Relevance of time‐varying properties of the first formant frequency in vowel representation

Maria‐Gabriella Di Benedetto
1985 Journal of the Acoustical Society of America  
Invited Papers 8:05 AI. The effects of se• surface roughness on the transmission of sound through an air-ocean interface. J. This is a review of recent experimental and theoretical studies of the influences of the sea surface on the transmission of airborne sound into the ocean. Most applications will depend upon conditions where the refracted grazing angle is greater than approximately 30' since the loss is very large for smaller angles. For grazing angles larger than 30', two parameters serve
more » ... fairly well to characterize the transmission loss through the sea surface. The first is a roughness parameter R which is equal to the sea wave height measured in airborne sound-speed wavelengths. The second is a diffraction parameter DL which is equal to the Fresnel zone diameter measured in correlation lengths of the sea surface slopes. When R is less than unity and DL is less than 3, there is usually no effect due to surface roughness. When either condition is not satisfied, there is usually more transmission loss than there would he through a flat surface. Exceptions to this depend upon the refracted grazing angle and upon the angle between the wave crests and the incident sound. 8'.3O A2. Review of source spectrum level density of wind-generated noise. The source spectrum level density of wind-generated noise has been deduced from ocean measurements, laboratory experiments, and theoretical studies in the past 25 years. Selected contributions are reviewed and the results are compared. Since the source level results compare favorably, at least in deep water, an average source spectrum level density as a function of frequency and surface wind speed is proposed for use in noise models. The differences between the dipole and "rocking" dipole source directivity patterns is discussed and wind-generated noise modeling issues are addressed. There seems to he little agreement on the underlying physical mechanisms of wind-generated noise. The question of what information would be added to the source spectrum level densities if a theory for the underlying physical mechanisms were known is also addressed. 8:55 A3. Low-frequency ocean s rface noise sources. W.M. Carey •Naval Ocean R&D Activity, NSTL, MS 39529) and M.P. Bradley (Planning Systems Inc., Slidell, LA 70458} The interaction of the wind with the ocean surface has long been recognized as the major source of ocean ambient noise. High-frequency noise data 1200 to 2 kHz) has consistently been found to have strong winddependent characteristics associated with spray, splashes, bubbles, and rain. Recently, wave-wave interaction has been shown to be a source of infrasonic (0.2 to 2 Hzl noise and ocean bottom microseisms. Generally, the low-frequency noise (2 to 20 Hz) is associated with noise from distant ships. However, narrow-hand {as opposed to !/3-octave) measurements show in addition to the noise from ships a wind-dependent characteristic. Furthermore, mid-ocean basin vertical directionality measurements show noise intensity near the horizontal with a broad frequency characteristic in diverse geographic locations. These results suggest a wind-generated noise due to a mechanism such as wave-wave interaction, wind turbulence, or the interaction of surface waves with turbulence is coupled into the mid-basin sound channel by either a shallowing sound channel such as found at high latitudes or a down-slope conversion process due to the basin boundaries and sea mounts. Theoretical Sl J. Acoust. Soc. Am. Suppl. 1, VoL 78, Fall 1985 110th Meeting: Acoustical Society of America Sl expressions are derived from first principles following the approaches of Yen and Pertone [Naval Underwater Systems Center TR5833 { 1979}] and Huon Li [Naval Ocean R&D Activity, TN89 ( 1981)] yielding the frequency-dependent radiation characteristics for wave-wave interaction, wind turbulence, and wave-turbulence interaction. These results show that wave-turbulence interaction is a possible source of wind-driven noise in the 10-to 200-Hz regions. Other possible mechanisms such as nonlinear capillary wave interactions are discussed and compared to this wave-turbulence mechanism. A4. Verticaldirectionalityafambientnoise--Areview. W. vonWinkleandD. Browning(NavalUnderwater Systems Center, New London, CT 06320) The vertical directionality of ambient noise in the sea falls into three frequency regimes: 0-200 Hz, dominated by long range sources; 200-10 003 Hz, relatively small directionality; 10 kHz and above, downward rays and heavily attenuated. A review is made of data mechanisms and our present prediction capability. Contributed Papers 9:.45 AS. The air/sea interface as a wave source. The surface waves on the ocean are coupled to the underwater sound field, the coupling arising either because the one generated the other, or because they were both created cohcrently by a common source. This paper discusses the various mechanisms of sound generation at and near the ocean surface and identifies parameters controlling the regimes where the various sources are dominant. Reference will be made to the results of some exact model problems involving aerial turbulence-induced surface waves and their associated sound field in an infinitely deep ocean. These model problems utilize the techniques of aero-acoustics on the ocean sound problem and indicate that some aspects have not previously been adequately modeled when the surface is vigorously disturbed in the main noise-producing conditions. 10:00 A6. On the wind-wave interaction mechanism. R. H. Meilen {PSI Marine Sciences, New London, CT 16320) and D. Middleton 027 E. 91 Street, New York, NY 10128} Sonar and radar backscattering measurements indicate that the winddriven sea surface does not behave aeccording to linear theory at high wavenumbers. Backscatter strengths at small grazing angles are much greater than expected and the scatterers appear to move without the usual dispersion associated with "capillary" waves. Theory shows that nonlinear effects of snrface-drift and wind-wave interactions can account for the shocklike properties. Nonlinear wind-wave effects have been observed in wave-flume measurements by spectral analysis of wave-gauge data. Initially periodic ripples develop coherent harmonics with increasing fetch; however, subharmonic growth leads to rapid degeneration and a continuous spectrum. The mechanism appears to involve intermodulation between wind and wave leading to chaotic behavior and redirecting the energy cascade to lower wavenumbers. In the equilibrium stage, further deveiopment of longer gravity waves ceases as they outrun the source, incoming energy being balanced by high-wavenumber dissipation. Unstable disturbances generated on the surface should degenerate into ensembles of solitons, explaining the Doppler spreads and backscatter cross sections observed at high frequencies. 10:15 A7. Infrasonie ambient ocean noise: Northeast Paeifi½ Ocean. Rudolph H. Nichols (Physics Department, Naval Postgraduate School, Monterey, Measurements ofocean ambient noise were made at three widely separated deep-water bottom locations in the N. E. Pacific, at eight frequen-ties in the range from 2.5-20.0 Hz, for 40 consecutive days. Concurrent data on wind speed and wave height were collected. Analysis indicates that the spectrum level of infrasonic noise is linearly related to the log of the wind speed above a threshold level. There is evidence that the noise can be directly associated with the wind, rather than through the surface waves it produces. [Work supported by ONR.] 10'•0 AS. Underwater noise caused by precipitation. The characteristics of underwater noise in the ocean generated by precipitation are important to weather forecasters and occanographers since they permit the detection and measurement of rain over the ocean by remote {i.e., buoyed or bottom-mounted) acoustic sensors. We have recently observed the character of the underwater noise generated by rain, hail, and snow. The spectrum of rain noise, for wind speeds below 1.5 shows a peak of 13.5 kHz with a sharp cutoff on the low-frequency side and a gradual falloff{7 dB per octave) on the high-frequency side. Stronger winds smear the peak. Hail spectra show a peak at 3.0 kHz with a gradual {roughly 11 dB per octave) falloff on both sides. The spectrum of snow noise is unique. Our instrumentation permitted the measurement of the drop Ior stone} size distributions in the precipitation. These findings will enhance the art of remote acoustic sensing in oceanograpy. [Work supported by: Supply Several mechanisms by which bubbles can contribute to ambient noise in the ocean are described and their effectiveness estimated. At low frequency, up to a few tens of Hz, bubbles are driven into oscillation by oceanic turbulence. The normal quadrupole radiation mechanism of turbulence thus acquires a monopole character. At frequencies from around I to a few kHz single bubbles formed by breaking waves radiate in free oscillation. In the range of hundreds of Hz the acoustic emission may be due to collective oscillations of systems composed of many bubbles. The efficiency of all these mechanisms is estimated on the basis of an adaptation of Lighthill's theory of aerodynamic noise. Finally, at frequencies above several kHz, drop impact and free oscillations of bubbles thereby produced appear to be responsible for the ambient noise. [Work supported by SACLANT ASWR, Low-frequency ambient noise directionality measurements were made by a towed line array in the Gulf of Mexico at two different sites. The first site was seaward of the continental shelf and north of the Yucatan Peninsula. The second site was near the entrance to the Guff of Mexico. The presence of a shallow shelf at the first site effectively shields the array from long-range acoustic propagation from the south. This shielding, combined with favorable propagation conditions and high-density shipping to the north, is responsible for the high degree of anisotropy in the measured ambient noise horizontal directionality pattern at the first site. The measured ambient noise directionality pattern at the second site shows the strong influence of the shipping entering and exiting the Gulf of Mexico. Because the acoustic propagation conditions and the shipping in the Gulf of Mexico are approximately repeatable annually, these two directionality measurements are considered to be good estimates for the ambient noise for future years. A discussion of the horizontal directionality as well as additional ambient noise statistics for the two sites will be given. 