Volume 59, Issue S1, April 1976
Index of content:
- PROGRAM OF THE 91ST MEETING OF THE ACOUSTICAL SOCIETY OF AMERICA
- Session JS. Joint Inter‐Noise '76‐Acoustical Society of America Technical Session: Advanced Techniques—Applications to Noise Control (Noise I)
- Invited Papers
59(1976); http://dx.doi.org/10.1121/1.2002490View Description Hide Description
Before the advent of realtime one‐third‐octave analyzers, which give a better insight into different noisecharacteristics and provide a significant increase in the speed with which acoustic data can be processed, noise signals were analyzed in serial fashion into octave, one‐third octave, or “narrow‐band” results. More recently, narrow‐band real time analysis techniques have provided an entire new dimension to our ability to interpret acoustic phenomena and isolate sources of noise. However, the most significant advance has undoubtedly been the introduction of the Fast Fourier Transform and its implementation in a real‐time processing mode. This gives the ability to process signals with respect to their phase, as well as cause and effect, to a degree that was previously impossible. This paper deals with the application of two‐channel signal processing techniques, using FFT, to several signal and system related measurement problems.
59(1976); http://dx.doi.org/10.1121/1.2002491View Description Hide Description
This paper discusses the application of coherence function techniques in developing semiempiricalmodels of acoustical systems, with particular application to diesel engine noise.Pressure transducers, mounted in each of the combustion chambers, measured the cylinder pressures and a microphone mounted 1 m to one side of the engine measured the sound pressure. The frequency responses between each cylinder and the noise were computed. The frequency responses were used to predict the effect on the engine noise of varying combustion parameters such as injection timing. Previously, the effect on noise of combustion parameters has been determined by empirical means only. It is also shown that even though the inputs are highly coherent, it is still possible to determine the frequency responses for the case where the inputs are identical but delayed in time by fixed amounts, as is the case for multicylinder pumps, engines, and compressors.
59(1976); http://dx.doi.org/10.1121/1.2002621View Description Hide Description
Noisesource strength measurement and location for the case of large machines in situ can be difficult with conventional methods. This is due, in part, to the complicated motion of the large surface areas as well as the variable acoustical conditions around the machine in a typical workspace. A noise‐source identification technique based on the well known results for the sound powerW= ∫IdS, where I is the surface acoustic intensity and where the integral is taken over the component surface S, is described. The method avoids all of the problems of the contamination of the pressure field around the machine caused by surface reflections, proximity, and reverberant effects. An example of the method applied to a large centrifugal chiller is described. It is shown that the sound power spectra of the two major radiating components could be obtained. These were found to be in good agreement with sound pressure spectra measured at the AR1575 standard positions. The need for further development of the method is discussed.
59(1976); http://dx.doi.org/10.1121/1.2002622View Description Hide Description
The application of mechanical mobility and radiation factor methods to augment standard structure‐borne and airborne noise surveys in diagnosing structural contributions to vibroacoustic deficiencies on high performance vehicles is considered. Emphasis is placed on the development of a field‐oriented analysis method based on a normal mode interpretation of mobility and radiation factor measurements to characterize the dynamic response of the structural elements of such vehicles. These characteristics (i.e., natural frequencies, modal loss factors, mode shapes, relative noise transmission path strengths, modal density, termination mobilities, and acoustic radiation efficiencies) provide information which influences the selection, design, and performance of all transmission path noise control methods that are applicable to corrective redesign. The effects of modal density on such an analysis approach are briefly examined along with techniques to resolve modal information from normal modes closely spaced in the frequency domain. Applications of the methods to shipboard noise problems are presented. Mobility/radiation factor evaluations were found to provide an economical means of obtaining vibroacoustic design information for high performance vehicles where transverse modes govern the vibratory response of the vehicular shell.
