Volume 34, Issue 10, October 1962
Index of content:
Demonstration of a Speech Processing System Consisting of a Speech Analyzer, Translator, Typer, and Synthesizer34(1962); http://dx.doi.org/10.1121/1.1909035View Description Hide Description
A public demonstration of a rudimentary speech processing system consisting of a speechanalyzer, language translator, type decoder, and speechsynthesizer was given at the Meeting of the Acoustical Society of America in New York, May 23–26, 1962. Each word spoken into the microphone was analyzed, converted to a syllable code, transmitted as a syllable code, synthesized to speech, or converted to the typed page either in the same language or other languages. The two language inputs were English and French, the synthesized speech outputs were English, French, and Spanish and the typed outputs were English, French, German, and Spanish.
34(1962); http://dx.doi.org/10.1121/1.1909037View Description Hide Description
A method is presented for increasing the aperture of solid, polygonal delay lines based on the method of linear programming. If it is assumed that there is no spreading of energy so that the acoustic beam may be defined by two parallel lines, the radii of the polygon can be computed which result in the maximum value of the aperture (the distance between the parallel lines) for fixed angles, subject to the constraints of the problem. Small variations in the angles which lead to a further increase the aperture can also be computed by linear programming. In addition, the method can be applied if the beam is assumed to spread in a linear fashion.
34(1962); http://dx.doi.org/10.1121/1.1909039View Description Hide Description
The acoustic intensity absorption coefficient of tissue of the central nervous system has been determined at the sound frequency of 1 Mc/sec in the temperature range from 2° to 28°C at incident sound intensities ranging from 5 to 200 W/cm2. The absorption coefficient exhibits an increase with increasing temperature and no variation with the acoustic intensity.
34(1962); http://dx.doi.org/10.1121/1.1909041View Description Hide Description
A theory is developed for the diffraction of light by ultrasonic waves of sufficiently high frequency, large amplitude, and/or large beamwidth that the emerging light wavefront is significantly amplitude modulated in addition to the phase modulation considered in the Raman‐Nath approach. The ultrasonic beam is considered to consist of N adjacent sections and the final diffractionspectrum to result from N successive diffractions. The diffraction orders emerging from a given section are considered to be sources for further diffraction by the next section. Only phase modulation of the separate plane waves (diffraction orders) is considered within a section. Refraction of light is not considered as such; it is characterized by successive redistribution of light in the diffraction orders. Numerical results are obtained by iterations using an electronic computer. These results are compared with measurements.
34(1962); http://dx.doi.org/10.1121/1.1909043View Description Hide Description
By using some approximations it is possible to get fairly simple approximate formulas for the number of resonance frequencies and for the point impedance of thin cylinders. Measurements of the resonance frequencies and of the number of resonances within a certain frequency range confirm the theoretical results.
34(1962); http://dx.doi.org/10.1121/1.1909045View Description Hide Description
The performance of an array is affected by the interaction of the array elements; this interaction is expressed as a radiation impedance. The radiation impedance for a circular piston in an infinite planar array is calculated as a function of distance between elements and the angle to which the array is steered.
34(1962); http://dx.doi.org/10.1121/1.1909047View Description Hide Description
The influence of loading of a test item on a rigid shock‐Machine platform was investigated to determine whether any peaks or notches are introduced into the acceleration shock spectrum which would cause excessive over or under testing at various frequencies. The results showed that the shock spectrum for a terminal peak saw tooth pulse is quite insensitive to shock‐machine loading. The very low frequency end (below 100 cps for a 6‐msec pulse) is most influenced, whereas the higher frequencies are virtually not influenced at all. The results also showed that even for a very high loading condition the primary spectrum is in all cases less severe than the residual spectrum.
Curves have been generated to depict both the primary and the residual spectrum for various resonators to shock platform mass ratios, for the case of a terminal peak saw‐tooth pulse. This particular shock pulse is of interest since the resulting shock spectrum satisfies the requirements of a number of U.S. Air Force missileelectronic equipment specifications.
34(1962); http://dx.doi.org/10.1121/1.1909049View Description Hide Description
A cross‐sectional taper is described which provides a bar with the possibility of uniform stress in longitudinal oscillation. It is shown that this kind of motion is a special case of a more general kind of motion in which it is possible by properly terminating the bar to make the displacement and the stress functions related polynomials of any desired order. A theory which describes the behavior of the bar in longitudinal motion is presented and compared with the results of experiment.
34(1962); http://dx.doi.org/10.1121/1.1909051View Description Hide Description
An elastic energy method is presented for finding the fundamental torsional vibration frequency of solid rectangular beams without boundary constraints. A Rayleigh‐Ritz method is employed, and the results, which were calculated on an IBM 650 computer, are shown. Also illustrated is the effect a change in the Poisson's ratio of the beammaterial has on the fundamental frequency. A comparison is made with results of torsional vibration tests performed at the National Bureau of Standards. This comparison indicates that the analytical method employed gives fairly accurate results.