11:15 All. Computer simulation of the vertical structure of mld-oceen ambient noiseThe acoustic field for a realistic distribution of noise sources was calculated at a deep-water location in the Northeast Pacific corresponding to the CONTRAK VI measurement site [M. Z. Laurence and D. J. Ramsdale, J. Acoust. Soc. Am. Suppl. 1 77, S?0 (1985)]. The calculations were made with the NORDA Parabolic Equation programs along four horizontal directions representing distinctly different environments and historical shipping distributions. The calculated field at the CONTRACK VI site was sampled by vertical line arrays to determine the vertical arrival angles of the noise. This vertical structure was studied as a function of array depth, array length, and frequency for each horizontal direction and compared with the CONTRACK Vl measurements. The shape and depths of the resulting vertical noise notch are heavily affected by arrivals from sources over continental slopes and in cold water. [Work supported by DARPA.] 11:30 AI2. Calculations of ocean ambient noise over a continental slope. D. R. Del Balzo, M.J. Authement, and C. T. Mire (Naval Ocean Research and Development Activity, Ocean Acoustics Division, Code 240, NSTL, MS 39529) Computer calculations of slope effects on surface noise at I00 Hz were made using the NORDA high-angle, implicit finite difference acoustic field model. A geoacoustic description of a continental margin, which is typical of the east coast of the U.S., was chosen for analysis (vertical slope of 2.8 deg and covered by 500 m of silt}. In order to enhance the effect of bottom interaction, a Summer Sargasso Sea soand-speed profile was used. The calculations indicate the presence of a deep noise notch over the slope (for certain basin and shelf noise generating locations) which is deepest near the ocean surface (25 dB} and which decays with depth. The potential for signal-to-noise ratio gain is suggested. [Work supported by NAV-AIR.] 11:45 A13. Arctic Ocean [main theory. Low-frequency ambient noise sources observed in the Arctic are discretely located in space and time. With a large number of events per unit area per unit time, the total number of contributions to any one observation grows with horizontal range but also shrinks in intensity inversely with range. This standoff therefore suggests that limits on observed ambient noise are set by basin bathymetry and/or by sound absorption. In one useful model we find the source spectral density typical of low-frequency events is changed via the basin by anf • shaping at frequencies below, and by anf-• shaping at frequencies above a characteristic frequency. [Work supported by ONR.] TUESDAY MORNING, 5 NOVEMBER 1985 DAVIDSON ROOM A, 8:15 TO 11:50 A.M. Invited Papers 8:15 BI. A tribute to James H. Botsgord (1925-1984) and a histor'cal Feview of impulse noise measurement in the steel industry. A tribute to Jim Botsford who passed away on 18 May 1984 will be presented. A historical review of noise measurement in the steel industry, including impulse noise, highlighting Jim's endeavors will be given. The instrumentation and methods used for noise measurements along with rating information will be discussed. Problems associated with impulse noise measurement will be reviewed and questions regarding proper instrumentation, analysis, and rating for impulse noise will be raised. $3 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America S3 8.'• B2. Clas•itieation of impulse noise: Comparison of crcst factor and kurtosi• as metrics. In this session honoring Jim BoLLford, I will address one aspect of a problem with which he was concerned, whether impulsive noise should be (or could be) considered separately from continuous noise exposures. Recently, ! presented a method for classifying noise as impulsive based on a statistical distribution measure of "peakedhesS," the kurtosis (b2). This is the first step in addressing the "separability" of noise types. It is likely that BoLLford and others of a practical demeanor would ask what advantage, if any, this has over familiar measures such as crest factor (cf). By defining a ratio, a, such that a = b•/cf •, we show that the measures are related such that for a given impulse form, a is constant for any duration signal. Only the form of the impulse changes the value of a. Thus, for a rectangular impulse, a = I and for an exponentially decaying sinusold, a = 3/2. The crest factor may, therefore, have some utility for classifying impulsive environments. This utility is diminished, however, for cases at the border between impulsive and continuous noise. For signals with 3 < b 2 < 6, the crest factor is unreliable for classifying the environment because of its dependence on a single peak. Examples of the relative efficiency of the metrics for classification will be drawn from actual samples of industrial noise. 9:05 B3. Noise doslmeter transient response eharacterlstics. John J. Earshen (Metrosorties, Inc., Noise dosimeters used to monitor worker exposures to noise must have prescribed transient response characteristics to satisfy the requirements of the U.S. Department of Labor, Occupational Safety and Health Administration. In addition, computation of noise dose or average sound level must be based on a 5 dB per time doubled trading ratio. Various individuals have concluded that there is an apparent discrepancy between measured and computed doses when industrial noise contains impulsive components. The discrepancy results when computation neglects the effects of the prescribed transient response characteristics. In p•rticular, "slow" response has a critical effee• when trading ratios of 4 or 5 dB are used. A dosimeter is required to produce measurement results which correspond to results that can be computed from SLiM readings. This paper shows the relationship between transient response and measured dose resulting from noise containing impulses. Furthermore, experimental results are presented which show that dosimeter and SLM derived results are consistent and correct. Reported results purporting to show that dosimeters give incorrect measuremeats of impulsive noise are shown to be incorrectly evaluated because theoretically computed reference values fail to account for U.S.D.O.L. OSHA prescribed transient response. The effect of Substituting "fast" response is also presented. 9:30 ]M. Assess'ng the hn,•rds of impulse noise superimposed on background noise. D. Henderson and R. P. Hamernik (Cailler Center of the University of Texas at Dallas, 1966 lawnod, Dallas, TX 75235) Impulsive noise, either in the military or industrial setting, rarely occurs without some substantial t•ckground noise. This paper reviews a series of animal experiments which show synergistic interactions between relatively "safe" continuous and impulsive noise. In all the experiments, chinchillas were either exposed to (1) 60 rain of background noise (either 2-4 kHz or 0.5-1 kHz) at levels ranging from 95 to 89 dB; (2) 50 spark discharge-generated impulses with peak sound pressure levels ranging from 158 to 137 dB and A durations of 32 to 64/zs; 13) various combinations of(l) and 12). The animals' hearing was measured immediately after the exposure, and at regular intervals over a 30-day period. After 30 days, the cochleas were analyzed for losses of sensory cells. The results show that certain combinations of individually "safe" impulse and continuous noise can (1) produce a synergistic interaction; (2) that these interactions behave systematically as either the level of the impulse or the continuous noise is reduced; and (3) that the interaction effect is enhanced as the frequency spectra of the two noises begin to overlap. The results of the experiments will be discussed in terms of the mechanisms of noise-induced hearing, implications for measurement of noise environments loss, and the implications for public health standards. [Work supported by NIOSH.] 9:55 HS. Hearing protector attenuation as intlueneed by frequency and time variation of the no'se. Daniel L. Johnson and Myron D. Smith (Larson-Davis Laboratories, 280 S. Main, Pleasant Grove, LIT 84062} Over a decade ago, Jim BoLLford proposed two innovative concepts that provide the basis of this study. First, he proposed correcting the attenuation of hearing protectors by a "CA" adjustment [J. H. BoLLford, Sound Vib., 32-33 {Nov. 1973)]. Using three separate time history dosimeters, sc• to 5-s periods, the C-weighted and A-weighted levels at the shoulder were measured; also, the level under the muff was recorded. The results show that most of the time the C-weighting outside the muftis the best predictor of the A-weighted level under the muff when used with a single value of noise reduction. Another concept of Botsford was the TTS meter, in which the intermittency pattern of a noise could be assessed [J. H. BoLLford, Am. Ind. Hyg. Assoc. J. 32, 92-95 (1971}]. Harris and the first author further enhanced this concept by providing a TTS-based intermittency correction factor to adjust a noise dose [D. L. Johnson and C. S. Harris, Proc. Noise Expo, Chicago, IL, 16-122 (! 