Volume 31, Issue 1, January 1959
Index of content:
31(1959); http://dx.doi.org/10.1121/1.1907606View Description Hide Description
Two captive porpoises (or bottle‐nosed dolphins) were given a variety of tests for their ability to perceive and avoid objects in water by reflected sound signals. The tests were made in a specially constructed anechoic pool in which visual perception was eliminated by the turbidity of the water. The results showed that trains of underwater sound pulses were emitted periodically by the animals in this situation whether they were swimming or at rest. The signals became more numerous when reflecting targets were lowered into the water. Sound pulses could also be triggered by surface splashes, but they were not sustained after a splash if no target was present. These porpoise sonarnoises are similar in some ways to man‐made sonar, although they differ from it in several important characteristics.
The porpoises would locate and approach food fish by this method but would avoid nonedible objects. They were also able to swim accurately through a field of obstructions without colliding with them—even in darkness. The ability to avoid obstacles by sound did not appear to be affected by artificial jamming noises.
31(1959); http://dx.doi.org/10.1121/1.1907617View Description Hide Description
The identification of the frequency of a tone was examined under four training procedures. The training procedures attempted to determine whether identification of tonal frequency could be improved by attaching distinctive motor responses to elements of the display. For relatively short training periods, such procedures were little more effective than simply providing verbal information of the correctness of the identification.
31(1959); http://dx.doi.org/10.1121/1.1907619View Description Hide Description
Poststimulatory threshold shift has been measured for 1000‐cps stimuli of varying duration and sensation level, turned on and off either without audible transients or abruptly. The measurement was performed by means of test stimuli of varying duration and following the prime stimulus at various time intervals. The results show that the poststimulatory threshold shift depends little on the duration of the test stimulus, but that it depends critically on the time interval between the termination of the prime stimulus and of the test stimulus. The poststimulatory threshold decreases as this time increases. The poststimulatory threshold increases with sensation level and it depends in a complex fashion on the duration of the prime stimulus. The time pattern of the threshold changes radically when a gradual cutoff of the prime stimulus is replaced by an abrupt one. Replacing the 1000‐cps tone by a random noise leads to the same result as an abrupt cutoff.
31(1959); http://dx.doi.org/10.1121/1.1907604View Description Hide Description
The effect of symmetrical speech peak clipping upon speech intelligibility in noise was examined. Over a wide range of conditions, intelligibility is independent of the level of peak clipping if the postclipping speech power is held constant. Peak clipping thus achieves protection of the ear against painfully loud speech without demanding penalty of intelligibility. Indeed, under restricted ranges of conditions, peak clipping may actually improve intelligibility with a constant speech power.
31(1959); http://dx.doi.org/10.1121/1.1907605View Description Hide Description
The scattering, and hence attenuation, of plane dilatational and shear waves due to random orientation of the grains is calculated on the assumptions that (1) the wavelength is large compared to the grain size and (2) the variations in elastic moduli from grain to grain are small compared to the average values of these moduli for the bulk material. Previous treatments of this problem are incomplete; for incident dilatational waves they are valid for a fluid (rigidity modulus=0), and for incident shear waves they describe inadequately the angular dependence of the intensity of the scattered shear waves. It is shown in this paper that for both of these types of incident waves both the scattered dilatational and shear waves are produced; the energy carried away from the incident beam by the latter is times that by the former. Here Vl and Vs are the velocities of the dilatational and shear waves, respectively. The agreement between the theoretical and observed values of the attenuation coefficient is much better than that previously obtained by Mason and McSkimin.
The attenuation due to local fluctuations in the density of a medium caused by thermal agitation is also calculated. For solids and liquids this effect is completely negligible. However, for a gas near its critical point, this attenuation is appreciable and rises sharply as the critical point is approached.
