Index of content:
Volume 30, Issue 7, July 1958
- PROGRAM OF THE FIFTY‐FIFTH MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session A. Noise Control
- Contributed Papers
30(1958); http://dx.doi.org/10.1121/1.1929927View Description Hide Description
Because of the sensitive nature of the environment, noise control criteria suitable for use in residences, differ from those applicable to commercial or industrial use. The trend to smaller, more compact, residential air conditioning systems has developed a need for noise control criteria which orient new product design in proper relationship with the ultimate environment. A series of products, having various levels of known customer acceptance, have been subjected to comprehensive acoustical analysis using the reverberation room technique to determine the sound power level (PWL) spectrum of each unit. A modified form of the Beranek NC contours has been found to provide an encouraging correlation with field experience. The best correlation is obtained from 1/3 octave band data which show up important changes in spectrum shape in addition to providing resolution of line spectra in terms of adjacent broad band noise. The largest variability in correlation has been found to depend on the pure tone content. In general, a plus 5‐db correction for line spectra exceeding the level of adjacent 1/3 octave bands seems to provide a more consistent agreement with field experience. The language barrier which the acoustical engineer must bridge when influencing product design is well known. This barrier can be made less formidable when control criteria can be presented relative to levels of customer acceptance. A moderately successful method of doing this has been found.
30(1958); http://dx.doi.org/10.1121/1.1929928View Description Hide Description
Attainment of acceptable air conditioning noise levels results from (a) low vibration and noise output, and (b) acoustic enevironment at installation. Commercially available equipment requires extraneous acoustic treatment for 5 to 50% of domestic installations (depending on building construction) in order to eliminate annoyance. Acoustical surveys of site conditions prior to installation permit most economical application of acoustic treatment and constructions. Classification of installation features and identification of respective transmission paths facilitate prior analysis of site conditions and assessment of treatment requirements. The most important transmission paths are through (a) the foundations, (b) the air discharge and return openings and ducting, and (c) the equipment casings. Effective methods for blocking transmission along each path by means of field applied treatments are discussed.
30(1958); http://dx.doi.org/10.1121/1.1929929View Description Hide Description
Quieting reciprocating compressors for refrigeration involves all the classical techniques and devices used in high‐speed electromechanical machinery—vibration isolation, balancing, mufflers, flexible connectors, etc.—and an important number of special techniques, involving thermodynamics with temperature changes of 200 to 300°F, pressure changes of the order 20 times atmosphere, high flow velocities, and media where only metallic components with very little inherent damping can be used. The working medium must be carried in pipes under pressure and the flow is modulated by complex pulses due to the high‐speed pumping and fast valves, which act like fast switches in a transmission line, generating noise. A study of acoustic transients (pulses) in pipes has been made on an analog scale model (linear acoustical pressures up to 100 dyne/cm2 114 db), then on an actual compressor (high nonlinear pressures up to 107 dyne/cm2) with different acoustical components (inductance, capacitance, resistance) in the line, and different input and termination impedances of the line. The influence of these components on pressure pulsation and outside noise, and how the analog helps to identify noise sources, which have same frequency and amplitude noise output (rpm, intake, exhaust) will be described along with structural noise transmission, through flexible “corrugated metallic hose” under different pressures used in refrigerating compressors. Mechanical impedance measurements carried out on compressor shells to find points with high mechanical impedance for optimum locations of exhaust lines will be shown, with vibration modes and resonant amplification of a typical compressor shell.
30(1958); http://dx.doi.org/10.1121/1.1929930View Description Hide Description
The noise reduction afforded by ducts or parallel baffles depends not only on the physical properties of the duct and its lining, but also upon the angle of incidence of the sound waves impinging on the ducts. The noise reduction per duct width or per unit length provided by a duct for sound waves impinging at several angles of incidence is found to vary with the length of the duct. The variation of noise reduction per duct width decreases with increasing duct length. For sufficiently long ducts, however, the noise reduction per duct width approaches a constant value For a fairly wide range of duct lengths, the noise reduction can be expressed as a + bl, where b is the noise reduction per duct width, l is the length of the duct measured in duct widths, and a is a source distribution correction factor. Several sets of experimental data are presented to show that a may be comparable in size to bl for typical duct lengths. Several implications of a noise reduction of the form a + bl are discussed. Procedures are given for estimating the noise reduction for ducts of finite length by use of Morse's theory and the data presented here. Some unsolved problems are discussed. [Supported in part by the U. S. Air Force under Contract Nos. AF 33(616)‐2151, AF 33(616)‐3335, and AF 33(616)‐3938.]