979)]. Using this adjusted dose, our current data indicate better performance by muff-type hearing protectors than simply calculating inside and outside average noise levels might predict. We also noted that a person's S4 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America $4 voice can disproportionately influence the attenuation predicted by an under the muff measurement. We believe that Botsford's early concepts are still valid and technology now exists which can make his ideas useful. Contributed Papers lO:2O B6. Development of a new artificial head for impulse noise attenuation measurements of hearing protective devices. While psychoacoustical methods to evaluate the sound attenuation of heating protective devices have been standardized {CSA Z94.2-MI984; ASA STD 1-1975} and proved to be practical for continuous noise, no standard exists for impulse noise. Microphone-based techniques seem to provide the best avenue for the measurement of impulse noise attenuation at the present time. This paper presents our progress in the design of a new artificial head simulating the relevant mechanical and acoustical parameters of the human head influencing the protected and unprotected eardrum sound pressure. These parameters include: {1} the diffraction of sound by the human head, (2) the mechanical impedance of human tissues, {3) the circumaural contours, {4) the external ear features such as pinna, ear canal, and eardrum impedance, and {5} the vibration of the head in the sound field. {Work supported by the Defence and Civil Institute of Environmental Medecine, Canada.] 10'•5 B7. The dependence of critical level upon duration. W. Dixon Ward and Dorisann Henderson {Hearing Research Laboratory, University of Minnesota, 2630 University Ave. SE, Minneapolis, MN 55414) In the chinchilla, the critical level for single exposures to 700-to 2800-Hz noise of 20 min or longer duration has been found to be between 112 and 120 dB SPL. For example, a 22-min exposure at 120 dB produced considerably more permanent threshold shift, PTS {50 dB at 2000 and histological damage [80% destruction of the outer hair cells (OHC}] than a 220-rain exposure at 112 dB or any single uninterrupted exposure longer than 220 min but equivalent in energy to the 220-rain 112-dB exposure (20 dB PTS, 10% OHC destruction}. However, in the present experiment, daily exposure for 9 weeks, Monday through Friday, to 40 7.2-s bursts of 120-dB noise spread over an 8-h "workday" (hence with 12 rain between bursts) produced negligible PUS and OHC destruction. Clearly, therefore, the critical level depends on duration and temporal pattern, and so regulatory systems governing exposure of workers to industrial noise that put a fixed ceiling on exposure levels regardless of temporal pattern-whether this ceiling be 115 dBA, 130 dBA, or 140 dB peak--will often, if not usually, be incorrect. [Research supported by NIH Grant 12125.] 10:50 BS. The effect of impulse intensity and the number of impulses on hearing and eneMear pathology in the chinchilla. James H. Patterson, Jr., Ilia M. Lomba-(•autier, Dennis L. Curd University of Texas at Dallas, 1966 Inwood Road, Dallas, TX 75235) Forty-one chinchillas, divided into seven groups, were exposed to 1, 10, or 100 noise impulses having peak intensities of 131, 135, 139, or 147 dB. Hearing thresholds were measured in each animal prior to exposure using an avoidance conditioning procedure. Threshold shifts were moni-toted at regular intervals over a 30-day post-exposure period. A surface preparation of the cochlear sensory epithelia was performed approximately 90 days after exposure. There was generally an orderly relation between the amount of permanent threshold shift and the severity of exposure, and a general agreement between averaged histologlcal data and the audiometric data. Detailed relations between temporary and permanent threshold shift, cochlear pathology, and exposure variables are discussed, as are the implications of these data to the development of exposure criteria. 11:0S B9. Impulse noise repetition rate: Implications for the equal energy hypothesis. The equal energy hypothesis {EEH} for impulse noise is based on the assumption that the hearing hazard associated with an impulse noise expusure is a monotonic function of the total A-weighted energy received by a listener. A series of experiments will be described in which the EEH was evaluated using exposure paradigms in which the repetition rate {4/s to 1 every 16 s} and the intensity { 107-to 131 <lB peak SPL} of the impact noise were varied. Each of the various exposure paradigms had the same total energy. If the magnitude of the hearing loss is determined by the amount of acoustic energy, then all groups should develop the same amount of hearing loss. For equal intensities, the degree of damage increases with increased repetition rate. For example, the level ofTIS increased approximately 25 dB when the repetition rate was increased from I every 16 s to 4/s. The effect of increasing the repetition rate also resulted in greater PTS and cochlear damage. The results are discussed in terms of the EEH and its appropriateness to impulse/impact noise. [This work was supported by NIOSH-R010H 1152.] 11:20 BIO. Temporary functional changes in men due to exposure combinations of noise and low-frequency vertical vibrations. Olavi J. Manninen {The The main purpase of this study was to produce additional information on the connections between the temporary hearing thresholds, body upright posture sway, and cardiovascular changes induced by complex exposure conditions. Seven healthy male students were exposed consecutively five times to mere noise, to mere whole body vibration, and simultancousiy to noise and vibration at 20 'C. The noise categories were: ( 11 no noise and {2) noise of 90 dBA. The categories of low-frequency whole body vibration (Z axis} were: ( 1} no vibration, (2) vibration within the range 4.4-5.6 Hz, (3) vibration within the range 2.8-5.6 Hz, {4) vibration within the range 2.8-11.2 Hz, {5} vibration within the range 1.4-11.2 Hz, and {6} sinusoidal vibration with a frequency of 5 Hz. The {rms} acceleration in all the vibration models was 2.12 m/s 2. The TYS 2 values at 4 and 6 kHz increased as a result of simultaneous exposure to noise and vibration signdicantiy more than as a result of exposure to noise alone. The means of the sway variances in the X and Y directions at 0.1 Hz and within the range 0.06-2.00 Hz increased only when the vibration in the noise-vibration combination was a sinusoidal one. The changes in the heart rate, R-wave amplitude, blood pressure, and the hacmodynamic index scores also depended on the bandwidth of the vibration, the number of consecutive exposures, and on whether the subjects were simultaneously exposed to noise in addition to vibration. [Work supported by The Academy of Finland.] S$ J. Acoust. Soc. Am. Suppl. 1, VoL 78, Fall 1985 110th Meeting: Acoustical Society of Amedca S5 11:35 BII. A •tion of room no'se levels et telephone locations and the effect upon employees' work and communication. George R. Auten, Jr. (United States Air Force, HQ SAC/DERI, Offutt AFB, NE 68113-5000), Noise surveys were conducted at telephone locations in four types of work environments: State government office buildings, fast-food restaurants, automobile dealers/maintenance shops, and shopping mall retail stores. Two types of surveys were taken. A type I survey was conducted during normal working conditions and a type 1I survey was conducted at selected sites in order to better define the dominant noise sources. Acoustic measurements made included A-weighted, C-weighted, F-weighted (flat], and octave-band sound pressure levels. General room descriptors were also noted for each survey location. A total of 270 telephone locations were surveyed at 51 sites. The employees were asked to assess their work place with respect to noise and its effect on their work, face-to-face, and telephone communication activities. Summary statistics for each type of sound survey and the responses from the questionnaires given out duro ing the surveys will be presented. TUESDAY MORNING, 5 NOVEMBER 1985 REGENCY BALLROOM IV, 8:30 TO 11:45 A.M. We discuss the design of a text-to-speech synthesizer, which accepts any type of English text as input, and creates an appropriate speech signal as output. Effective algorithms for converting text to sound must make use ofinterrnediate data structures that systematically encode the degrees of freedom available to speakers of the language being synthesized. These data structures are an engineering approximation to what linguists call phonological representations; we will call them "P-structures." Any TTS system must: (1) define its version of P-structures; (2) design and implement algorithms for transforming input text into P-structure•" (3) design and implement algorithms for transforming P-structures into sound. Task • 1 is mainly a problem in applied linguistics; task #2 can best be seen as applied AI; task #3 is an application of phonetics and signal processing expertise. Because of the interdependence of approximate solutions in different portions of the system, integration of the various parts is a nontrivial problem. We will analyze an example of a TTS system, showing how these problems were handled in building and combining its numerous pieces. Finally, we point out some areas where better solutions are needed, and suggest how to find such solutions. 