Index of content:
Volume 51, Issue 1A, January 1972
- PROGRAM OF THE EIGHTY‐SECOND MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session A. Speech Communication I: Perception
- Contributed Papers
51(1972); http://dx.doi.org/10.1121/1.1981612View Description Hide Description
Temporal order judgment (TOJ) in dichotic listening can be a difficult task. Previous experiments that used two speech stimuli on each trial (S/S) obtained sizable error rates when subjects were required to report which ear led (TOJ‐by‐ear). When subjects were required to identify the leading stimulus (TO J‐by‐stimulus), the error rate increased substantially. Apparently, the two speech stimuli were competing for analysis by the same processor, and so were overloading it. The present experiment used the same TOJ tasks, but presented a speech and a nonspeech stimulus on each trial (S/NS). The error rate was comparable to that of S/S for TO J‐by‐ear, but did not increase for TO J‐by‐stimulus. This would be expected if the speech and nonspeech stimuli are being sent to different processors, each of which performs its analysis without interference from the other. The interpretation of the data given here is consistent with the results of standard identification experiments reported elsewhere: when asked to identify both stimuli on each dichotic trial, subjects made many errors on S/S, while performance was virtually error‐free on S/NS.
51(1972); http://dx.doi.org/10.1121/1.1981613View Description Hide Description
Experienced Morse code operators showed significant right‐ear superiority, indicating left‐hemisphere lateralization, for the perception of dichotically presented Morse code letters. No significant lateralization was found for rapid monotically presented Morse code words. Subjects who did not know Morse code did not show significant lateralization when dichotically presented with a set of stimuli which included all Morse code letters; however, they showed a tendency toward left‐hemisphere lateralization with a set of Morse code letters which were restricted in duration and presented at relatively low intensity. Our results show that articulability is not a necessary property of stimuli lateralized to the left hemisphere in dichotic listening; a possible further interpretation is that it is language rather than speech that is lateralized to the left hemisphere.
51(1972); http://dx.doi.org/10.1121/1.1981614View Description Hide Description
The distinction between categorical and continuous modes of perception has played a central role in discussions of speech perception. Recent findings of Fujisaki and Kawashima [J. Acoust. Soc. Amer. 47, 57 (1970)] suggest that these two modes are not necessarily dichotomous but may be related to the degree to which some brief auditory short‐term memory store is employed during discrimination. The present experiment examined discrimination performance within and between phoneme boundaries using a delayed comparison AX same‐different task. The delay interval between presentations of the standard and comparison sounds varied over a range from 0–2.0 sec for each of four types of synthetic speech stimuli. For steady‐state vowels accuracy of discrimination both within and between boundaries was related to changes in the delay interval. Discrimination of stop consonants remained relatively stable both within and between boundaries. The results suggest that vowels (even as short as 50 msec in duration) are stored differently than stop consonants in auditory short‐term memory. [Research supported in part by grant to Haskins Laboratories from NICHD.]
51(1972); http://dx.doi.org/10.1121/1.1981615View Description Hide Description
Spoken dialogue was systematically interrupted to prompt listeners' recall for one of the final two sentences which they had just heard. Besides change of speaker, three experimental conditions were defined, based on the more immediate speech. These were overt recurrence, implied but linguistically deleted recurrence, and lack of recurrence of critical phrases from the sentences to be recalled. Shared linguistic information in immediate memory by speakers and listeners as they talk is proposed to account for facilitation of recall in the implied recurrence condition. This information is coded to permit the deletion and recovery of linguistic segments in the production and perception of speech. Implications for regulation of the flow of information in conversational speech are noted.
51(1972); http://dx.doi.org/10.1121/1.1981616View Description Hide Description
Recent research has suggested that auditory perceptual processing and short‐term memory are interrelated and temporally biased. A recommended manner for investigation of this concept is to co‐vary stimulus duration (SD) and inter‐stimulus interval (ISI) in recognition and recall tasks. The purpose of this study, then, was to investigate the recall accuracy of seven‐word first‐ and second‐order sentential approximations, co‐varying SDs of 200, 300, and 400 msec, with ISIs of 100, 200, 300, and 400 msec. Ten monosyllabic sententlal strings of each order were read by a trained male speaker under controlled experimental conditions. Twelve experimental tapes representing the 12 possible SD/ISI combinations were then manually prepared. One hundred and twenty college‐age normal hearing listeners heard the 20 sentential strings as processed under one of the 12 conditions and were required to recall the seven‐word strings. The results of an ANOVA revealed significant main effects for ISI, WD, and Order, as well as interactions of WD×ISI, and WD×Order. Recall accuracy increased as WD increased, as ISI increased, and as Order increased. The WD×ISI and WD×Order interactions revealed that recall proficiency could be enhanced via a trading relationship between the interacting factors. The results are discussed in terms of their theoretical and clinical implications.
