Volume 35, Issue 2, February 1963
Index of content:
35(1963); http://dx.doi.org/10.1121/1.1918429View Description Hide Description
Masked thresholds were obtained by the method of behavioral audiometry fur each of four cats at eleven frequencies over the range from 125 to 16 000 cps. Bands of noise, which were approximately one to two octaves wide, as well as two broad bands of noise, were used as masking stimuli. Experiment I shows that the function relating masking to noise level has exactly the same form for cat as for man; the signal to‐noise ratios, however, are greater for cat than for man. The critical ratio (K) is defined as the ratio of signal power to spectrum level at the masked threshold. The function relating K to frequency was determined for cat in experiment II; tiffs function parallels that for man, but lies 4 to 5 dB above it at most frequencies. The masking data for the cat are shown to be consistent with measurements of frequency discrimination for this animal.
35(1963); http://dx.doi.org/10.1121/1.1918430View Description Hide Description
Principle, theory, and an experimental development program for a new type of siren, capable of generating wide‐band noise, are discussed with special emphasis on its application to the economic and realistic simulation of high‐intensity jet and rocket‐noise environments. In contrast to conventional sirens with a single rotor, this wide‐band noise siren uses a series of overlapping, slotted rotors rotating at different speeds to produce the modulation of an air flow through a nozzle. The sound fieldgenerated by this siren can, mathematically, best be approximated by an “almost periodic” function; that is, a function whose spectrum is a line spectrum but with infinitely many lines in each interval of frequency. A series for the power spectrum can be derived and used to guide the design. For most practical purposes, the resulting acoustic field radiated by the siren represents random noise. Data on acoustic power, spectrum, efficiency of noise generation, and fine structure of the noise for various experimental siren types are presented. The potential value of this principle for large‐scale installations (sonic fatigue and missile‐component testing and bio‐acoustic applications) is evaluated.
35(1963); http://dx.doi.org/10.1121/1.1918431View Description Hide Description
The finite size of a transducer‐sensing element limits its space resolution of a pressure field associated with a local turbulent flow. Such pressure fields are translated at a speed comparable to the characteristic velocity of the flow. Consequently, a lack of resolution in space causes an apparent inability to resolve in time. This problem—an example of the mapping of a random function of several variables by a linear operator—is examined here. With the help of a formalism which has been previously discussed and of some recent experimental information about the spatial structure of turbulentpressure fields in boundary layers, the mapping or distortion of statistical quantities associated with the second‐order moments of the pressure field is given. The attenuation of the frequency spectral density and of the cross‐spectral density is given explicitly in table form and in asymptotic form. The numerical results indicate that the attenuation caused by the finite size of transducers is generally more severe than previous computations had suggested, mainly because the lateral correlation of pressure is highly frequency‐dependent, a typical turbulent pressure‐wave component being inclined to the stream direction at roughly 45 degrees.
The results are applied to an evaluation of contemporary measurements of turbulentpressure fields in shear flows. It is shown that the transducer size used introduces undesirable large errors in these measurements, which lead to doubts even about the magnitude of the intensity of turbulentpressurefluctuations.
Asymptotic formulas for the attenuation of large transducers are given which yield estimates of the degree to which a flush‐mounted sonar receiver immersed in a boundary layer is able to reject the background noise provided by turbulentpressurefluctuations.
35(1963); http://dx.doi.org/10.1121/1.1918432View Description Hide Description
The following interpretation of the word frequency effect—the tendency for common words to be perceived correctly at much lower speech to‐noise ratios than uncommon words—was proposed and verified experimentally. When a stimulus word is only a few decibels below its intelligibility threshold, many of its features—the number of syllables, for example—are still perceived correctly. If enough such features are heard, then only a small number of English words will be consistent with them. Subjects' incorrect responses will be taken from that small set of words, and furthermore they will usually be the relatively common words in the set. If the stimulus is an uncommon word and there are common English words that are phonetically quite similar to it, then these common words will usually be given as (incorrect) responses except at quite high speech‐to noise ratios. This interpretation is confirmed by an analysis of errors in an articulation test. The results show that all subjects tend to give the same incorrect responses, that these incorrect responses are usually more common than the stimuli to which they were given, and that, when a word is not confused with another in this way, its threshold is relatively low regardless of its frequency of occurrence in English usage.
35(1963); http://dx.doi.org/10.1121/1.1918433View Description Hide Description
An automatic system is described for high‐speed extraction of speech parameters. Data obtained from a high‐resolution, real‐time spectrum analyzer of the correlation type are converted to digital format and processed by a large‐scale computer. An effective program has been written in FORTRAN and FAP languages for the extraction of spectrum power, amplitudes and frequencies of the first three formants, and other functions of these parameters. Typical output data are presented.
35(1963); http://dx.doi.org/10.1121/1.1918434View Description Hide Description
The elastic constants of three compound semiconductors,ZnS, PbTe, and GaSb, were measured at 29°C using the ultrasonic pulse‐echo technique. The constants are found to be
c 11 c 12 c 44
(in units of 1011 dyn cm−2)
ZnS 9.76 5.90 4.51
PbTe 10.72 0.768 1.300
GaSb 8.97 4.12 4.48.
These results are compared with previously reported measurements made by other techniques.