31(1959); http://dx.doi.org/10.1121/1.1907607View Description Hide Description
In order to evaluate the method for measurement of the ultrasonicpressure amplitude in progressive waves which was developed by Loeber and Hiedemann for stationary waves, a comparison between it and two other optical methods was made. Pressure measurements under identical conditions by the three methods are given. Experimental considerations for the use of optical methods for measurement of sound pressure amplitudes are discussed.
31(1959); http://dx.doi.org/10.1121/1.1907608View Description Hide Description
Measurements of ultrasonic absorption have been made in dry , and in with hydrogen admixtures, between 19°C and 146°C and at pressures from 0.1 to 650 atmos, and frequencies from 300 kc to 7 Mc. The absorption is described within experimental error by the relaxation formula throughout this range using either (a) a single relaxation time and a single relaxing specific heat representing the total of the vibrational modes, or (b) two relaxation times somewhat more than a factor of two apart, and two relaxing specific heats given by certain combinations of the component vibrations. The relaxation time, or times, varies inversely as the density throughout. The maximum absorption per wavelength comes at 32 kc/atmos at 25°C. Addition of hydrogen shifts the relaxation time by 110 kc/percent by volume of ; this amount is in agreement with the shift determined by Knudsen and Fricke, provided their data is interpreted as being by volume and not by weight. The “classical absorption” at low pressures is 1.5 times the theoretical value.
31(1959); http://dx.doi.org/10.1121/1.1907609View Description Hide Description
Acoustic propagation in plate‐shaped solid delay lines is analyzed using diffraction theory. The analysis yields the dependence of group velocity (time delay), phase velocity, pulse shape, pulse rise time, output amplitude ripple spacing, and ripple amplitude on the parameters of frequency, delay line length, free‐medium soundvelocity, and plate thickness. Data are presented corroborating the theory quantitatively for time delay dependence on frequency and plate thickness, ripple spacing dependence on frequency and mode, and pulse rise‐time dependence on frequency. The diffraction theory is compared to the wave equation plane‐wave solutions for infinite plates. The results of the diffraction analysis are more appropriate for the description of sound propagation in plate‐shaped delay lines.
31(1959); http://dx.doi.org/10.1121/1.1907610View Description Hide Description
The study of the asymmetry of the light diffracted by ultrasonic waves progressing with moderate amplitudes in liquids is used to detect and measure the finite amplitude distortion in water and carbon tetrachloride. A theory is developed which relates the distortion present to available isothermal compressibility measurements. This development includes the presence of absorption.
Measurements of the second harmonic were made in water at 2 Mc and 4 Mc, at distances up to 50 cm, and at peak ultrasonicpressures between zero and 1.5 atmos. More detailed measurements were made in carbon tetrachloride at frequencies of 2 and 3 Mc, at distances up to 21 cm, and at pressures up to 0.7 atmos. For both liquids, good agreement is obtained between the experimentally measured values and the theoretically predicted results.
Small amounts of distortion were easily measured. The second harmonic could be measured when the intensity of the harmonic was less than 0.04% of the intensity of the fundamental. Distortion could be measured when the fundamental was much less than 0.2 atmos.
31(1959); http://dx.doi.org/10.1121/1.1907611View Description Hide Description
In the research reported here an attempt has been made to discover by experiment what physical assumptions and approximations are appropriate in the theory of cavitation microstreaming, especially for cavitation bubbles located near solid boundaries. A systematic investigation of the phenomenon has been made and its dependence on certain parameters (e.g., amplitude of sound) has been determined. The investigation has disclosed that as the sound amplitude is varied, other conditions remaining the same, the streaming changes discontinuously through several stable regimes. It appears that in order to account for the generation of vorticity one needs to assume different conditions at the boundaries for each regime. For at least one regime, a theoretical model due to Nyborg seems to be applicable; comparison made with experimentally determined streaming velocities.