8:45 C2. A hybrid domain articulatory speech synthesizer. Articulatory speech synthesizers model the human vocal tract by means of its geometrical and functional properties. It is believed that this approach can be advantageous in speech coding at bit rates below 4800 b/s. Existing articulatory synthesizers work in the time domain. They either solve a system of differential equations for the vocal tract and the glottis, or synthesize speech using wave digital filters. The first approach is computationally very cumbersome. Both approaches have difficulties in incorporating important acoustic parameters, for example, the radiation impedance at the lips, wall vibration, and other losses. So far a realistic glottis model suitable for the wave digital filter approach does not exist. We have combined a nonlinear time-domain model of the vocal cords [K. lshizaka and J. L. F!anagan, Bell Syst. Tech. J. 51, 1233-1268{1972)] with a linear frequency-domain chain-matrix model of the vocal tract [e.g., M. M. Sondhi, paper •4.5.1, Proc. lnt. Congress on Acoustics, Paris, France, 1983, Vol. 4, pp. 167-170]. The interface between these two models consists ofconvolving the glottal flow in the time domain with impulse responses of the tract obtained by inverse FFT. Examples of synthesized speech using manually generated and measured tract areas will be given. 9.-OO C3. Townrds developing a imrametrie model Acoustic parameters, such as LPC parameters, vary continuously as functions of time. The underlying speech (phonetic) events, however, are discrete. How do we go from continuous acoustic representation to discrete phonetic representation? We will discuss in this paper a parametric model for representing variations of LPC parameters with time. We define a matrix Y to express the LPC parameter values at successive time instants with different columns representing time. We then write matrix Y as a product of three matrices A, Z, and P: a central speech information matrix P representing underlying speech events, a left matrix A relating the speech events to the LPC parameters, and a right matrix P representing the spreading of speech events along the time axis. The number of columns in the matrix Zis equal to the number of speech events in a given speech segment and is much smaller than the number of time frames in that segment. We assume that there is overlap between the spreading functions of only neighboring speech events, that is, the matrix ppr is a tridlagonal matrix. We will discuss procedures to determine the matrices A, Z, and Psuch that the mcan-squared error between the elements of the matrix Y and the product JZP is minimized. We will also discuss conditions under which the selection of the matrices is unique. 9:15 CA. Aligning phoneroes with the corresponding orthography in a word. As part of a speech synthesis project a technique has been developexi for aligning the phonemes in a phonetalc transcription with the graphemes in a word. For example creationism/kXle[j'n. •zam/can be aligned as e.r.e.n.t.ion.i.s.m with /k.r.x.eLj'•m/. The technique has been used to determine in a machine-readable dictionary, where an entry is S6 J. Acoust. Soc. Am. Suppl. 1, VoL 78, Fall 1985 110th Meeting: Acoustical Society of America S6 given with multiple orthographies and one pronunciation, whether the pronunciation may be used for all the orthographies. For example, in the Collins English Dictionary there appears the entry abutment [•bAtm•nt/or abuttal'. We know these two words are synonymous but are not pronounced the same. aligns with/•.b.A.t/but the al cannot be aligned with/m•nt/. The pronuciation of abutment cannot be used as the pronunciation ofabuttal. The aligned forms of words and their pronunciations, together with statistics on frequency of rule use, is seen as an invaluable aid for developing a set of text-to-phoneme rules, where the ordering of rules, based on frequency of occurrence, is critical. It may also be useful as part of a pattern matching process in a speech recognition system. This paper will describe the algorithm and will give some results of using it against a phonemically tagged version of the Lancaster-Oslo/ Bergen Corpus. CS. A dictiunary-intensive letter-to-sound program. We treat the letter-to-sound problem as that of deriving all regularly inflected forms of English words from a dictionary ofbase forms. A computer program must perform all lexical operations a human might have to perform in using an abridged desk dictionary. Of course this includes addition of"s." "es," "d," "ed," "er," and "ing," with their accompanying sound changes, deletions of mate "e," and changes from "y" to "i." The program must also handle other forms. Stress-neutral endings like "less" and "ness" are quite simple. Four other types of endings, exemplified by "ation," "ity," "gan," and "ency," are more complex--reordering stress within the base word, and causing vowel shifts and consonant sound change. The problems with these endings are much like those for straight letter-to-sound rules, but with pronunciations of base form and ending as available information. With prefixes and compounds, the relative stress of constituent parts will depend on the intended part of speech for the target word, and on the properties of the parts. Final vowels of Greek prefixes may or may not reduce to schwa, depending on properties of the base forms they modify. With a programmed dictionary driver, we are currently finding •99.9% of words in the Brown corpus (excluding proper names and misspellings)--• 75% from exact dictionary entries, and the rest by derivation. We default to Mark Liberman's Name-Say program for words not found or derived. 10.•0 C7. Speech synthesis of Japanese from written Esperanto. With a subsystem designed to extract meaning from written Esperanto being added to the speech synthesis system of Sapanese generated in the semantic/syntactic base IS. Sato and H. Kasuya, J. Aconst. Soc. Am. Suppl. ! 73, S5 (1983}], we present an overview of this translation-synthesis system including a knowledge base editor for implementing linguistic information needed. Allowing a wide range of data modification and file manipulation operations, the frame editor can edit on a frame/slot basis the rules and dictionaries used in parsing Esperanto and generating Japanese. The main portion of the knowledge base is occupied with descriptions about case structures, selecttonal restrictions, hierarchial relations among items. etc. for detecting anomaly and ambiguity of a given Esperunto sentence. The scmantic• gives to the meaning a case frame expression, from which the sentence generator produces a surface form with some trimming with phonological and accentuation rules. The synthesis program further converts the surface into acoustic paramete• rendering a time series of control parameters for tbe terminal analog synthesizer. 10:lS cg. Text-to-speech synthes's of Japanese based on the Cepstrum-CV method. Yasushi Ishikawa and Kunio Nakajima (Mitsubishi Electric Corp., A text-to-speech synthesis system for Japanese is described. In synthesis of Japanese, generation of fundamental frequency pattern is one of the most important problems. We proposed the prosody rules that generate intonation pattern of any Japanese text using syntactic information. The prosody rules consist of modification rules, syntax rules, and a generation rule. Connecting several words, the system makes an accent phrase and gives it an accent using the dictionary and the rules (modification rules). The parameters used in a subsequent generation rule are modified on the basis of the syntactic structure of the input text (syntax rules}. Then, a fundamental frequency pattern is generated from these parameters (generation rule}. In this system, the synthesizer parameter is Cepstrum. Parameter rules assign segment duration and concatenate CV syllables Cepstra. Thus, this system translates any text of Japanese written with the syntactic symbols into good quality and natural speech. 9:4S C6. Stress assignment in letter-to-sound rules for speech synthesis. This paper will describe a program that attempts to determine word stress from spelling. Four sources of constraints will be discussed: (1) syllable weight, (2) part of speech, (3) morphology, and (4) etymology. Stress assignment is a well-established weak point for many speech synthesizers because stress dependencies cannot be determined locally. It is impossible to determine the stress of a word by looking through a five-or six-character window, as many speech synthesizers do. Well-known examples such as degrhde/d•gradhtion and t•legraph/telbgmphy demonstrate that stress dependencies can span over two and three syllables. In addition, examples such as calculi and tottoni demonstrate that stress is sensitive to etymology, a nonlocal property of the entire word. Calculi, like most other latinate forms, obeys the main stress rule of English, which assigns antipenultimate stress to nouns ending in two light syllables. In contrast, tortoni, like most other Italian and Japanese loan words, receives penultimate stress. Stressing Italian and Japanese surnames with the main stress rule of English as some commercial synthesizers do, leads to unacceptable results, as will be demonstrated. Our program produces much more acceptable results by estimating the etymology (by counting three letter tri-grams} and then adjusting the stress rules appropriately. 