Memory Effects in Auditory Presentation of Digit Sequences Related to Presentation of Verbal Material51(1972); http://dx.doi.org/10.1121/1.1981617View Description Hide Description
Two experiments investigated accuracy of recall for digit sequences of length 7 and 10 which were presented with other verbal material. In study 1, subjects listened to a complete recorded message before attempting to recall the digit sequence imbedded in the message. Study 2 required that subjects only “listen to as much of the message as you need….” More errors were made when recall was delayed by additional verbal material. When 10 digits were presented, but only 7 were required to be recalled, error rates were higher than when 7 digits were presented. Requiring the subjects to reorder the digit sequence before recall increased error rate. Systematic effects of alternative message phrasings were observed on both error rate and listeners' subjective evaluation of relative difficulty. Error rates in the two studies were generally similar.
51(1972); http://dx.doi.org/10.1121/1.1981618View Description Hide Description
The purpose of this investigation was to examine the relationship postulated between illusory changes of repeated words (“Verbal Transformations” or VTs) and illusory presence of phonemes replaced by noise (“Phonemic Restorations” or PhRs) [R. M. Warren and R. P. Warren, “Auditory Illusions and Confusions,” Sci. Amer. 223, 30–36 (1970)]. Separate groups of 20 subjects each were presented with one of four taped variations of the repeating stimulus word magistrate: (1) stimulus intact; (2) speech sound /s/ removed and replaced with 100‐Hz octave band noise; (3) syllable “gis” removed and replaced with 100‐Hz octave band noise; (4) speech sound/s/removed and replaced with a silent gap. When PhRs are evoked, most VTs involve the restored portion of the word, suggesting the usefulness of PhRs combined with verbal reorganizations in listening to speech in the presence of noise. Other effects of PhRs upon VTs were found. Implications of these results for mechanisms of speech perception are discussed.
51(1972); http://dx.doi.org/10.1121/1.1981619View Description Hide Description
Are meaningless sounds learned and identified in a speech mode? Sixteen sounds were generated differing on four dimension, each having two values. These were: the source waveform contained either all harmonics or odd harmonics; the fundamental frequency was either 90 or 142 Hz; there were either one or two formant frequency regions; and the formant(s) was (were) either low, 600 (and 1550) Hz or high, 940 (and 2440) Hz. Twenty‐four listeners identified these sounds. Fewer confusions were made between sounds (1) as the number of dimensions on which they differed increased, and (2) as the dimension(s) changed, from formant number to ferment frequency region, to fundamental frequency, to source waveform. Similarly, fewer confusions were made between sounds whose formant frequencies and fundamental frequency were in a fixed ratio than between those with the same formant frequency. In essence fixed ferment sounds were more often called the same sound than were fixed ratio sounds. The data fit a hypothesis that classification was made on a generalized “pitch” dimension.
51(1972); http://dx.doi.org/10.1121/1.1981620View Description Hide Description
Eighteen right‐handed and eighteen left‐handed young normal listeners identified the location of clicks embedded within sentences. The sentence‐click stimuli were presented monotically and dichotically to right and left ears; subjects wrote down click location with a slash mark on the sentence. Results indicate no significant difference in correct location of clicks (for either right‐ or left‐handed listeners) as a function of mode of presentation, monotic or dichotic, to either ear, i.e., no laterality effect. Size of target (click on correct syllable, correct word, but not for word and/or word boundary) significantly affected correctness of location for right‐ and left‐handed subjects. Left‐handers were generally more accurate in over‐all click location on word and syllable sized “targets.” There is also evidence of a handedness by mode of presentation interaction. Results are discussed in terms of models of speech perception and laterality‐handedness effects.