30(1958); http://dx.doi.org/10.1121/1.1936027View Description Hide Description
A field measurement program was performed to determine the background noise levels as a function of time and location for a number of selected community areas in one city. Measurements of background noise were obtained according to a randomized sampling schedule. Concurrent observations were obtained of street traffic flow and other significant noise sources. This paper discusses variations in mean background noise spectra with time and location and short term fluctuations in noise levels. Inferences are drawn for the number and disposition of necessary field measurements to establish background noise levels with desired precision. The degree of correlation between background noise levels and street traffic flow is discussed and related to a previously reported procedure for estimating the noise from vehicular traffic [J. Acoust. Soc. Am. 29, 779(A) 1957]. Comparisons are drawn between observed aircraft noise and background noise from all other sources. [This work was supported in part by the U. S. Air Force under Contract Nos. AF 33(616)‐3335 and AF 33 (616)‐3938.]
30(1958); http://dx.doi.org/10.1121/1.1929931View Description Hide Description
Field measurements have been made of the transmission loss of about a dozen commonly used partitions. The test conditions were exceptionally good so that the flanking problem was not serious. The test results can be summarized as follows: (1) For most of the partitions the measured transmission loss (TL) can be explained, in at least a qualitative way, in terms of the partition weight, stiffness and damping; the partition size; and the size and shape of the test rooms. (2) For most of the partitions the measured TL was appreciably less than the value one would predict if the wall stiffness is neglected. (3) For most of the partitions the measured TL was appreciably less than the values published by testing laboratories (which utilize the standard, ASTM E90‐50S test method).
30(1958); http://dx.doi.org/10.1121/1.1929932View Description Hide Description
There is considerable confusion in the public mind on the noise reduction efficiency of folding partitions. Some of the reasons for this state of affairs are (1) exaggerated performance claims by some manufacturers, (2) some unrealistic aspects of the method of testing used, (3) the large variation of data reported by the reference testing laboratories, and (4) the paucity of field data. This paper is a report of field measurements on a number of actual installations. The technique used, which has been applied to a large number of field tests of sound transmission, is very convenient and gives consistent data. The sound source consists of a portable playback of successive magnetic tape loops of half‐octave bands of random noise.Measurements are made in both source and receiving rooms with a sound level meter and half‐octave band filter. Data are usually obtained in the 106–150, 150–212, 212–300 … 3400–4800 cps bands, thus giving essentially the same measurement frequencies (125, 175, 250, … 4000) that are used in the reference laboratories. The most important parameter in the performance of the partitions is the effectiveness of the seals at the top and bottom as provided by sweep strips, if they are used at all. Using the noise reduction in the critical 600–825 cps band as a reference, for example, the figures vary between 7 to 22 db, depending mostly upon the seal. There are no conclusive differences among manufacturers. Acceptable acoustical isolation for meeting spaces can be made with folding partitions, provided (1) extreme privacy is not needed, (2) the spaces on each side are not too large, (3) efficient sweep strips are used with precautions in installation and maintenance, and (4) there is a reasonable background noise for masking.