10:30 C9. Fast seq ential closed-phase glottal inverse filtering based on optimaily weighted recursire least squares. T. C. Luk (Baldwin Technology Corporation, $118 S. Dansher Road, Countryside, IL 60525) and J. R. Deller, Jr. Presented in this paper is the theoretical basis, with simulation verification, for a sequential method of deconvolution of the glottal waveform from voiced speech. The technique is based upon a linear predictive model of the vocal tract, and the assumption of a "pseudo-closed phase" (PCP} (noisy closed phase} of the glottis during each pitch period. Existing techniques for closed phase glottal inverse filtering (CPIF} employ "batch"type methods which are generally slow, highly user interactive, and restricted to the use of one cycle of data in the analysis. The basic ideas underlying CPIF, and a brief review of existing methods, will be presented in the first part of the paper. In the second part of this paper, requisite theoretical results and the new method will be developed. In particular, these will include a unified theory of CPIF in which the selection of closed phase points is viewed as a data weighting process. This viewpoint readily admits the use of more than one cycle of data in the analysis {advantageous $7 This paper proposes a new hackward-type speech coding system (ADPCM-AB} for wideband {7 kHz} speech and sound signals. In this system, a split-band adaptive predictive coding scheme with gradient PARCOR lattice filter and a dynamic bit allocation scheme are employed, where quantization bits are dynamically allocated over the subbands {frequency), the subintervals {time), and the channels (stereo) in accordance with the distribution of the residual energies. They serve to remove the redundancies due to the periodic concentration of the prediction residual energy and the nonuniform nature of the speech spectrum. The ADPCM-AB needs neither longer delay time than 4 ms nor transmission of side information parameters because the parameters for predictive coding and dynamic bit allocation are calculated with the locally decoded signal• It is clarified that the ADPCM-AB system has the best speech quality among the conventional backward-type coding systems. It is also shown that this system provides speech quality subjectively equivalent to 1 l-bit linear PCM {176 kb/s} at 64 kb/s. 11.•0 CII. Objective quality measures applied to enhanced speech. Previously reported experiments have indicated that computer enhancement of single channel speech in wideband noise generally does not result in increased intclligibility. Our research goal has been to investigate the effect of such enhancement on speech quality. Our experimental framework employed natural speech degraded by additive white, nonwhite, and aircraft cockpit noise over a wide range of signal-to-noise ratios the report of Kaiser and David, in no case was speech intelligibility improved as a result of processing. Apparently, the benefits of removing remote interference {temporal and spectral} do not outweigh degradations of the target signal. [Work supported by NIH.] S8 Control rooms designers typically measure and specify rooms according to their physical structure and acoustic properties. They are unable, however, to measure or predict how well the room will support the subjective qualities of stereo imagery produced over loudspeakers. As the quality and sallenee of stereo imagery improve through the use of more sophisticated recording and processing techniques, control room requiremeats become more stringent. Beyond speaker placement, there are three primary factors that influence the perception of stereo images: time-energy-frequency characteristics of the speakers, spatio-teraporal distribution of early reflections, and the inclusion of acoustic diffraction. These are easily measured through the use of time delay spectrometry {TDS), but at present an adequate model for predicting subjective response from these physical measurements is lacking. Ensuring the perception of optimal stereo imagery requires the application of standardized subjective evaluation techniques. Currently under development at Northwestern Computer Music (NCM} is an evaluation technique using the Listening Environment Diagnostic Recording {LEDRXM), which enables the immediate assessment of changes in stereo imagery that result from progressive changes in control room acoustical treatment. Field tests indicate that LEDR TM is valuable in the design and modification of control rooms for optimizing stereo imagery. 9:• D2. Recording control room design incorporating a reflection-free zone and reflection phase grating acoustical The acoustical diffusing properties of fiat, mono, and bieylindrical surfaces, alternating absorptive and reflective panels, and a wide variety of reflection phase grating diffusors (RPG) will be discussed. Experimental TEF measurements including energy-time, energy-frequency, time-energy-frequency, and polar angle-energyfrequency curves plus theoretical Kirehhoff modeling calculations will be presented. A new control room design incorporating a reflection-free zone (RFZ), in the front half of the room, and a diffuse zone, in the rear half of the room created with RPG diffusors, will be described and documented. The RFZ minimizes the speaker boundary interference over a wide volume surrounding the mix position. The RPGs provide a diffuse sound field which enhances the perception of spatial textures and helps maintain the stereo perspective across the entire width of the mixing console and in the rear of the room. The RFZ/RPG design minimizes frequency coloration, image shifting, and provides accurate stereo imaging. 10.'00 D3. Optimization of monitoring signal refieetiun i•atterns in recording studio eontrol rooms. L M. Wrightson IJoiner-Pelton-Rose, Inc., 4125 Centurion Way, Dallas, TX 75244} and Brad S. Brubaker IDepartment of Psychology, University of Wisconsin, Milwaukee, WI 53201} Recent control room design practice has often included rear/side wall reflective surfaces to provide a discrete, delayed reflection of the direct monitor speaker signal. This delayed signal is typically directed to the ear contralateral from the direct sound. To assess the perceptual effects of this practice, listeners were presented with pulses, speech, and music to both ears with a delayed repetition to a single ear. The level and delay duration of the tooneural signal were varied. Under some conditions, 50 additional channels of delay at low intensity were presented diotic, a!ly to airnulato room r•fiectiona. Listeners were asked to make judgments of detectability for the motmural signal and to describe any lateralization changes in the undelayed signal when differences were heard. These conditions were repeated in a control room environment. The results of these judgments indicate that addition of the tooneural delayed signal is easily detected for most conditions and substantially alters the spatial perception of the undelayed signal. $9 Physically small rooms are often erroneously analyzed and measured according to the classical statistical acoustical equations and techniques. Evidence is presented herein to support the position that a truly diffuse reverberant field does not exist in such spaces at a level above that of the ambient noise. 11.'00 DS. Isolating music and mechanical equipment sound sources with gypsum board partition systems. H. The 100-Hz low-frequency limit for sound transmission loss measorements and the STC rating system have seriously limited the development of a vocabulary of practical lightweight constructions which can effectively isolate music and mechanical equipment sound sources. This paper discusses the results of wcent research on gypsum board partition systems which include sound transmission loss measurements down to 50 Hz and the evaluation of a number of design factors such as balanced and unbalanced constructions, panel damping, cavity resonances, and cavity insulation. Measurements made at 50-, 63-, and 80-Hz one-third octave bands are in good agreement with calculated TL at these frequencies. These data, plus calculated TL at 31 and 40 Hz, are used to further evaluate the proposed MTC (music and mechanical equipment transmission cla•) rating system which is based on 125-to 5000-Hz data. 11:30 D6. Review of some approaches to architectural-acoustic design of sound recording facilities in the U.S. S. R. Gregory A. Kacherovich (Jaffe Acoustics, Inc., 114A Washington Street, Norwalk, CT 06854) Rooms are classified for their purpose in relation to acoustics, based on the authors's experience in architectural-acoustic design and consulting in the U.S. S. R. for over 20 years. General acoustic requirements are set for each group of rooms. Auditocia are classified depending on the sound source origin. Special attention is brought to sound recording studios and related spaces used for radio, television, and motion picture facilities. General acoustic criteria, general approach in design, selection of size, shape, and materilds are discussed for different groups of rooms including motion picture and television stages, music and speech recording studios, re-recording studios, control rooms, etc. Music recording studios are considered to be of two major types depending on the use of distant and close pickup microphones. Several multi-purpose studios were designed and built in the 1970's providing the necessary acoustic environment for music and speech recording as well as for re-recording purposes. Systems for variable acoustics were developed providing a wide range of variability. The standard cosmoiogical model is based on the approximation that the universe is homogeneous and isotropic. The overwhelming success that this model has enjoyed is based on its ability to predict the primordial abundance of elements and the isotropy of the microwave background. However, the universe itself is not homogeneous. Any realistic model must include the perturbations from homogeneity that evolved into galaxies and galaxy clhsters. Perturbations on these scales behaved like sound waves at sufiicienfiy early times when the gas in the universe was almost completely ionized. In this talk I will review the work done on the linear and $10 J. Acoust. Soc. Am. Suppl. 