as a function of range and the order of the image source of a ray, is taken in the form given by Abelés for reflection of plane waves from a layered system. The theory yields calculated transmission loss functions which show reasonable agreement with experimental results. Comparisons of theoretical loss and data obtained at Panama City, Florida, are made over the frequency range 6–244 cps.