- Session B. Speech Analysis and Intelligibility
- Invited Paper
30(1958); http://dx.doi.org/10.1121/1.1929933View Description Hide Description
The objectives of acoustical analyses of speech are generally to obtain economical descriptions of the important characteristics of speechsounds, together with the distributions of these characteristics. The relevant acoustical properties cannot be determined and quantified easily without reference to data on the processes of speech production and perception. An understanding of speech generation indicates the constraints that are imposed on the vocal output, whereas knowledge of the perception process suggests acoustical properties that are likely to be important for understanding speech, as well as the accuracy with which they should be specified. It is probable that relatively inaccurate measurements of several “unreliable” features are adequate to distinguish reliably a given speechsound from others of the same language, provided these features are selected properly. These remarks are illustrated by existing data on the acoustical properties of vowel and consonant sounds, and areas are noted where further data are needed. The possibilities of utilizing the high‐speed digital computers in the analysis of speech are discussed.
- Contributed Papers
30(1958); http://dx.doi.org/10.1121/1.1929934View Description Hide Description
In order to take best advantage of the speed and flexibility of the digital computer as a research tool in speech analysis and recognition studies a real‐time input system is needed. Such a system, which makes power‐spectrum data directly available to the computer, has been constructed at the Lincoln Laboratory for use with the TX‐0 and TX‐2 computers. A speech signal is fed to 35 band‐pass filters covering the range 100 to 10000 cps. The output of each filter is envelope‐detected, and the envelopes are sampled periodically by a 128‐position mercury commutator switch. The output of the commutator is converted to binary digital form and fed directly to the computer, At present the system is capable of making 7620 samples per second with a signal‐to‐noise ratio of over 40 db (7 bits). The filters, made up of high‐pass and low‐pass sections in cascade, can be connected to give a variety of band widths and center frequencies Patch‐cord connections between envelope‐detector outputs and commutator inputs permit some frequency channels to be sampled at higher rates than others when desired. Utilizing the maximum sampling rate of the input system and the full storage capacity of the computer, up to 20 sec of continuous speech may be read into the computer and held for analysis. [The research reported in this talk was supported jointly by the U. S. Army, Navy, and Air Force under contract with the Massachusetts Institute of Technology.]
30(1958); http://dx.doi.org/10.1121/1.1929935View Description Hide Description
No necessary‐and‐sufficient stimulus, or stimuli, has been discovered for the identification of any given word, yet people can reliably recognize speech under a wide variety of conditions One implication of this fact might be that speech recognition involves a complex data processing which takes advantage of linguistic redundancy. Such a system could make possible a reliable identification on the basis of a number of discriminating conditions, no one of which was in itself dependable. An attempt was made to realize such a data processingsystem in a program written for an experimental computer at Lincoln Laboratory. The program input consisted of the spoken digits processed through the Haskins Laboratories' resonance vocoder. Ninety‐eight percent recognition was achieved on a sample of ten voices, mixed male and female. [The research reported in this talk was supported jointly by the U. S. Army, Navy, and Air Force under contract with the Massachusetts Institute of Technology.]
30(1958); http://dx.doi.org/10.1121/1.1929936View Description Hide Description
The present American standards for normal hearing for audiometric purposes imply that the difference between audiometric reference zero for a 1000‐cps pure tone and audiometric reference zero for spondee words is 6 db (i.e., approximately 16 db SPL for a 1000‐cps tone and 22 db SPL for spondee words). An accumulation of evidence from several sources suggests, however, that the difference between the two thresholds is considerably larger than 6 db. The present study was designed to assess the effects of sex, ear, order of test presentation, method of threshold measurement, and degree of prior familiarity with spondee vocabulary on the difference between the threshold for a 1000‐cps tone and the threshold for spondee words, as measured with W‐1 recordings. Subjects were 126 young adult college students. Results indicate that the difference varies from 12 to 14 db, depending slightly on degree of prior familiarity with the spondee words, and on the method used to measure the pure tone threshold. The average speech threshold SPL in the present group agreed closely with the current standard reference level of 22 db, while the average 1000‐cps pure tone threshold SPL was 6 to 8 db lower than the current standard reference level of 16 db. [This research was supported by the U. S. Air Force under Contract No. AF 41 (657)‐185, monitored by the School of Aviation Medicine, USAF, Randolph Air Base, Texas.]