Statistical Evaluation of Near‐Field Sound Pressures Generated by the Exhaust of a High‐Performance Jet Engine31(1959); http://dx.doi.org/10.1121/1.1907612View Description Hide Description
The operation of modern, high‐performance turbojet engines creates an environment that promotes rapid structural fatigue in many areas in jet aircraft. To facilitate the simulation of a valid test environment and, therefore, enable fatigue resistant structure to be readily developed, the characteristics of the near‐field sound pressures created by the jet stream must be established. This study is concerned with the establishment of the level and the rate of occurrence of peak pressures in relation to the commonly measured rms level. A method is presented which enables the determination of these qualities from rms sound pressures recorded on magnetic tape. Also included are typical results obtained from near‐field sound pressuresgenerated by the jet exhaust of a General Electric CJ805 turbojet engine. Noted conclusions: (1) Occurrence of peak pressures does not follow a Rayleigh Distribution. (2) Peak pressure distribution is not altered by the physical position related to the jet stream nor does it appear to be a function of frequency. (3) A maximum ratio of peak to rms pressure is shown to exceed four, but a physical limit is not established.
31(1959); http://dx.doi.org/10.1121/1.1907613View Description Hide Description
The results obtained in a previous paper, Part I of this study, are summarized and used to derive the theory of vibration and sound insulation of systems with many natural frequencies, such as machines with more than one resonance mounted on resonating springs, a motor spring‐mounted on a shell, or plates at higher frequencies. At lower frequencies, the quality of the insulation is a function of the ratio of the square of the lowest fully excited resonant frequency to that of the frequency of the force. Weakly excited natural modes cause spikes in the transmission curve. Resonances in the vibrating masses may annihilate completely the effect of the vibrator mounting, whereas resonances in the springs are harmless as long as the mass is small in comparison to that of the machinery. Spring mounting a vibrator on a shell or plate reduces the noise level considerably. Any measure that increases the fundamental natural frequency to the housing helps to improve vibration insulation. Mounting the vibrator in a rigid cage or onto a rigid girder seems to be expedient; and welding these along a whole circumference or along their whole length onto the housing, to simulate a one‐dimensional vibration pattern, is expedient.
The noise radiation of a shell or plate to greater distances can be reduced by increasing the mass of the radiating shell or plate and by decreasing its size sufficiently to reduce the number of modes that radiate sound. Damping diminishes the sound radiation at lower frequencies, but has very little effect on it at higher frequencies.
American Standard Recommended Practice for Laboratory Measurement of Air‐Borne Sound Transmission Loss of Building Floors and Walls31(1959); http://dx.doi.org/10.1121/1.1907615View Description Hide Description
The American Standards Association recently issued the standard which is reproduced below, complete and unabridged, with their permission.
This new standard can be regarded as an up‐to‐date revision of the ASTM Tentative Recommended Practice for the same measurement (ASTM E 90‐50T, later ASTM E 90‐55), which was printed in this Journal [23, 686, (1951)]; in 1955 the “Tentative” was deleted from the title, the text remaining unchanged.
Some other countries, including Britain and Germany, have standards on this topic (British Standard 2750:1956, and German Standard DIN 52210, both obtainable from the American Standards Association), and a draft exists for an international standard; the latter includes provisions for field measurements as well as laboratory measurements, and for measurements of impact as well as air‐borne sounds. At present ASTM Committee E‐6 is working to prepare a new Recommended Practice for the measurement of the sound transmission loss of building structures, and will consider the inclusion of field and impactmeasurements.
31(1959); http://dx.doi.org/10.1121/1.1907616View Description Hide Description
Parties are classified as loud or quiet, and the distinction is shown to depend often upon a critical acoustic relationship rather than upon the guests themselves. An explicit formula is found for the maximum number N of well‐mannered guests compatible with the quiet party. When this number is exceeded, the party will become a loud one within a calculable time T.