- Session A. Psychological Acoustics I: Masking
- Contributed Papers
59(1976); http://dx.doi.org/10.1121/1.2002623View Description Hide Description
Forward masking, backward masking, and gap masking were employed to investigate temporal processing by hearing‐impaired and normal listeners. Masked thresholds for a 5‐reset 2500‐Hz tonal signal in the presence of a narrow‐band (1175–1725 Hz) noise masker were measured with a 21FC procedure. (This study represents an extension of previous work in a lower frequency region.) In the gap masking conditions, the noise was on continuously except for a brief silence, or “gap,” of 200‐, 50‐, or 20 msec. In general, hearing‐impaired listeners had less steep backward masking functions than did the normal Ss. They also showed less decrease in masking during the gap. As the duration of the gap decreased, the difference in the gap masking patterns between normal and hearing‐impaired Ss increased. The implications of these results for temporal processing by hearing‐impaired listeners will be discussed.
59(1976); http://dx.doi.org/10.1121/1.2002624View Description Hide Description
For listeners with hearing loss due to cochlear damage, we have shown that in frequency regions displaying threshold elevation, there is a marked alteration in the psychophysical tuning curve [J. Acoust. Soc. Am. 58, S71 (A) (1975)]. Of major importance was the observation of a complete disappearance of the low‐threshold finely tuned segment in regions of cochlea damage. In the vicinity of the probe, in contrast to the normal tuning curve, the effectiveness of the masker was sharply diminished. The “notch” in the tuning curve has now been confirmed with several listeners having various types of cochlear pathology. Follow‐up investigations of the frequency‐selective properties of sensorineural listeners were also carried out. In general, the results of experiments examining frequency discrimination and additivity‐of‐masking are in good agreement with the unusual frequency response revealed by tuning‐curve measurements. For example, in regions of threshold elevation, as much as a 10‐fold decrease in frequency discrimination was observed. Additivity‐of‐masking effects also revealed a close correspondence between musket effectiveness and region of cochlear damage. These findings are being used to build a profile of the auditory capacities of listeners with sensorineural hearing loss.
59(1976); http://dx.doi.org/10.1121/1.2002671View Description Hide Description
Unmasked and centrally masked thresholds were measured for normal listeners and for listeners with unilateral lesions involving the auditory temporal cortex. The signal was a 1000‐Hz pure tone with duration of 20 or 100 msec. The masker was a 60‐dB SPL, 1000‐Hz pure tone and was either continuous or gated with a 200‐reset duration. Signal and masker were presented to different ears. Both signal and masker were presented via hearing aid receivers to maximize interaural attenuation. In the gated masker condition, delay between masker and signal onset ranged from 20 to 150 msec. Data from the normal listeners agreed well with those from previous studies. Masked thresholds for the cortical‐lesioned subjects were much higher than for the normal listeners. Generally, maximum threshold shift was observed when the masker was presented to the ear ipsilateral to the lesioned hemisphere. When the masker was presented contralateral to the lesioned hemisphere, performance was normal. One of the cortical‐lesioned subjects had greater than normal threshold shifts for both ears. As signal onset delay increased, masker effectiveness decreased for all subjects.
59(1976); http://dx.doi.org/10.1121/1.2002672View Description Hide Description
The masking of silent intervals, or gaps, by surrounding noise bursts was investigated. The durations (T1 and T2) and the spectrum levels (L1 and L2) of the two noises were varied and the just‐detectable gap, ΔT, determined. The paradigm incorporated a replication of Plomp's study [J. Acoust. Soc. Am. 36, 277–282 (1964)]. There were three primary findings. First, by randomizing T1 with T2 fixed and then by randomizing T2 with T1 fixed, it was determined that, if L2<<L1, the difference in the perceived duration of the second noise burst could provide a detection cue. Second, for T1 = 200 msec, changes in T2 did not greatly affect ΔT, suggesting a close relationship between forward masking and gap detection. Third, the slope of the line relating L2 (in dB) to log ΔT was determined as L1 increased with T1 fixed and as T1 increased with L1 fixed. The data indicate that the slope (which reflects the decay of auditory sensation) is controlled by L1 (the spectrum level of the first noise) rather than by T1 (which determines the sensation level of the first noise). It is suggested that the temporal averaging mechanism itself, as well as its output, depends upon the spectrum level of the input. [Work supported by NIH.]