30(1958); http://dx.doi.org/10.1121/1.1929937View Description Hide Description
The perceptual confusions among English compounds of two consonants were examined. One defined class of syllables, made up of 15 initial compound consonants in conjunction with three vowel sounds (i), (a), and (o), and another class of 15 final compound consonants in conjunction with the same three vowel sounds, were employed. Recorded syllables were played back against a white noise background and against a low‐frequency noise background. Confusion patterns among the compound consonants depended upon the articulatory dimensions of the individual consonant members which formed the compound That is to say, there was little evidence of interaction among the members of the compound. The confusion patterns indicated that the low speech frequencies, i.e., those frequencies heard above the white noise, convey the consonant distinctions of nasality, stop vs liquid, and place of articulation of the liquids. On the other hand, the higher speech frequencies, heard above the low‐frequency noise, conveyed the distinction of affrication. The different vowels had minor effects on the perception of the consonants adjacent to them.
30(1958); http://dx.doi.org/10.1121/1.1929938View Description Hide Description
The employment of directional information of the desired talker in a multitalker environment has been termed the “cocktail party effect.” We have attempted to determine whether the addition of (1) “comb” filtering, which enhances the perception of spatiality [M. R. Schroeder, J. Acoust. Soc. Am. 29, 774 (1957)], or (2) splitting the speech frequency range, which enhances multichannel listening [W Spieth and J. Webster, J. Acoust. Soc. Am. 27, 866 (1953)] will improve the cocktail party effect, as previously defined [I. Pollack and J. M. Pickett, J. Acoust. Soc. Am. 29, 1262 (1957)]. The introduction of frequency separation does not remove the cocktail party effect. Intelligibility decreases as the frequency subdivision becomes more gross. However, it may be noted that a communication system based upon crude frequency separation and binaural listening may still be effective for certain instances of low‐information communications. Specifically, PB monosyllabic word intelligibility of over 30% may be achieved with 4 background talkers in each of two split‐frequency channels. This is equivalent to employing 11 transmitters, shared in pairs among 55 different talkers.
30(1958); http://dx.doi.org/10.1121/1.1929939View Description Hide Description
A small megaphone was tested for use in direct voice communication at high voice levels against backgrounds of noise. The voice transmission characteristics of the megaphone were measured. Intelligibility scores were determined for ten talkers, speaking with and without the megaphone, over a range of field voice levels of 70 to 100 db SPL (at 1 meter). A white noise and a low‐frequency noise background were employed. Nearly the entire over‐all acoustical gain produced by the megaphone was realized in improved intelligibility above both noise spectra. The acoustical gain of the megaphone permitted a lower voice level, and, hence, delayed the deterioration in intelligibility associated with distortions due to shouting. Frequency shifts in the voice spectrum were measured over a range of vocal efforts from conversational maximum shout. Control tests demonstrate that the voice‐frequency shifts associated with shouting are not the primary factor in the deterioration of intelligibility; the frequency shifts appear to be only symptomatic of serious vocal distortions in shouting.
30(1958); http://dx.doi.org/10.1121/1.1929940View Description Hide Description
Acoustical and nonacoustical factors have been observed which suggest that words in one position in a sentence may be more intelligible than words in other positions. An experiment was undertaken to determine the effect of some of these factors upon word intelligibility in sentences heard in noise. One‐hundred six‐word sentences, drawn from magazines, were read to 10 listeners The intelligibility of words in the first, fourth, and sixth positions was determined and their acoustic characteristics were measured.Measurements of intensity, duration, and total energy were obtained for each word. The words in final position were lowest in intelligibility despite the fact that they were highest in average intensity, duration, and energy. This apparent contradiction was resolved by a grammatical analysis of the words. The words in the final position included a larger proportion of content words (Nouns, verbs, adjectives, adverbs) than words in the other positions. The content words are less predictable, less intelligible, and contain more information per word than pronouns and function words (articles, prepositions, conjunctions). The intelligibility observed at each position was approximately equal to the intelligibility calculated from (1) the intelligibilities of content and function words, irrespective of position, and (2) the relative proportions of content and function words.