51(1972); http://dx.doi.org/10.1121/1.1981621View Description Hide Description
In continuation of an experiment described previously [D. Terbeek and R. Harshman, J. Acoust. Soc. Amer. 50, 147(A) (1971)], judgments of perceptual similarity and difference for 12 natural vowelsounds were obtained from native speakers of Turkish and Swedish. These new data, like the earlier data from English, Thai, and German subjects, were analyzed using the PARAFAC multimode factor analysis method [R. Harshman, UCLA Working Papers in Phonetics No. 16 (1970)]. The resulting perceptual spaces are roughly interpretable in terms of traditional vowel descriptions, but certain difficulties remain. The results for the five languages tested, together with results reported by other investigators, were subjected to several further types of analysis. The patterns which emerge suggest that the function relating perceived distances between vowels to their positions along underlying perceptual dimensions is non‐Euclidean in two ways. First, the perceptual dimensions do not lie orthogonally to one another, implying that they are related in meaning. Second, they interact nonlinearly, producing a curved space, an effect which causes the extraction of an uninterpretable dimension when linear models are used.
51(1972); http://dx.doi.org/10.1121/1.1981622View Description Hide Description
This study reports subject auditory reaction times to syllables (CVC, CCV, and VCC in construction) and to their constituent segments. A target syllable or segment was located randomly in each of 288 lists each containing 10 nonsense syllables. Stimulus material was presented via earphones to 12 normal hearing subjects who heard the target syllable or segment spoken before each list. Reaction time was measured from the beginning of the target to the depression of the response button by the subject. Response times, false alarms, and misses were recorded from a frequency counter on a data printer. Analysis of data indicated that, within limits, subjects reacted equally fast to syllables of different constructions, given similar durations. Contrary to the findings of Savin and Beret [J. Verbal Learning and Verbal Behavior, 1970], subject response to syllable initial segments was significantly greater than to the entire syllable. Segment response times within a syllable generally increased according to segment position from initial to final.
Contour Plots of Speech with Subglottal‐Rate Time Resolution Provide New Insight Into Perceptual Cues51(1972); http://dx.doi.org/10.1121/1.1981623View Description Hide Description
An electronic analog ear having 24 sections has a space‐time pattern that can be converted to a contour pattern in the manner of a “Voiceprint.” Each section signal is half‐wave rectified and low‐pass filtered to 400 Hz so as to retain several harmonies of fundamental voice pitch. Contours are generated with high speed commutation of the 24 section signals, a delay line smoothing device, and a level clipper. The actual experimental system, which is frequency scaled downward by a factor of 20 is described. A number of example patterns are shown for both speech and nonspeech sounds. Cues are revealed which allow the central nervous system to make localization measurements on the basis of envelope periodicity. A voice pitch mechanism is demonstrated which depends upon periodic pattern projections into the cochlea in a manner which is not well described with spectra.
- Session B. Engineering Acoustics I: Transducers and Arrays I
51(1972); http://dx.doi.org/10.1121/1.1981624View Description Hide Description
Acoustic Radiation from a spherical source eccentrically positioned within a fluid sphere embedded in another fluid is evaluated. This configuration is an idealization of a spherical acoustic lens used as a sound projector. To effect a solution, the translational addition theorem for spherical wavefunctions is used to express series of wavemodes centered at one sphere in terms of modes centered at the other, thereby facilitating the task of satisfying the boundary conditions. Because spherical wavefunctions centered at one point are not orthogonal over the surface of a sphere centered elsewhere, each mode associated with the first sphere couples into every one of the modes associated with the other. Hence, the evaluation of the modal coefficients in the series for the radiated pressure necessitates solving an infinite set of equations. Approximate solutions are obtained by truncating the set to finite size and solving numerically. The source radiation impedance and farfield directivity have been calculated for representative values of the many parameters that characterize the problem, such as wavelength size of fluid sphere and source, relative characteristic impedances and soundvelocities of the two fluids, and position and velocity distribution of the source. [Research was supported by the Ordnance Research Laboratory under contract with the U. S. Naval Ordance Systems Command.]
51(1972); http://dx.doi.org/10.1121/1.1981625View Description Hide Description
This paper presents the results of a theoretical study of the effects of water loading on a turbulent‐flow‐excited, thin, finite plate. The water loading is defined as the pressure field produced in the water, at the plate surface, as a result of the plate vibration. The propagating component of this pressure field acts in temporal phase with plate resistive forces, and the nonpropagating component acts in phase with plate inertial forces. Thus, water loading may be interpreted as additional inertia and additional resistance of the plate. In an example, comparison of computed, turbulent‐flow‐excited water loaded and in‐vacuo plate vibration spectra revealed the additional inertia to be 1.5–12 times the inertia of the plate alone. The additional resistance was found to be approximately 1% of the critical value of the in‐vacuo plate. The additional inertial and resistive forces caused the “mode” shapes of the water‐loaded plate to differ from the corresponding mode shapes of the in‐vacuo plate.
