The masking of silent intervals, or gaps, by surrounding noise bursts was investigated. The durations and the spectrum levels of the two noises were varied and the just detectable gap A T determined. The first experiment demonstrated the conditions in which duration judgments may have provided a cue. The remaining experiments were therefore designed so that this duration cue was not available. The second experiment indicated that changing the second noise duration from 200 to 2 msec did not greatly affect A T, suggesting a close relationship between forward masking and gap detection. Finally, Plomp [J. Acoust. . 36, 277-282 (1964)] has argued that a graph of the level of the second noise in decibels as a function of log iX T traces the time course of decay of the first noise. If so, the data from the third experiment indicate that this decay rate depends on the spectrum level of the noise but not on its duration. Soc. Am

This research investigates the human observer's ability to discriminate between the durations of two silent intervals, each interval preceded and followed by noise bursts called markers.The markers are separated· by T or T +~T' msec and T ranges from 0.3 to 1,000 msec.T is defined as that value of~T' for which the probability of discriminating T from T + AT' is 0.75. We compared the value of AT for conditions in which the markers were fixed in amplitude and duration with conditions in which the marker amplitudes and durations were randomly chosen.~T increased by as much as a factor of 4 when the amplitude and duration of the markers were randomized. The performance decrement was primarily due to randomizing the first marker duration.Models of duration discrimination relate discriminability to the duration of the stimuli presented. Stimuli that differ in duration, however, differ in other ways as well, and the idea that it is only their temporal extent which provides a discrimination cue is open to question. In a recent paper, Abel (1972) investigated the human observer's ability to discriminate between the durations of two silent intervals, each interval preceded and followed by noise bursts called markers. This paradigm is an important one, because the silent intervals differ in temporal extent only. There are no cues related to the total energy or to the energy spectrum of the stimuli. It is therefore of particular interest whether or not discriminability is independent of marker amplitude and duration in this paradigm.In Abel's paradigm, the marker amplitude and duration were fixed from trial to trial and varied between blocks. For these fixed markers, discrimination was affected by marker amplitude indicating that some nontemporal cue might be used (Allan & Kristofferson, 1974). In the present experiment, we investigate an extreme case of marker variability. In our work, the parameters of each of the four markers bounding the two intervals are chosen randomly and independently. For large marker variability, we reopen the question of whether it is only the physical duration of the silent temporal interval which provides the discrimination cue. METHODS SubjectsTwo undergraduates at Arizona State University were paid to participate. The author was also a subject. Two of the three This research was supported in part by a grant from the National Institutes of Health and in part by a faculty research award from the City University of New York. subjects were practiced listeners with at least 3 years of experience in psychoacoustic experiments. ApparatusThe durations of the noises and the gap between them were controlled by a PDP-IS computer. The noise was generated by a random noise generator (General Radio Model 1383), gated, filtered (0 to 20 kHz; Krohnhite Model 3SS0R), and then served as input to TDH-39 headphones. Two observers were run simultaneously in two double-walled lAC chambers. ProeedureThe data were collected using a two-interval forced-choice procedure. A trial began with a flash of a warning light...

For patients with noise-induce sensorineural hearing loss, the results of matching a binaurally presented comparison tone to subjective tinnitus during a 20-days test period are reported. As a control, results of matching an external comparison tom, to a standard tone. are also presented. The variability for tinnitus measurements was extremely large relative to comparable measures for a objective stimuli The relevance of this finding to the nature of tinnitus and to the construction of tinnitus maskers is discussed.

No abstract

Recent work on temporal masking and temporal resolution indicates that acoustic stimuli are combined to a degree determined by their temporal proximity [D. Ronken, J. Acoust. Soc. Am. 47, 1091-1099 (1970); J. H. Patterson, Jr., Ph.D thesis (1970) (unpublished); M. J. Penner, Percept. Psychophys. 18, 114-120 (1975)]. For broadband stimuli, the process of combining inputs has been modeled as resulting from an exponential integrator with a time constant of 2-4 ms. Such models seem to suggest that stimuli more than, say, 20 ms apart are treated largely as separate inputs. In contrast, the classical view, based on the results from time-intensity trades, holds that broadband noise stimuli are integrated together for 100-200 ms [W. R. Garner, J. Acoust. Soc. Am. 19, 808-815 (1947)]. This difference in the period over which stimuli are integrated therefore varies by about two orders of magnitude according to research in the two areas. It may be that the difference in the magnitude of the time constants reflects the plasticity of the auditory system. However, in this paper we show one way to eliminate the apparent discrepancy by demonstrating that there exists a class of short-term integrators, operating on some fractional power of the stimulus intensity, that can produce the time-intensity trade data. PACS numbers: 43.66.Dc, 43.66.Mk, 43.66.Ba INTRODUCTION A series of studies have been reported indicating that acoustic stimuli are combined if they occur near each other in time (Ronken, 1970; Patterson, 1970; Penner, 1975)o In each of these reports, the stimulus' spectrum is relatively broadband and flat. For example, Penner (1975) explored the detectability of a low-pass filtered click embedded in the temporal center of a noise burst as the duration of the burst varied. She showed that only the noise immediately surrounding the brief click masked it and that this masking could be modeled as resulting from an exponential integrator which combined the signal and noise. The time constant of the integrator was 3.7 ms so that 99% of the noise which was effective in masking the click occurred within 15 ms of it and the time constant seemed independent of the noise level. Other investigators have also proposed an exponential integrator with a 2-4-ms time constant (Duifhuis, 1973; Siebert, 1968). Such brief time constants are in striking contrast with the traditional view that power is accumulated for 100-200 ms (Garner and M•ller, 1947; Green, Birdsall, and Tanner, 1957; Green, 1960). The principal technique for studying long-term accumulation in detection experiments is to determine combinations of signal intensity I s and signal duration Ts, that result in constant detectability.For a considerable range of signal duration, the signal intensity may be decreased as its duration increases. This phenomenon is usually described by saying that intensity and time can be "traded" to preserve constant detectability. In particular, the following summary statement is often made: log/s decreases in a linear fashion with log Ts, implying ...

The detection of a silent interval, or gap, placed in the temporal center of a gated noise burst was investigated. The gated noise masker ranged from 2 to 400 msec in duration. For long noises, the duration.a. of the just-detectable gap remained fixed at about 2.8 msec. Progressively shortening the duration of the noise did not affect A until the duration was approximately 20 msec; thereafter. decreasing the noise duration improved detectability of the gap. In a second experiment, continuous noise filled the temporal gap. although the decibel difference between the noise in the gap and the noise surrounding the gap was always at least 5 dB. The level of noise filling the gap did not greatly affect a. The third experiment was similar to the first, except that the signal was a click rather than a gap. The results for both gaps and clicks were fitted by a model assuming a sliding integrator.

No abstract

A model of temporal summation and intensity coding relates the subject's internal percept y(t) to the stimulus input x(t) by the equation y(t) = g(St - oof[x(tau)] h [t, tau, x (tau)]d tau). In words, some transformation f[x (t)] of the stimulus intensity is weighted by a function h and integrated; the result is transformed into the internal percept by a function g. This system postulates a linear integral operator preceded and followed by transformations which may be nonlinear. Based on forward masking of clicks by white noise, we (1) show that the above characterization of the model is appropriate (which involves showing that there is a linear temporal summation stage), and (2) derive certain characteristics of the system's nonlinearities. In particular, the integral of h times f is shown to be a nonlinear function of the input intensity exhibiting more compression than a power function. It is also shown that h must depend upon the intensity of the stimulus.