Masking-level differences and the form of the psychometric function

Abstract

The detectability of an auditory signal that is presented in noise may be increased markedly by an appropriate change in the interaural relations of the signal or of the noise. The effects are so large that the enhanced detectability , or "release from masking," is typically measured, not as an increase in the detectability of the signal, but rather, as the change in the signal energy that is necessary in order to maintain the listener's performance at a constant level. This change in signal energy, expressed in decibels, is called a "masking-level difference," or MLD. The investigations of Licklider (1948), Hirsh (1948), and Hirsh and Webster (1949), mark the beginnings of this line of auditory research, and since the publication of these early studies, there has been a steady increase in our knowledge of binaural masking phenomena. In particular, Jeffress and his coworkers (e.g., Jeffress et al, 1952Jeffress et al, , 1956Blodgett et al, 1958;Robinson & Jeffress, 1963; Langford & Jeffress, 1964) have made many significant contributions to this area of audition. Recently, Durlach (1963), Green (I 966a), and others have developed theoretical approaches to these phenomena different from that developed by Jeffress, and these various models of binaural masking have contributed to an increased interest in MLDs.As an example of a masking-level difference, suppose that the "masked absolute threshold" is determined for a sinusoidal signal when it and a continuous masking noise are presented to the right ear only, a condition called Nm-Sm. Next, the noise waveform that is presented to the right ear is presented simultaneously to the left ear, so that the instantaneous sound pressure of the noise in the left ear is the same as that in the right ear. With this perfectly correlated "binaural noise," the masked threshold for the monaural signal in the right ear is redetermined. It is typically found under this condition, termed NO-Sm, that the intensity of the sinusoid must be decreased by about 9 dB (at 300 cps) in order to reestablish the masked threshold. This difference in signal intensity of 9 dB is the masking-level difference, or MLD, of NO-Sm relative to Nm-Srn. Although the magnitude of the MLD is a function of the frequency of the signal, the effect is easily measured over the range from 200 to 1000 cps.It might be conjectured that the MLD obtained for NO-Sm is due simply to a facilitative, or synergistic, effect from the noise in the nonsignal ear. That this is not the case is evidenced by the fact that if the two noises are uncorrelated (NU-Sm), the detectability of the signal is the same as with the monotic condition.The symbols used in the current literature on masking-level differences are sometimes confusing and it may help the reader to understand this nomenclature if the symbols employed in this paper to describe the various interaural conditions are all defined at this time. The symbols used are: N for the masking noise, S for the sinusoidal signal, 0 for an interaural phase shift of zero degrees, tt for...

Forward masking in gerbils: The effect of age

“…Based on physiological data (Harris and Dallos, 1979;Yates et al, 1985) it is reasonable to assume that the masker used in the present study (400 ms, 40 dB SPL) reduces spontaneous and evoked neural activity in sensitive auditory nerve fibers tuned to the masker frequency after the end of the masker in the time window relevant for the present forward masking analysis. If this reduction of neural activity following the masker is associated with a reduced variability of internal noise, one might expect to see steeper slopes in the condition with the masker (Egan et al, 1969;Watson et al, 1972). However, our data showed no significant difference in the slopes of the psychometric functions obtained in the presence or absence of the masker.…”

“…Another reason for determining the slope of psychometric functions is related to the variability of internal and external noise. The reduced spontaneous and evoked activity at the level of auditory nerve fibers (Harris and Dallos, 1979) that follows the cessation of the masker might also affect the variability of internal noise and consequently the slopes of the psychometric functions (Egan et al, 1969;Watson et al, 1972).…”

A quantitative analysis of psychometric functions for different auditory tasks in gerbils

“…The slopes of psychometric functions are expected to depend on the variability of external and internal noise (Egan et al, 1969;Watson et al, 1972) that interferes with the detection of the signal. Penner (1986) showed that the slopes were more shallow for test-frequencies associated with tinnitus compared to test-frequencies distant from the tinnitus and tinnitus was suggested as a source of variable internal noise that caused more shallow slopes of psychometric functions.…”

A new psychoacoustical masking model for audio coding applications

“…CG• the distortion detectability is equal to the distortion-to-masker-power ratio. This is in line with the so called sensitivity index rf which is used in signal detection theory as a measure for the detectabi lity of a signal close to its masking threshold [8]. This ratio also implies that if the powers of the dis tonion and the masker are increased by the same amount (in dBs) that the detectability.…”