59(1976); http://dx.doi.org/10.1121/1.2002673View Description Hide Description
Plomp [J. Acoust. Soc. Am. 36, 277–282 (1964)], in explaining results on the detection of gaps in noise bursts, suggested that auditory sensation decays as a power function of time. According to Plomp's interpretation, the sensations resulting from more intense stimuli have faster rates of decay than sensations resulting from less intense stimuli, in such a way that all sensations reach threshold at the same moment after stimulus offset, independent of intensity. This suggestion is made quantitative and the resulting model is applied to Homp's data and to new data exploring the effects of intensity on reaction time to offsets of noise bursts. [Work supported by NSF and NIH.]
59(1976); http://dx.doi.org/10.1121/1.2002674View Description Hide Description
Previous research has shown that subjects can respond to the onsets of sinusoidal signals more rapidly than to the offsets of such signals. The difference in the speed of response to onsets and offsets of monaurally presented signals was investigated for three different signal frequencies: 500, 1000, and 4000 Hz. For the two higher signal frequencies it was found that this difference in reaction time to onsets and offsets was greater for signals presented to the right ear than for signals presented to the left eat. This eat difference resulted mainly from slower responses to the offsets of signals presented to the right ear than to the offsets of signals presented to the left ear. This ear difference is unusual in that it was obtained with monaural presentations of simple stimuli: the ear differences most frequently reported are typically found with dichotic presentations of complex (speech) stimuli. Previously reported ear differences have been attributed to differences in hemispheric organization of the brain. The basis of the present results is unclear.
59(1976); http://dx.doi.org/10.1121/1.2002675View Description Hide Description
The relationship between pulsation threshold curves and temporary threshold shift (TTS) was studied in normal hearing listeners following exposure to a 105‐dB SPL octave band of noise centered at 2000 Hz. Pulsation thresholds were obtained for a 10‐dB SL and a 20‐dB SL 3000‐Hz probe at six masker frequencies (1500, 2250, 2700, 3000, 3300, and 3750 Hz). For each probe sensation level, plots of masker intensity at pulsation threshold comprised the baseline pulsation threshold curves. These pre‐ and postexposure curves indicated an elevation of threshold and a degradation of frequency selectivity during TTS. While recovery from TTX followed the expected logarithmic pattern, pulsation thresholds remained elevated well beyond the recovery interval for TTS. Our interpretation of these findings supports the analogy between psychoacoustic and physiological measures of frequency selectivity. [Work supported in part by a training grant awarded to the first author by the Veteran's Administration Hospital, Indianapolis, IN.]
59(1976); http://dx.doi.org/10.1121/1.2002676View Description Hide Description
Two‐tone masking of a 1000‐Hz narrow‐band noise by high‐level masking tones (60–100 dB SPL) yielded notches in the resulting two‐tone masking functions. As orginally demonstrated by Zwicker [Acustica 4, 415–420 (1954)], those notches occurred at large frequency separations between masking tones. In a search for an adequate explanation of those notches, reversed masking functions were obtained for the 1000‐Hz narrow‐hand noise masked by constant‐intensity single‐tone maskers at frequencies above the 1000‐Hz narrow‐band noise and at frequencies below the 1000‐Hz narrow‐band noise. In further experiments, selective off‐frequency masking was employed at frequency regions of suspected distortion components. The results indicated (1) that the notch obtained in two‐tone masking functions is due primarily to the lower‐frequency tone of the two‐tone masking complex; (2) that the frequency separation at which the notch occurs is dependent upon the level of the masking tones; and (3) that the notch can be eliminated with selective off‐frequency masking. Alternative explanations for the notch are discussed which include the detection of off‐frequency distortion components and/or limitations in the temporal resolving power of the auditory system. [Work supported in part by NIH grants NS‐10561, NS‐12125, and NS‐04105].