1, VoL 78, Fall 1985 110th Meeting: Acoustical Society of America $10 ES. Gravitational waves: A new wfndow for astronomy. Peter F. Michelson (Physics Department, Stanford University, Stanford, CA 94305} Gravitational waves were predicted more than 50 years ago by Einstein as a consequence of the general theory of relativity. Because of the weakness of the gravitational interaction, efforts to directly detect gravitational waves have focused on astrophysical sources rather than terrestrial sources. In laboratories around the Sl I J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America Sll world, second-generation cryogenic acoustic detectors and laser interferometric detectors are being developed with sensitivity and bandwidth sufficient not only to verify directly the existence_of gravitational waves, but also to study the received waveforms. Because the gravitational radiation emitted by an astrophysical source contains information about the source that is orthogonal to information obtained from electromagnetic signals, the direct detection of gravitational waves will open a new window for astronomy. In this review, the likely astrophysieal sources of gravitational radiation and the technology of second-generation acoustic detectors will be discussed. 11'.30 E6. Vnbrations in the environment of large future telescopes. E. T. Pearson (National Optical Astronomy Observatories, Advanced Development Program, 950 North Cherry Avenue, Tucson, AZ 85726) As larger telescopes are designed with more servo-controlled elements, the chance of undesirable vibrations increases. What are the sources of the vibrations and how can the telescape be designed to avoid problems? If problems can be created by acoustical and mechanical vibrations, might there also be a use for input vibrations? TUESDAY MORNING, 5 NOVEMBER 1985 A new hypothesis for the processing of acoustical information by bony fish is presented. The hypothesis proposes that the ear itself can perform most of the calculations required to localize a sound source, and discriminate frequency and beam form for signal to noise enhancement. Prior theories assume all of the analysis occurs in the central nervous system S13 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America S13 (CNS}. A mechanism is demonstrated by which the ears may resolve the particle velocity into three vectorial components and determine the time derivative of the acoustic pressure. This information can be used to localizc a sound source, assumiffg that the CNS is capable of evaluating the ratio of the appropriate velocity components. If localization is accomplished in this manner, it follows that the CNS would also be capable of finding the ratio between the magnitude of the acoustic particle velocity and the time derivative of the acoustic pressure. This ratio is shown to be proportional to frequency in the farfield, and the product of the frequency squared and the distanc• from the source in the nearfield. Thus discrimination between frequencies is possible at any distance, absolute frequency can be determined in the farfield, and some capability to discriminate range might also be possible. Supporting evidence from the literature is presented. [Work supported in part by ONR and NIH.] TUESDAY AFTERNOON, 5 NOVEMBER 1985 The physical characteristics of the available magnetic media which recommend them for audio sound recording and reproduction, by both analog and digital processes, will be discussed. Inherent characteristics of each medium which affect sound quality, dynamic range, and signal-to-noise ratio will be described. G2. Optical media, their characteristics and potential for sound recording and reproduction. Ronald Uhtig and Allan Marchant (Eastman Kodak Company, Rochester, NY 14650) An overview of the present-day optical processes, including film photography, magneto-optics, and laseroptics, will be discussed. The features of the processes and local defects of the media which affect sound quality, dynamic range, and signal-to-noise ratio will be described. 3,-OO G3. Error correction and concealment in digital recording. Bart N. Locanthi (Pioneer North America, Inc., Development Laboratory, Duarte, CA 91010) No medium is perfect. The imperfections range from accumulations of dirt on a film or disk to the microscopic absence of the recording medium on its supporting substrate. In digital reproduction, bits, or even words, are lost with potentially spectacular results. The nature of these defects, and the technologies that have been developed to overcome them, ,will be discussed and demonstrated. 3'.3O G4. Technical-esthetic ennsiderations for sound recording. John M. Eargle IJBL Inc., 8500 Balboa Boulevard, Northridge, CA 91329 and $ME Associates, Los Angeles, CA 90068) From the very beginning of sound recording, physical limitations of the medium have dictated certain esthetic constraints in the use of the medium. The major technical epochs have offered advantages to both engineer and producer, and the historical progression has provided solutions to problems in playing time, dynamic range, distortion, and space perspective. The problems associated with major format changes are almost predictable at this point. With each major change, producers and engineers have had to redefine the best fit of the message into the medium. They have invariably stumbled at first, and then rapidly learned to work the medium to the advantage of the music. In this paper we will review the fascinating interplay which has taken place between esthetics and technology over the years, concentrating on those problems assaciated with the coming of digital recording. S14 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America S14 G5. State of the art commercial music recording prsctiee. Richard P. Blinn (Capitol/EMI Corporation, Hollywood, CA 90028) The variety of steps which take place between the artist's performance and the production of the final product for sale to the consumer will be explained and their effect on the final reproduced sound will be demonstrated by means of specially prepared recordings. Panel Discussion An overview of the research on tactile aids for the deaf, blind. and deaf-blind will pay particular attention to the problems of encoding the environmental signals in ways that the skin senses can interpret them efficiently. Following a brief summary of the early approaches to the development of systems for sensory substitution, there will be a closer examination of the efforts mounted in the past 20 years. The talk is intended to include illustrations of successful and unsuccessful efforts at device and system deYelopmcnt from scientific and engineering studies. In addition, there will be recounted entrepreneurial experiences of a number of investigators and developers of devices and methods for improving communicative skills in the sensorily disabled population. [Work supported by NIH Grant 04755.] $15 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of America $15 2.'O0 H2. The applimtion of princip•!-enmponents specfral analysis to sensory aids for the hear•g impaired. In order to present speech information through an alternate sensory modality--such as touch or vision-extensive processing of the speech signal is required. This processing should represent a large amount of speech information with a small number of slowly varying parameters. These parameters should be relatively speaker independent, robust in the presence of noise, and closely linked to the perccptually importsnt features of speech spectra. One such set of parameters are speech spectral principal components. In this paper we will discuss the application of principal-components analysis of speech spectra to speech receding for use with sensory aids. A brief overview of the analysis procedure will be included. An interpretation of principal components, and a comparison with other speech parameters such as speech formants, will be presented. Both potential benefits and drawbacks of this analysis procedure will be discussed. In addition, the implementation of a vowel articulation training aid based on principal components will be presented. Although the vowel training aid utilizes a visual display format, with color as a primary information-bearing parameter in the display, the procedure used for obtaining principal components applies equally well to a tactile aid. 2'.3O tt3. T•ztile aids: A comparison of single and multiehannel devices. The paper describes a case study of Dimitry Kanievski, a deaf individual who has been using a vibrotactile aid for appro 'xmmtely 13 years. He has acquired the abil/ty to lipread speakers in three languages, using the Kanievski speech-nnnlyzing device. The report describes his communica-110th Meeting: Acoustical Society of America $16 $17 S19 2:• JS. Perceived pitch (tonality) of five unfiltered nonsense syllables and their identification and quality assessment through six l-oct filter bands. S20 Research involving the perceived pitch (tonality) [M. M. Peterson and C] has supported the notion that phonemes may be "frequency specific." The purposes of this study were to have normal-hearing listeners (1) rate the perceived pitch of five unfiltered nonsense syllables with phoneroes homogeneously grouped according to a pitch model [C. W. Asp, J. S. Berry, and C. S. Bessell, J. Acoust. Sec. Am. Suppl. 1 64, S20 (1978)]; and (2) identify and assess the quality of the nonsense syllables each filtered through six 1-oct filter bands with center frequencies of 250, 500, 1000, 2000, 4000, and 8000 Hz. Results indicated that (l) the nonsense syllables could be ranked accord-110th Meeting: Acoustical Society of America $20 2:50 K6. Bottom penetration from a range-and depth-dependent oceanic sound duct, J. Miller, A. Nagl, and H. 