Models of auditory masking: A molecular psychophysical approach

Abstract

Gilkey et al. [J. Acoust. Soc. Am. 78, 1207-1219 (1985)] measured hit proportions and false alarm proportions for detecting a 500-Hz tone at each of four starting phase angles in each of 25 reproducible noise samples. In the present paper, their results are modeled by fitting the general form of the electrical analog model of Jeffress [J. Acoust. Soc. Am. 48, 480-488 (1967)] to the diotic data. The best-fitting configurations of this model do not correspond to energy detectors or to envelope detectors. A detector composed of a 50-Hz-wide single-tuned filter, followed by a half-wave rectifier and an integrator with an integration time of 100 to 200 ms fits the data of all four subjects relatively well. Linear combinations of the outputs of several detectors that differ in center frequency or integration window provide even better fits to the data. These linear combinations assign negative weights to some frequencies or to some time intervals, suggesting that a subject's decision is based on a comparison of information in different spectral or temporal portions of the stimulus.

An evaluation of models for diotic and dichotic detection in reproducible noises

“…3͒. The negative weights for the MD model found above and below the target frequency were consistent with weighting patterns in previous MD model results ͑Ahumada and Lovell, 1971;Ahumada et al, 1975;Gilkey and Robinson, 1986͒, but only positive weights were possible in the suboptimal weighting scheme because these weights were derived from rms metrics. Note that the weights that fit to the responses of S3 for MD with the 100-Hz bandwidth were close to those fitted to the MDS model.…”

“…The first two are related: single critical-band ͑CB͒ models and multiple-detector ͑MD͒ models that linearly combine outputs of multiple critical bands. To date, detection patterns estimated under diotic conditions have been best predicted by a MD model ͑Ahumada and Lovell, 1971; Ahumada et al, 1975;Gilkey and Robinson, 1986;Davidson et al, 2006͒. The MD model accounted for up to 90% of the variance in one subject's responses in Ahumada and Lovell ͑1971͒ and up to 72% of the variance in one subject's responses in Gilkey and Robinson ͑1986͒.…”

“…Such data present a more rigorous test for models of masked detection because, in addition to predicting average threshold, the models must predict detection statistics for individual waveforms. As shown here and in other works, models that accurately predict average thresholds may fail to predict responses to individual waveforms ͑e.g., Isabelle, 1995;Isabelle and Colburn, 2004͒. The models tested in this study were selected because they have successfully predicted reproducible noise data in the past ͑e.g., Fletcher, 1940;Ahumada and Lovell, 1971;Ahumada et al, 1975;Gilkey and Robinson, 1986͒, because they have been used with some success to predict thresholds for a broad spectrum of psychophysical detection tasks ͑e.g., Dau et al, 1996aDau et al, , 1996bBreebaart et al, 2001aBreebaart et al, , 2001b, because they are straightforward adaptations of ob-served physiological phenomena ͑e.g., McAlpine et al, 2001;Marquardt and McAlpine, 2001͒, or because they use a processing strategy that involves an interaction between stimulus envelope and fine structure ͑e.g., Goupell and Hartmann, 2007͒. The interaction of envelope and fine structure was of particular interest because such an interaction has been suggested by recent empirical studies of detection of low-frequency tones in reproducible maskers ͑Davidson, 2007; Davidson et al, 2009͒.…”

The temporal weighting of loudness: effects of the level profile

“…Thus, molecular, rather than molar, analyses were conducted on the data (Green, 1964) in order to gain insight into the decision process, rather than obtaining molar measures such as loudness or accuracy. Molecular, or perceptual weight, analyses have been used for several decades (Ahumada & Lovell, 1971;Berg, 1989;de Boer & Kuyper, 1968;Gilkey & Robinson, 1986) and have found increasing application in several domains (e.g., Ahumada, 2002;Berg, 2004;Neri, Parker, & Blakemore, 1999;Oberfeld, 2009;Yu & Young, 2000). We applied this technique to a loudness judgment task, in order to estimate the influence of the sound pressure level of different temporal portions of the stimulus on global loudness.…”

Binaural detection with narrowband and wideband reproducible noise maskers. IV. Models using interaural time, level, and envelope differences

“…In order to test model predictions and compare the effectiveness of different cues, it is useful to consider detection performance on a waveformby-waveform basis (molecular-level) for each listener (e.g., Sch€ onfelder and Wichmann, 2013). Gilkey et al (1985) and Gilkey and Robinson (1986) found that averaging detection performance across masker waveforms obscures the differences across individual waveforms and listeners, suggesting the utility of a more molecular-level approach. However, molecular-level predictions are difficult to obtain because of the unknown internal noise for each listener and the possible use of different cues by different listeners.…”