Description:
<jats:p>Recently, a new approach to modeling binaural interaction was described [Breebaart et al., J. Acoust. Soc. Am. 103, 2844–2845 (1998)]. This model is based on a subtractive mechanism and is sensitive to interaural time differences as well as interaural intensity differences. In order to extend this model to masking data for partially correlated noise maskers, the following experimental findings have to be considered. (a) For broadband maskers with correlations &lt; 0.95, the power ratio between signal and masker difference intensity (i.e., masker intensity at the output of the subtractive mechanism) is constant. (b) In broadband conditions, frozen- and running-noise maskers lead to the same thresholds, indicating that masker correlation uncertainty does not influence detectability of the signal. (c) For narrow-band running-noise maskers, the masker correlation uncertainty is the dominant factor to explain signal thresholds. Result (a) can be implemented by applying a logarithmic compression to the output of the substractive mechanism and adding an internal noise with a constant rms value. With this modification, the model is capable of describing the differences between frozen- and running-noise maskers as listed under (b) and (c). [Work supported by the Dutch Organisation for Scientific Research (NWO).]</jats:p>