'Experiments have been carried out with 30-/•s-1ong sinusoidal wave trains incident on an aluminum plate submerged in water. The excitation of plate resonances by such pulses has been observed, in a fashion used previously for cylinders [G. Maze et el., J. Aeoust. Soc. Am. 77, 1352 (1985)]. The pulse distortions (in the form of initial transients, a quasi steady-state regime, and a final transient} have been interpreted by the interference of reflected and multiple internally refracted pulses, as previously done by us for a layered ocean bottom [A. Nagl et el., Inverse Problems 1, 99 ( 1985)]. The final transient, showing a step structure, rep-S35 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1085 110th Meeting: Acoustical Society of America S35 resents the ringing of the resonance. Overlapping resonances are shown to lead to beat effects in the ringing. [Work supported by the Direction des Recherches, Etudes et Techniques, France, and by the Office of Naval Research and the Naval Research Laboratory, U.S.] 9:30 QS. Transmission and reflection of acoustic waves through a combination of elastic and liquid layers. $acob George (Mail Stop 181, Submarine Signal Division, Raytheon, Portsmouth, RI 02871} We extend the formalism of Folds and Loggins [J. Acoust. Sec. Am. 62, 1102 {1977}] to calculate acoustic transmission and reflection coefficients for a plane wave at oblique incidence on a .combination of elastic and liquid layers of plane parallel plates. We discuss the modified boundary conditions and show how the acoustic field is calculated at any point inside an elastic or liquid layer. We present numerical results for the case of steel and water layers. 9:45 Q6. Surfnee wnve velocity men•urement• nt fluid-porous solid interface. The presence of ultrasonic surface waves of various modes on a fluidporous solid interface is demonstrated and their velocities are measured. The experimental technique developed earlier [A. Jungroan et eL, 1. Appl. Phys. 53, 4672 { 1982)] for a fluid-isotropic solid interface utilizes reflected broadband spectra from periodic surfaces. Three sharp minima corresponding to mode coupling of incident waves into surface waves at the fluid-porous solid interface are observed. The velocities of these surface waves are in qualitative agreement with theoretical predictions IS. Feng and D. L. Johnson, J. Acoust. Sec. Am. 74, 906 { 1983}] and are identified as pseudo-Rayleigh, pseudo-Stoneley, and true StoneIcy waves. [This work is supported by the Experimental results are presented for the scattering of an acoustic pulse from a region ofturbulcot flow. The investigation was performed in the 7-m-deep section of the U•NA Hydrodynamics TOW Tank facility. Turbulent flow is produced by a submerged nozzle of 4. Scm-diam and 100-em/s exit velocity. Flow is directed perpendicular to the water's surface. The 1.9-MHz acoustic pulse is generated from a 2.54-cm-diam ultrasonic transducer located 18-nozzle diameters downstream from the jet exit and at a perpendicular distance of 84 cm from the jet axis. An identical transducer receiver located at the depth of the sender is positioned 217 cm from the jet axis. Angular scattering measurements arc made in a plane parallel with the water's surface. Scattering results are presented for the temporal broadening, the spectral broadening, and the energy distribution of the incident pulse using a digital data collection and analysis system. Results are compared with theoretical predictions. [Work supported by NRL (Physical Acoustics Branch) and USNA.] S36 Center.] 9:15 R4. Dynamic and perceptual constraints of loud speech. In an investigation of the production of loud speech, it is found that, in comparison with normal speech, these productions are characterized by a substantially lowered jaw position for all vowels. On the basis of these results, we examine whether, for loud speech, speakers maintain tongue constrictions typical for normal vowel production in spite of the lowered jaw position, or if the tongue follows the jaw using constrictions atypical for the particular vowel. Acoustic analyses of the loud tokens show an increase in F 1, thus indicating a lowering of the tongue. Tests show, however, that listeners perceive the intended vowel quality, and not that which might be inferred from the loud formant values. The characteristics of the loud speech targets will be discussed in light of the dynamic constraints of the articulatory system and the perceptual constraints set by the listener. 9:30 RS. Acoustic-phonetic properties of vowels produced in noise. David It is well known that speech produced in noise has higher intensity, longer duration, and higher fundamental frequency than speech produced in quiet. However, recent studies in our laboratory have indicated that changes in the acoustic properties of speech produced in noise are not limited to changes in the glottal source but may also include significant changes in the acoustic-phonetic structure of speech reflected in the characteristics of vocal tract transfer function. We investigated the effects of wideband noise on the acoustic properties of vowels. Talkers produced ten American-English vowels in /hvd/context. Ten tokens of each vowel were produced in quiet and in the presence of masking noise 190 dB SPL I. The results indicate a rather complex pattern of changes in formant frequencies. First, with the exception of the vowels/•a,•/, whenever a significant change in F 1 was observed. F 1 frequency was higher in the noise condition. Second, F2 frequencies decreased in front vowels and increased S37 110th Meeting: Acoustical Society of America $39 108 consists of members of S2. S3, and other persons not necessarily members of those committees). There will be a report on the meeting of ISO/TC 108 and Subcommittees 1, 2, and 4, held from 2-13 September 1985, in Vienna, Austria. S50 The quality of •ound pickup in large rooms---such as anditoria, conference rooms, or classrooms•is impaired by reverberation and interfering noise sources. These degradations can be minimized by a transducer system that discriminates against sound arrivals from all directions except for that of the desired source. A two-dimensional array of microphones can be electronically beam steered to accomplish this directivity. This report gives the theory, design, and implementation of a microprocessor system for automatically steering a two-dimensional microphone array. The signal-seeking transducer system is implemented as a dual-beam, "track-while-sean"array. It utilizes signal properties to distinguish between desired speech sources and interfering noise. The complete automatic system has been tested in anechoic and medium-sized auditorium environments, and its performance is discussed. '• Present address: Dornier System, D7990 Friedrichshafen, West Germany. 8:50 AA2. A simple second-order toroid microphone. 7. E. West (AT&T Bell Laboratories, Murray Hill, NJ 07974) and G. M. Sessler {Technical University, Darmstadt, West Germany) SS2 J. Acoust. Sec. Am. Suppl. 1, VoL 78, Fall 1985 tom wall, while the exterior surface is in contract with the liquid. Although the bottom wall is designed to satisfy the pressure code (ASME Code for Pressure Piping B31, 1980 edition), it is still sufficiently thin to permit ready excitation of the axisymmetric plate modes of vibration. The liquid depth is measured by a conventional pulse-echo technique. The external threads of the plug serve as the primary pressure seal. A prototype gauge, constructed with a 3/4-in. stainless steel plug and a 10-MHz piezoelectric transducer, was tested successfully in a 600-gallon water vessel at pressures up to 5200 psi. 9.'• AA4. Experience with the two-microphone method for measuring the terminating impedance of a ducted burner during operation. R. Previous papers on the two-microphone method establish its superiority over the classical impedance tube method for measuring the terminating impedance of a duct. The technique has been adapted here to measure the terminating impedance of a ducted burner with combustion noise acting as the sound source. The measurement system encounters a large dynamic range (30 dB) and high temperatures (900 K}. Flow effects are negligible since Mach numbers remain below about 0.03. Results are compared to similar data from the literature, obtained using the classical impedance tube method, for duct terminating impedance in the presence of a hot flow. The method permits in situ measurement of the terminating impedance of a ducted burner during operation in spite of the hostile environment within the duct. AAS. Comparison of the tharmuaeoustic efficiency of premixed and diffusion hydrogen-flame ducted burners. It has long been known that confining a flame within a duct has an influence on the efficiency with which it converts chemical energy to acoustic energy. The present paper describes an experimental program aimed at explaining and quantifying this influence. Results are presentod for premixed and diffusion hydrogen-flame ducted bumera of otherwise similar design operating in the 5-to 10-kW range. The results show that the premixed flame yields the higher thermaacoustic efficiency at a given thermal power level, and that its efficiency is sensitive to the power level. The thermoacoustic efficiency of the diffusion flame is lower and relatively insensitive to the power level. The results are explained in terms of Rayleigh's criterion and current combustion noise generation theories. AA6. The electroacoustic description of an underwater msgjletohydrodynsmic transducer. Stephen C. Schreppler and 11ene The Loreritz law, electromagnetic force in a conducting-dielectric, compressible, fluid medium, is investigated as a source of underwater acoustic radiation. The transducer system investigated consists of a duct which confines a conducting-dielectric fluid medium. A time-steady magnetic field and mutually orthogonal time-varying current density are established across the transverse dimensions of the duct. The system radiates from the longitudinal ends of the duct in a dipole, two-point array mariner. The electroacoustic system equations in two port form, theoretical and experimental projector sensitivity, and pressure field directivity patterns will be discussed. [Work supported by ONR.] 10:05 A.AT. A coupled finite difference model for n fluid-loaded, free-flooded, thickness-polarized, piezoelectric cylindrical shell transducer. Chwei-This paper presents