34(1962); http://dx.doi.org/10.1121/1.1909053View Description Hide Description
An experimental investigation has been made of a transducer‐reflector combination consisting of a 90° conical reflector with a line transducer placed along the axis. Both far field and near‐field data indicate that, except at low frequencies, when a uniform line is used, the radiation patterns are equivalent to those from a uniform circular aperture. Evidently other effects compensate for the geometrical shading. By a suitable choice of line‐element connection, however, a shaded aperture distribution can be obtained. The horn reflector increases the sensitivity of the transducer but not to the extent predicted from the geometry. The interaction of elements composing the line transducer is increased by the addition of the conical reflector.
34(1962); http://dx.doi.org/10.1121/1.1909055View Description Hide Description
It is well known in sonar work that the pulse form of a direct echo from a target bears little relation to the form of the original signal. This is true even for regularly shaped bodies, such as a sphere. In this paper, the case of a homogeneous elastic sphere in water is examined theoretically and it is shown in comparison with experimental results, that the observed effects originate from vibrations induced in the sphere by the incident sound. Calculated results are presented for a variety of solid materials and it seems that echo forms could possibly provide information about the size and constitution of a sonar target.
34(1962); http://dx.doi.org/10.1121/1.1909057View Description Hide Description
The scattering strength of the sea surface was measured for a range of wind velocities, grazing angles, and frequencies, in octave bands in the frequency range from 400 to 6400 cps. An empirical equation was obtained relating the scattering strength of the sea surface to the above variables, for grazing angles below 40°. At low grazing angles, scattering of sound from a subsurface layer of isotropic scatterers, probably of biological origin, frequently masked the reverberation due to scattering from surface roughness. For a given wind speed, the scattering strengths measured in this study at grazing angles below 20° were appreciably less than those obtained by other observers at higher frequencies. At higher grazing angles, of the order of 40°, there was little systematic difference between the measurements made at high and low frequencies.
Acoustic Signal Detection by Simple Correlators in the Presence of Non‐Gaussian Noise. I. Signal‐to‐Noise Ratios and Canonical Forms34(1962); http://dx.doi.org/10.1121/1.1909059View Description Hide Description
The detection of underwater acoustic signals by simple auto‐ and cross‐correlation receivers in the presence of nonnormal, as well as normal background noise is examined on the basis of signal‐to‐noise ratios calculated from a generalized deflection criterion. Particular attention is devoted to the effects of impulse noise and mixtures of impulse and normal noise on system performance. Comparisons between system behavior vis‐à‐vis the two types of interference are made. For impulse noise equivalent in spectral distribution and average intensity to a Gaussian noise background it is found that the output signal‐to‐noise (power) ratios are related by the canonical expression , where Λ(⩾0) is the “impulse factor” and μ is the fraction (in average intensity) of the total noise background that is attributable to normal noise.Impulse noise always degrades system performance vis‐à‐vis normal noise in the autocorrelation reception of stochastic signals, characteristic of applications where passive receiving methods must be used. This degradation can be considerable [O(10 dB or more)] if the noise is highly impulsive (large Λ) and if large values of are required (for high accuracy of decision). On the other hand, when coherent (i.e., deterministic) signals are employed, so that cross‐correlation reception is possible, the degradation may be reduced essentially to zero (i.e., Λ → 0) under realizable conditions of operation. It is observed for impulsive, as well as normal noise backgrounds, that cross‐correlation receivers are linear in their dependence on signal‐to‐noise ratio, i.e., if sufficiently strong injected signals are employed. The analysis is carried out largely in canonical form, so that the general results for can be applied to other, special types of nonnormal noise backgrounds. Specific relations are included, along with a detailed summary of the principal results, showing the dependence of on , filtering, delay, noise and signal spectra, etc., for weak and strong inputs, little or heavy postcorrelation smoothing and for Gaussian as well as for impulse noise.
34(1962); http://dx.doi.org/10.1121/1.1909061View Description Hide Description
A series of experiments designed to elucidate the differences between pure tones and octave bands of noise in terms of their ability to arouse temporary threshold shift (TTS) is described. It is shown that the critical‐band concept is not relevant at high intensities, and that therefore the 10 dB correction factor currently employed, being based on the critical‐band concept, must usually be incorrect. Instead, it is demonstrated that the differences in TTS producing ability of tones and noises can be attributed mainly to their differences in ability to arouse the acoustic reflex; since noises cause more sustained contraction of the middle earmuscles than do pure tones, less TTS will usually be produced by noise. Because this difference in contraction, and hence in effective attenuation, is a variable function of frequency, intensity, and time, no one single decibel correction factor is appropriate. It is suggested that damage‐risk criteria for pure tones should be developed empirically by measuring TTS directly rather than by attempting to modify existing criteria for octave‐band noise.