35(1963); http://dx.doi.org/10.1121/1.1918435View Description Hide Description
Axial vibrations of whirling bars are studied by using undeformed or Lagrangian coordinates. In general, whirling lowers the natural frequencies of the bars. In the static case, an interesting and previously unreported result is obtained as a special case of the dynamic problem. When the angular velocity of the bar approaches certain critical values, “static resonances” or instabilities occur, and the axial displacements everywhere in the bar tend to become unboundedly large. These resonances are explained physically. The same problem is also solved by using deformed or Eulerian coordinates, and it is shown that the static resonances are concealed in such an analysis.
35(1963); http://dx.doi.org/10.1121/1.1918436View Description Hide Description
Within the framework of the normal‐mode approach, a solution is presented for the transient response of an arbitrary continuous structure to stationary forcing fields. No limitation is imposed on the types of forcing correlations considered; therefore, the solution is believed to be more general than those previously reported in the literature. It is shown that under stationary excitation the transient response is nonstationary, and it approaches the stationary solution of Powell in the limit as the time of excitation increases. An additional verification of the general theory is provided in that it is reducible to the known result given by Caughey and Stumpf for a single‐degree‐of‐freedom system. The general solution is further simplified when the forcing field is also homogeneous. The statistical dependence in the response of different modes is illusstrated in an example of a simply supported uniform beam.
35(1963); http://dx.doi.org/10.1121/1.1918437View Description Hide Description
The response to vibration of simply supported damped beams has been determined theoretically. Expressions have been derived for the transmissibility and input impedance and, in one case, the transfer impedance of such beams vibrating in their transverse modes, and for the transmissibility of one‐ and two‐stage systems of beams loaded by lumped elements of mass. The effect of varying beam damping is considered in detail.
35(1963); http://dx.doi.org/10.1121/1.1918438View Description Hide Description
The problem of wave propagation in an infinite piezoelectric plate belonging to crystallographic class C 6v , situated between shorted electrodes, is rigorously analyzed for two special orientations of the sixfold axis. The solution is derived using the linear piezoelectricequations. The analysis shows that for a given frequency and wave number in the propagation direction there are three independent solutions of the differential equations, and, furthermore, that these three solutions couple at the traction‐free boundaries of the plate. The dispersion curves can be calculated from the resulting transcendental equations. It has previously been shown that the resonant frequencies at infinite wavelength of a plate, in which the electromechanical coupling factors are high, deviate considerably from those of the purely elasticsolution. Thus the dispersion spectrum calculated from this analysis will deviate considerably from that of the purely elasticanalysis. It is shown in an appendix that the solution for the plate of infinite width is also applicable in the case of the plate of infinitesimal width.
35(1963); http://dx.doi.org/10.1121/1.1918439View Description Hide Description
A quantitative theory for backscattering or reverberation caused by the irregular sea surface is developed and compared with experimental results. Experimental results for forwardscattering or surface loss are also presented to compare with earlier theory.
35(1963); http://dx.doi.org/10.1121/1.1918440View Description Hide Description
The acoustic characteristics of a hydrogen‐oxygen explosive mixture are investigated in an attempt to find a substitute for solid explosives (e.g., TNT) which will achieve a simple underwater acoustic pulse uncomplicated by bubble oscillations. Data are shown relating the intensity and form of the acoustic signal to the proportions and amounts of gas mixture, point of ignition, hydrostaticpressure, length to‐diameter ratios of chambers, and various chamber appendages and terminations. It is shown that for certain conditions the bubble pulse can be almost suppressed or dissipated. Some experimental techniques and difficulties associated with practical application are included.
- LETTERS TO THE EDITOR
35(1963); http://dx.doi.org/10.1121/1.1918441View Description Hide Description
35(1963); http://dx.doi.org/10.1121/1.1918442View Description Hide Description
The tube method has been used to measuresound absorption and velocity in pure oxygen at 23.5°, 208.7°, and 301.3°C. At the elevated temperatures, the absorption was in excess of the classical value. This was attributed to thermal relaxation of the vibrational energy. At 208.7° and 301.3°C, the peak value of the relaxation‐absorption curve is estimated to fall at 100 ±4 and 175 ±13 cps/atm, respectively. These values correspond to relaxation times of 1.71 × 10−3 and 1.01 × 10−3 sec and are considerably longer than would be expected by interpolation from previous data at higher and lower temperatures.
35(1963); http://dx.doi.org/10.1121/1.1918443View Description Hide Description
Photographs are presented showing the principal effect of dynamic asymmetries on shell vibrations: splitting of the natural frequencies into two nearby values, with each frequency having different locations for its nodes. The experimental procedure is briefly outlined.
35(1963); http://dx.doi.org/10.1121/1.1918444View Description Hide Description
Part of the attenuation observed in the propagation of acoustic energy at frequencies below 1 kcps through the deep ocean is attributed to a relaxation absorption by an eddyviscosity mechanism. A relaxation frequency of about 630 cps is estimated from the available data, and an absorption dependence on frequency of slightly more than the first power is estimated for the process near the relaxation frequency.
Determination of the Power Radiated by a Piston‐like Underwater Sound Transducer from Near‐Field Axial‐Pressure Measurements35(1963); http://dx.doi.org/10.1121/1.1918445View Description Hide Description
This letter describes a method for obtaining the acoustic power radiated from a pistonlike underwater soundtransducer, where only near‐field measurements are possible due to boundaries. This method can be applied, for example, in the determination of the power radiated from underwater soundtransducers being calibrated in small pressure tanks. It involves measurements of the acoustic pressure along the axis of the source.