59(1976); http://dx.doi.org/10.1121/1.2002732View Description Hide Description
The threshold of a pulsed sinusoid centered in a notched‐noise masker was measured as a function of the bandwidth of the notch. Assuming that the amount of masking reflects the amount of noise passing through an auditory filter centered on the signal frequency, the derivative of a curve fitted to the data is an estimate of the auditory filter shape [R. D. Patterson, “Auditory filter shape,” J. Acoust. Soc. Am. 55, 802–809 (1974)]. Estimated of the bandwidth of the auditory filter, defined as the half‐power bandwidth, were found for three frequencies (1000, 2000, and 4000 Hz) and five noise spectrum levels (10, 20, 30, 40, and 50 dB). The bandwidth of the auditory filter increases with masker intensity above 30‐dB noise spectrum level. At these higher intensities, an interaction with signal frequency is also observed, the bandwidth being relatively greater at higher frequencies. [This research was supported by the NIH.]
59(1976); http://dx.doi.org/10.1121/1.2002733View Description Hide Description
Previously, we found masking by a remote portion of vowels on discrimination of a transition in the second formant, particularly at sound levels required by listeners with severe hearing loss. Would such masking also extend from a vowel to the noise‐burst cue of a consonant? We measured threshold shifts for the detection of a brief high‐frequency noise burst that preceded or followed a two‐formant synthetic vowel. The listeners had normal hearing or long‐term, moderate‐to‐severe sensorineural hearing losses. Threshold shift was the difference between the burst threshold for the noise burst alone and threshold when the burst was preceded or followed by the vowel. The vowel level was at the MCL of each listener and also at 95 and 105 dB for the normal‐hearing listeners. For both groups of listeners there was little or no threshold shift of the noise burst when the vowel was heard at MCL. For the normal‐hearing listeners, greater masking of the noise burst occurred when the vowel was presented at 95 and 105 dB, particularly when the noise burst followed the vowel.
- Session B. Speech Communication I: Stress and Timing Masking and Dichotic Listening
59(1976); http://dx.doi.org/10.1121/1.2002734View Description Hide Description
Traditional theories of phonology maintain that nominal compounds have a [1/3] stress pattern whereas nominal phrases are characterized by [2/1] stress [N. Chomsky and M. Halle, The Sound Pattern of English (Harper and Row, New York, 1968)]. However, there does not appear to be a sharp boundary between compounds and noun phrases, either in speaker's strategies for indicating stress, or in their perception of it. Indeed, there appears to be a continuum, where certain word pairs like “apple core” are definitely perceived and produced as compounds, others such as “boiling water” are usually noun phrases, and numerous others like “parcel post” or “box lunch” can be either. Durational and fundamental frequency data will be presented for three naive subjects who recorded 40 word pairs in sentence context. Additional subjects were given a written list of these sentences with the word pairs underlined and asked to circle the one syllable of each pair that should be most stressed. Certain classes of word pairs received more consistent stress assignments than others, but definite rules predicting stress seem to be difficult to formulate.
59(1976); http://dx.doi.org/10.1121/1.2002735View Description Hide Description
The perceptual effects of variations in the fundamental frequency contour were studies in some detail in identification of a disyllabic word as a noun or a verb. The F 0 contour of the first syllable had one of three slope values but had the same average frequency. Within the second syllable, both the slope (four values) and the average frequency (at least eight values) were varied. The verb “subject” was taken from a naturally spoken sentence which was analyzed and resynthesized with modified pitch contours by use of the LPC technique. Results show that the shift in categorization is relatively gradual with respect to small gradual changes (step size of 2 Hz) in average F 0. The boundary is dependent on the slope of both the first and the second syllable. These identification characteristics will be discussed in relation with the previously reported results of discrimination characteristics [H. W. Campbell, J. Acoust. Soc. Am. 58, S57(A), S92(A) (1975)].