34(1962); http://dx.doi.org/10.1121/1.1909063View Description Hide Description
The effects of monaural masking noise on loudness and on lateralization of clicks were compared for noise levels ranging between 35 and 65 dB SPL. For different fixed levels of unmasked clicks, sound‐pressure levels of the masked clicks were adjusted by the subjects to produce centered sound images. In a second experiment, similar procedures were used to generate equal‐loudness functions. For both loudness and lateralization, the effect of masking noise was greatest with high levels of noise and low intensities of click. In every case, the masking effect was greater for loudness than for lateralization.
34(1962); http://dx.doi.org/10.1121/1.1909065View Description Hide Description
Responses to glissando at 1500 cps at a 50‐phon level using trained and untrained listeners are analyzed. The sensitivity to this type of frequency modulation is surprisingly good. A constant of 5 cps change emerges indicating that sensitivity depends on a linear relationship between rate of change and stimulus duration. This constant is of the same order of magnitude as the traditional DL for pitch memory. The low observed relationship of glissando to either pitchdiscrimination by the constants method or with sensitivity to the presence or absence of sinusoidal frequency modulation indicates that probably a third distinct pitch factor is represented.
Free‐Field Threshold Shift and Temporary Threshold Shift Reduction as Measures of Efficiency of Ear Protective Devices34(1962); http://dx.doi.org/10.1121/1.1909067View Description Hide Description
The attenuation of two different earplugs and a helmet was measured by a procedure combining the free‐field threshold shift and the Békésy threshold tracking methods. Subjects were then exposed to impulse noise which would produce between 12 and 64 dB TTS at 6000 cps unprotected and their TTS was subsequently measured after identical exposure during which they wore the protective devices. Following this the subjects were exposed unprotected to continuous noise which would produce a TTS at 6000 cps within ±6 dB of the TTS they experienced after their unprotected exposure to the continuous noise. Finally, they were exposed to the same continuous noise while wearing the protective devices, and their TTS following exposure was determined.
It was found that the helmet afforded significantly less attenuation of low‐frequency sound than the ear plugs and roughly the same amount at higher frequencies. The three devices were found to be equally effective in eliminating the TTS following exposure to the continuous and impulsive noise. While the data did not lend themselves to definite conclusions regarding susceptibility to impulse and continuous noise, it appeared that susceptibility to one type of exposure was not necessarily significantly correlated with susceptibility to the other.
34(1962); http://dx.doi.org/10.1121/1.1909069View Description Hide Description
The application to organtones or other sustained musical sounds of the Stevens and the Zwicker methods for loudness‐level calculation from observed sound spectra is examined. It is found that both methods, although developed for dealing with noise of the continuous spectrum type, are applicable to organtones, a particular class of discrete component sound, subject to the use in both methods of accepted equal‐loudness contours. Examples of the loudness levels of some typical organtones are given. Questions needing further investigation are noted.
34(1962); http://dx.doi.org/10.1121/1.1909071View Description Hide Description
Relative differential intensive limens were obtained at three frequencies and two intensities for three stimulus patterns involving fluctuation of a steady tone and three patterns involving separate stimuli. The same observers were utilized in all cases; in Experiment I, signals were presented through earphones. Somewhat lower thresholds were obtained for fluctuating intensity patterns than for separate stimuli patterns, thresholds tended to be lower at the higher intensity level, and thresholds were lower for patterns allowing greater opportunity for comparison within the patterns. When similar thresholds were obtained for patterns presented through speakers (Experiment II), thresholds were smaller, the influence of intensity was smaller, and the influence of (presentation) pattern type was less consistent. Significance of the findings is discussed.
34(1962); http://dx.doi.org/10.1121/1.1909073View Description Hide Description
The most common device for analyzing sounds is the bandpass litter, usually with a bandwidth of one octave or a fraction there of, e.g., (one‐half or one‐third. Recently 1/3‐octave filters are also being used for vibration analyses. The effective bandwidth of these filters is affected not only by the obvious design parameters of filter cutoff/frequency and the number of resonant elements employed which determines the attenuation rate outside the passband but also on the slope of the noise spectrum which is being analyzed. This paper investigates the effect of all these variables on the effective bandwidth of a symmetrical bandpass filter by means of a normalized mathematical model and expresses the result in terms of bandwidth error. It is found (1) that the bandwidth error increases very rapidly for large values of spectrum slope; (2) that the bandwidth error is symmetrical about an equal energy per octave spectrum slope and that an optimum design point for attenuation at the filter band edges exists for minimizing bandwidth error as a function of spectrum slope.