51(1972); http://dx.doi.org/10.1121/1.1981626View Description Hide Description
Experiments have been performed with a small simple structure which is designed to implode under hydrostaticpressure, and to generate sound by such processes as tearing, inrush of water, compression of internal air, and impact. The associated theory predicts the motion and deformation of the structure, and also analyzes the radiated sound as the sum of multipole components. Experiments were conducted at depths of 40, 200, and 500 ft; the structure's motions were measured by accelerometers, and the radiated soundpressures versus time were measured with hydrophones at up to eight locations about the implosion. The measurements are in substantial agreement with the theoretical predictions. In particular, they confirm that the dominant contribution to the sound is simply the monopole component p(R,t) = ρV̈/(4πR), where V̈ is the second time derivative of the volume enclosed by the wetted surface.
Approximate Model for Predicting Operational Parameters of the Class IV Flextensional Underwater Transducer51(1972); http://dx.doi.org/10.1121/1.1981627View Description Hide Description
An approximate model for predicting operational parameters of the Class IV flextensional underwater transducer is developed. The model combines approximate solutions for the acoustic radiation problem, the forced shell vibration problem, and the electromechanical problem of the piezoelectric stack. The acoustic radiation is approximated by using a source density formulation technique and is coupled to a mathematical model which predicts the vibrations of the shell coupled to the piezoelectric stack. The forced vibration solution of the shell with the acoustical loads included is then combined with the electromechanical equations governing the piezoelectric stack. The model predicts operational parameters which includes the farfield radiation pattern and the stress variations and electrical impedance of the stack.
51(1972); http://dx.doi.org/10.1121/1.1981628View Description Hide Description
The combined integral‐differential equations of a piezoelectrically driven, radiating, oval shell are put into integral form and solved by an iteration process. The solutions for maximum pressure, drive‐point impedance, and maximum dynamic shell stress are calculated and shown to compare favorably with two experimental transducers that were separately tested over a 1‐oct bandwidth. Finally, the integrals in the drive point impedance expression are interpreted in a direct way as the oval equivalent of algebraically known impedances for the more elementary case of a circular ringtransducer. By so doing we apply bounds on the “modal” stiffness and inertia as functions of oval shape and also bound the acoustic loading components both external and internal to the shell.
51(1972); http://dx.doi.org/10.1121/1.1981629View Description Hide Description
Westervelt's theory for the end‐fire parametric radiator [P. J. Westervelt, “Parametric Acoustic Array,” J. Acoust. Soc. Amer. 35, 535–537 (1963)] includes only second‐order interaction of two primary frequencies to produce radiation at the difference frequency. When higher‐order saturation effects are included, harmonics of the difference frequency become important. Experimental measurements of the growth and radiation patterns of several difference frequency harmonics are presented.
51(1972); http://dx.doi.org/10.1121/1.1981630View Description Hide Description
Pump saturation and noiseeffects on the performance of underwater acoustic receivers utilizing medium nonlinearity for beam forming are discussed. Experimental data is presented illustrating the effect of medium saturation on the receiving beamwidth and the performance limitations resulting from spurious modulation of the pump source.
51(1972); http://dx.doi.org/10.1121/1.1981631View Description Hide Description
It is known that nonlinear effects in water cause demodulation of a pulsed carrier such that a signal proportional to the second time derivative of the square of the pulse envelope develops in the insonified medium. Previous experiments [B. A. Davy, E. L. Hixson, and T. G. Muir, J. Acoust. Soc. Amer. 50, 132(A) (1971)] indicated agreement with theory for several pulse envelope shapes. An extension of this work is to predistort the envelope function by time integrating it twice and taking its squareroot. The resulting signal was then used to drive a 3‐in.‐diam circular piston at a depth of 10 ft in a fresh‐water lake. Within the bandwidth limitations of the transmitter and receiver, the original signal was recovered. A low‐pass filter was used to eliminate any of the carrier not already absorbed by the water. [This work was sponsored by the National Science Foundation.]