Detection of spectrally complex signals in comodulated maskers: effect of temporal fringe.

This study tested the hypothesis that masking release for a complex signal under conditions where signal energy is present in all frequency regions occupied by the masker is attributable to an across-frequency-channel comodulation masking release (CMR) process. The approach was to identify a signature CMR trait, and to then determine if that trait was associated with the detection advantage for complex signals. The selected trait was the decline of CMR in the presence of a random temporal fringe. In experiment 1, a masking release was observed for a four-component harmonic signal presented in a comodulated masker, and this masking release was diminished by the random temporal fringe. A similar effect was observed in experiment 2 for a four-component inharmonic signal. These results support the hypothesis that a CMR can be measured for a complex signal even when there is substantial spectral overlap between the signal and its comodulated masker. This finding has consequences for CMR models since it demonstrates that the presence of "signal-free" cue bands is not a prerequisite for CMR, and that the presence of comodulation during the signal window is not sufficient to result in CMR.