- Session C. Propagation
30(1958); http://dx.doi.org/10.1121/1.1929941View Description Hide Description
An experimental attempt is made to determine the presence of scattering resulting from the nonlinear interaction of two finite amplitude sources, operating in water. The two beams cross each other at right angles. A detector crystal, tuned to the summation frequency of the two sources (13.4 Mc), and pivoted about a point above the interaction region, is used to investigate the scattered field. This crystal feeds into an RAO‐7 receiver, equipped with a calibrated s meter. The results of these investigations indicate no scatteredsound above the noise level of the detection system. Therefore, any scattering present would be several orders of magnitude below that predicted by the theory of Ingard and Pridmore‐Brown. The lack of scatteredsound is in agreement with the theoretical considerations recently presented by Westervelt. (This work was supported in part by the Office of Naval Research.)
30(1958); http://dx.doi.org/10.1121/1.1929942View Description Hide Description
Results of field measurements of scatterattenuation of sound in a turbulent atmosphere are reported. The observed attenuations are much smaller than might be expected from a scattering theory using the total scattering power of the turbulence in the calculation of the attenuation (Lighthill, 1952). That theory also predicts a strong dependence of the scatterattenuation on both mean wind speed and frequency. Again, this prediction is not supported by the experimental evidence. Attenuations calculated on the basis of backscattering alone are presented. They are in reasonable agreement with experiment. The backscattering is caused by eddies comparable in size to the wave length of the scattered sound. Consequently, the energy spectrum of the turbulence will affect the frequency dependence of the attenuation considerably. An extensive series of turbulence spectra has been obtained experimentally near the ground for different micrometeorological conditions. A discussion of these spectra and their dependence on mean wind speed and thermal gradients is presented. On the basis of the foregoing considerations, preliminary estimates of the attenuation of sound in the audio‐frequency range by atmospheric turbulence are given.
30(1958); http://dx.doi.org/10.1121/1.1929943View Description Hide Description
Undesired sound is frequently attenuated by having it pass through ducts with absorbing side walls. If the absorptive properties of the walls are described by a normal impedance, and if there is no air flow in the duct, the resulting attenuation is easily predicted on the basis of existing theory. In many cases, however (e.g., in ventilating ducts or the exhaust ducts in certain wind tunnels or jet‐engine test cells), there is also a flow of air through the duct which may affect the rate of sound attenuation. In order to investigate this phenomenon a study is made of the propagation of sound in both constant gradient shear flow and a turbulent shear flow above a flat surface. Curves are presented showing how, in the case of downstream propagation, the flow gradient tends to channel the sound energy into a narrow layer next to the wall. These results are used in estimating the effect of a flow on the attenuation of sound in a duct with absorbing side walls.
30(1958); http://dx.doi.org/10.1121/1.1929944View Description Hide Description
Studies have indicated that dc airflow through an orifice creates acoustic resistance. The creation of this resistance has been investigated by the measurement of the impedance of a perforated plate in an impedance tube. A method of calculating the acoustic resistance due to dc airflow has been prepared. The close agreement of theory and measurements points the way toward the use of this phenomenon for the improvement of all types of silencers where airflow is present.
30(1958); http://dx.doi.org/10.1121/1.1929945View Description Hide Description
A theory is developed for the propagation of plane, longitudinal waves through a material which is both flexible and porous. An idealized model is constructed consisting of an array of parallel elastic rods uniformly spaced in air. Viscous and inertial coupling forces, as well as pressure squeezing accompanying sidewise expansion of the rods, are responsible for interaction between the air and solid of the model. Two dilatational waves are found in the unbounded medium. The propagation constants describing these waves are written in terms of 6 dimensionless parameters characteristic of the model. Boundary conditions are introduced and an expression is obtained for the acoustic impedance at the front surface of a rigidly backed model of finite thickness. It is shown that viscous coupling forces are of predominant importance in controlling the amount of acoustic energy converted into heat at the internal surfaces of the material.Viscous coupling also determines the amount of energy transferred from the airborne sound waves to the interior of the solid where dissipation may take place if internal losses are appreciable. Preliminary data relating theory to experiment are presented.