Signal-Detection Analysis of Equalization and Cancellation Model

An analysis of the equalization and cancellation model from a signal-detection standpoint is presented. At low frequencies, say 250 cps, one can neglect the time-error parameter because any slight phase error is small as compared with the period of a sinusoid and all the error in the binaural mechanism can be treated as small-amplitude error. Using this assumption, it is possible to derive the equations presented in the text. Three main conclusions can be drawn from this analysis. (1) The signal-detection analysis generates the same equations as Durlach's signal-to-noise-ratio approach, except for a multiplicative factor that reflects the accuracy of the monaural energy measurement. (2) The parameter associated with the amplitude error, because of the multiplicative factor, is different from Durlach's by a factor of 5. (3) The new analysis predicts that shortening the signal duration or decreasing the noise bandwidth will increase the size of the masking-level difference. Both effects have been observed.