a coupled finite difference model for a fluid-loaded, free-flooded, thickness-polarized, piezoelectric cylindrical transducer. In order to find the vibration response of this shell when subjected to a sinusoidal voltage drive, the Donnel's thin shell equations of axisymmetric free vibration of a circular cylinder are coupled with the piezoelectric canonical equations and a surface Helmholtz integral model for the acous- S53 A new method for absolute calibration ofelectroacoustic transducers is described which permits 1ow-f-•luency reciprocity calibration in a test apparatus of small dimensions. The method takes advantage of the reduced wave speed in a cylindrical water column bounded by a compliant PVC tube to reduce the length of the standing wave to dimensions which can be easily handled in the laboratory. Results of a limited series of experiments show a mean reproducibility of + 0.8 dB in a l-m-long tube at frequencies between 750 and 1100 Hz and sound speeds from 330-365 m/s. The hydrophone voltage sensitivities obtained by the method, however, systematically differ by 1.5 dB from those measured by comparison with a standard hydrophone. Additional research will be required to resolve this discrepancy. [Work supported by ONR.] 11:50 AAI4. Improvement to the transfer function method for determining the complex dynamic modulus of polymer eompceites. Measurements of loss factor and storage modulus of diphasic transducer composites of polymers and piezoceramic were made using the transfer function method. This method consists of exciting a mass-loaded rod into longitudinal vibrations. The complex acceleration ratio between its ends is related to the complex modulns by two coupled, transcendental equations derived from the solution of the longitudinal wave equation with appropriate boundary conditions. In theory, these equations can be solved at any frequency by an iterative procedure. In practice, convergence problems reduce the valid solutions to those just at the longitudinal resonance frequencies. Improvements made during this study enable the equations to be solved at any frequency, except where the phase of the acceleration ratio is zero. Plots of loss factor and storage modulus are presented for a number of these new composites and other high-damping polymeric materials. A comparison between present results and those quoted in other studies indicates that this method produces reliable resuits. The improvernents enable results to be obtained more quickly and with much less effort than previously required. [Work suppored by ONR.] THURSDAY MORNING, 7 NOVEMBER 1985 REGENCY BALLROOM IV, 8:30 TO 11:45 A.M. An imaging technique is developed for the three-dimensional reconstruction of a desired sound field within a playback environment. S60 J. Acoust. Soc. Am. Suppl. 1, Vol. 78, Fall 1985 110th Meeting: Acoustical Society of Amedca $60 Through proper coupling of room design and speaker placement, and deduction of the needed inputs to the system to obtain the desired outputs (for a finite length time window of data), a discrete-sliding window algorithm is implemented to obtain the necessary time-domain inputs. As a result of these procedures, the feedback/reverberatory effects of the playback environment are annihilated, leaving any desired sound field, while maintaining the three-dimensionality of the original field. Between the temperatures of 2.2' and 5.2 'K, SHe is an ordinary liquid and the sound which occurs in it is ordinary sound. At 2.2 øK the lambda phase transition occurs and below T2 as a result of the Bose-Einstein condensation SHe is a superfluid with a galaxy of sound modes; first through fifth. Superfluidity and the acoustic modes persist as T•0. Their properties and use in determining the superfluid properties of 4He will be discussed. The critical point of 3He is 3.3 'K. It too, undergoes a superfluid transition but at much lower temperatures--at about 2 X 10-3 øK. Both above and below this temperature there is a new sound mode called zero sound. It exists both as a transverse and a longitudinal wave. Evidence that this mode arises bccause 3He is a viscoelastic liquid will be presented. Zero, first, second, and fourth sound have been observed in the superfluid phases of 3He. [Work supported by ONR.] 9:30 EE2. Nonlinear acoustics at low temperature. The anamolous dispersion, strong nonlincarity, and small transport coefficients of He 4 make it an ideal medium for establishing the frontiers of nonlinear acoustics. These properties make it possible for He s to exhibit a well-defined three-wave interaction, where the high-frequency wave is analogous to thc pump in a parametric amplifier (the low-frequency waves are the signal and idlcr). In fact, a macroscopic coherent sound wave will decay dramatically fast due to the parametric amplification of its first phonon of spontaneous decay [J. S. Foster and S. Putterman, Phys. Rev. Lett. 54, 1810(1985)]. It should bc possible to observe similar effects with capillary waves (ripplons). The properties of He s as well as the roton minimum in its dispersion law also make it an ideal medium for searching for two-and three-dimensional localized states which are generalizations of the nonlinear Schr'6dinger equation and Kadomtsev-Pctviashvili solitons. Finally, He 4 should be used to probc the spectrum of acoustic turbulence. The self-similarity of the stationary turbulent state is a symmetry which will be broken quantum effects at high frequency. 10:00 EE3. Surface acoustic wave investigation of superconducting films by means of the acoustoeleetric effect. The piezoelectric interaction of surface acoustic waves (SAW) with metallic films deposited on a piezoelectric substrate is quenched when the film becomes superconducting. This acoustoelectric interaction should be proportional to the sheet resistivity of the film in the limit where the SAW period is smaller than the acoustoelectric relaxation time. The attenuation coefficient of SAW propagating through superconducting films of granular Pb and NbN has been measured at 700 MHz. In both cases the attenuation is not proportional to the measured dc resistivity of the films. The attenuation remains finite in the superconducting state of the granular Pb film even when the sheet resistivity disappears. This difference is ascribed, with a percolation model, to the fact that the dc current measures the resistance of the whole film while the SAW measures the effective resistance of a sample with dimensions comparable to its wavelength. The attenuation in the NbN film attains its normal value when the temperature increases slightly above the Kosteriitz Thouless transition where antiparallel flux line pairs dissociate, while at this temperature the film resistivity has barely changed from zero. A model has been proposed that ascribes the attenuation to the presence of normal cores in the flux line pairs. Canahl [J. Acoust. Soo. Am. 50, 471-474 ( 1971 }] measured thresholds for a l-kHz sinusold masked either by two or by four surrounding tones. He reported four-tone masked thresholds that exceeded, by 5-7.5 dB, the energy sum of the masking produced by the individual tone pairs. The present paper reports on a series of experiments investigating the effects of several factors on this 5-7.5 dB "excess" masking. In each experiment, thresholds for a 1-kHz, 250-ms sinusoid were measured as a function of the overall level of two or four equal amplitude sinusolds with frequencies arithmetically centered around 1.0 kHz. For conditions similar to those of the Canahl experiment, 5-6 dB of excess masking were obtained indepen-S62 Standards Committees S3, Bioacousties. The current status of standards under preparation will be discussed. In addition to those topics of interest including hearing conversation, noise, dosimeters, hearing aids, etc., consideration will be given to new standards which might be needed over the next few years. The international activities in ISOfFC 43 Acoustics and IEC/TC 29 Electroacoustics, for which SI and S3 serve as the U.S. Technical Advisory Groups, will be discussed. There will also be reports on the meetings of ISOfrC 43 and IEC/TC 29 held in Budapest, Hungary, in April 1985. $65 63110} According to traditional thinking, the tectorial membrane (TM} is a gel with a matrix of proteins, but containing no collagen. We recently reported patterns of amino acids and proteins of the TM consistent with the presence of substantial amounts of collagen [I. Thalmann et al., J. Acoust. Soc. Am. Suppl. 1 77, S93 (1985)]. Using two-dimensional polyacrylamide gel electrophoresis with extended alkaline range, we have now found that the predominant protein of the TM of the guinea pig virtually superimposes with purified collagen standards. Moreover, peptide mapping of cyanogen bromide digests ("fingerprinting"} results in a pattern closely resembling that of collagen type II. Further procedures for identification and characterization of this protein (effect of collagenase, immunoblotting, etc.) are in progress. At this stage it is not possible to decide to what extent the amino and neutral sugars present in hydrolysates of the TM are part of the alleged collagen molecule. The cupula ampullaris, a vestibular superstructure thought to be analogous to the TM, exhibits a substantially different chemical profile. The functional significance of the findings will be discussed. [Supported in part by NIH.] S66 Synchrony of discharge to two-tone stimuli and "synchrony suppression" are analyzed by examining the implications of the definition of vector strength. "Synchrony suppression," defined as the reduction in the vector strength for one tone when a second is added, occurs by definition when half-wave rectification occurs in an otherwise linear system. The 10.'45 NNI0. Perceptual learning of synthetic words and sentences.
doi:10.1121/1.2023018 fatcat:3lbckbce2rglzj7zx3fp4nhiui