59(1976); http://dx.doi.org/10.1121/1.2002736View Description Hide Description
People can usually distinguish the word juncture in contrastive twins of word pairs such as play nice and plain ice. We can see many features which differ between the twins in spectrograms, but observation alone cannot tell us which features ate important to the perception of juncture. The contributions of pitch, amplitude, timing, and segmental features to juncture perception were studied as follows. Parametric representations of twins X and Y were obtained by analyzing natural speech, and the above features were exchanged between the twins in a complete factorial design. A feature f was considered important for juncture perception if exchanging f between X and Y also caused perception to be exchanged; that is, listeners heard Y when presented with X possessing feature f of Y, and vice versa. Listeners' responses indicated that all the features were important, but the importance of any particular feature or set of features depended on the juncture consonant. Features also interacted prominently for some consonants
59(1976); http://dx.doi.org/10.1121/1.2002785View Description Hide Description
A theory is proposed to account for semantic and syntactic effects on syllable duration. The theory is based in part on experimental studies by the author in which a sentence‐reading task was used to determine effects of semantic and syntactic variables, quasi‐independent of phonetic factors. The results of the studies motivated constraints on the operation of two processing levels, representing semantic and syntactic modes. The distinction between the two levels was prompted by an empirical difference in the domain over which timing effects operated. Syntactic effects were restricted to the domain of a single clause, whereas semantic effects were not. Predictions of the theory are discussed within the framework of a programmatic line of research. [Work supported by an NSF GraduateFellowship, NIH Grant HD‐05168, and grant from the Slosh Foundation.]
59(1976); http://dx.doi.org/10.1121/1.2002786View Description Hide Description
In rapid speech,acoustic analysis reveals that steady state vowel targets characteristically are not reached. Lindblom and Studdert‐Kennedy [J. Acoust. Soc. Am. 42, 830–843 (1967)] found in an experiment with synthetic speech that listeners showed a shift in the boundary between medial vowels /ε/ and /U/ with variations in the rate and direction of formant transitions. Apparently, perceivers compensate for simulated articulatory undershoot by perceptual overshoot. An experiment with natural speech demonstrated shifts in the acoustic criteria listeners employed in vowel recognition as a function of perceived rate of utterance. Nine American English vowels in /p‐p/ environment were produced by a panel of 15 talkers in a fixed sentence frame. The unstressed, rapidly articulated /p‐p/ syllables were excised from the tape recording and assembled into listening tests. In one condition, the /p‐p/ targets were prefaced by three slowly articulated syllables /hi, ha, hu/. In the other test condition, no precursors were present. Misperception of target vowels occurred with significantly greater frequency when they were preceded by precursors. The results showed evidence of perceivers' adjustment to tempo (as hypothesized by Lindblom and Studdert‐Kennedy) on the basis of preceding segments of speech. [Work supported by NIH.]
59(1976); http://dx.doi.org/10.1121/1.2002787View Description Hide Description
In experiments measuring reaction time (RT) to phoneme targets in sentences, earlier work produced slower RT when the target was temporally displaced by experimental intervention (computer editing of speech waveform), e.g., artificially extending by 100 msec a pre‐stop‐consonant silent interval two syllables earlier in the sentence. The result was interpreted in terms of disrupted timing expectancies based on distortion of the prosodic structure, i.e., an interaction between segmental and prosodic cue effects during ongoing perception. In further work, targets in six‐syllable nonsense sequences were temporally displaced by lengthening or shortening one vowel in the sequence by various amounts (50–130 msec). The perceptual impression when noticeable is of a sudden shift in speechtempo. RT was measured to phoneme targets in over 600 sequences during several hours of listening. Despite practice and potential cue effects of pre‐target manipulations, timing distortions within the first or second pretarget syllable slowed target RT. [Work supported by NIMH, ARIBSS.]