- LETTERS TO THE EDITOR
31(1959); http://dx.doi.org/10.1121/1.1907618View Description Hide Description
- PROGRAM OF THE FIFTY‐SIXTH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session A. Speech Communication: Perceptual Aspects
- Contributed Papers
31(1959); http://dx.doi.org/10.1121/1.1930098View Description Hide Description
The purpose of this experiment is to determine the extent to which both duration and formant frequencies play a role in the identification of American English vowels. A resonance analog synthesizer was used to generate a number of cvc syllables in which the initial and final consonants were always /d/ and /s/, respectively. The durations of the vowels were given values from 20 to 500 msec. Two series of stimuli were recorded and were presented to listeners in random order. In the first test the frequencies of the first three formants of the vowel varied in several steps to encompass values appropriate for the front vowel series /iɪɛæ/, and the listeners were asked to identify the vowel as one of these; in the second test the back vowels /uᴜʌɑ/ were studied. The results indicated that the /iɪ/ and /uᴜ/ distinctions are affected only slightly by vowel duration whereas the distinctions /ɛæ/ and /ʌɑ/ are strongly influenced by duration. In a similar experiment in which the stimuli were mixed with noise, the influence of vowel duration was even more marked. The results are discussed in the light of existing data on vowel perception. (This research was supported in part by the National Science Foundation.)
31(1959); http://dx.doi.org/10.1121/1.1930099View Description Hide Description
The extent to which the ear is sensitive to phase changes in the speech spectrum is uncertain. Evidence has been obtained previously to indicate that the ear does detect relative changes of phase in certain sinusoidal complexes, while on the other hand, many successful coding devices for speech ignore phase information. Therefore, it is the purpose of this study to obtain additional knowledge on phase sensitivity. A computer simulation is described in which speech samples have been modified so as to retain their original amplitude spectrum, but to have arbitrary phase spectrum. The study was carried out by sectioning the voiced portions of a speech sample into pitch periods, analyzing each pitch period to determine its harmonic components, and resynthesizing the pitch periods with various arbitrary phase patterns. The distribution of peak factors (ratio of range to average absolute amplitude for each pitch period) was computed to characterize the envelope for each phase pattern. The subjective effect of various phase patterns was evaluated by listening tests.
31(1959); http://dx.doi.org/10.1121/1.1930100View Description Hide Description
Five vowel‐like sounds, /i/, /ɛ/, /æ/, /ɑ/, /ɔ/, and /u/, were recorded from a series vowelsynthesizer with both male and female formant center frequencies and fundamental frequency varied in six steps from 100 to 800 cps. Each stimuli was inflected upward one semitone. Listeners, trained in phonetic transcription, were asked to identify which vowel sound was played. The data show interaction between formant frequency and vowel identification suggesting that information may be conveyed by the frequency spacing between the fundamental frequency and formants as well as the center frequencies of the formants.
- Invited Paper
31(1959); http://dx.doi.org/10.1121/1.1936079View Description Hide Description
In attempting to synthesize speech by rule, one must take account of the fact that the perceptually discrete phonemes are typically encoded at the acoustic level into segments of approximately syllabic length. It is, therefore, not possible to synthesize speech by stringing together prefabricated phonemes. By taking advantage of knowledge about the acoustic cues for speech perception, however, one can write rules for synthesis in terms of the constituent phonemes plus a few rules of combination. Thus, the number of rules can approximate the number of phonemes rather than the number of syllables. Indeed, one can reduce the number of rules still further by writing them in terms of subphonemic dimensions, viz., place and manner of articulation and voicing. Several complicating factors make it impossible to achieve an ideal minimum. First, rules must be added to take care of certain prosodic and positional variations. Failure to do so not only affects naturalness, but also impairs intelligibility, even at the level of segmental phonemes. Second, it is necessary in a few special cases to have different rules for a single consonant phoneme (or dimension) before different vowels. This reflects the occasionally complex relation between phoneme and articulation on the one hand and sound on the other; presumably, this complication would not affect the rules of synthesis for an articulatory model.