Further evidence against an across-frequency mechanism specific to the detection of frequency modulation (FM) incoherence between resolved frequency components.

Previously, Carlyon [J. Acoust. Soc. Am. 89, 329-340 (1991)] argued that there exists no across-frequency mechanism specific to the detection of frequency modulation (FM) incoherence, defined as a difference in FM phase, between pairs of resolved frequency components. Experiments are described which attempted to reconcile this conclusion with the results of two recent studies. Wilson et al. [J. Acoust. Soc. Am. 88, 1333-1338 (1990)] reported that the detection of FM imposed on a target component was impaired by the presence of an FM "interfering tone," and that the size of the effect depended on the FM coherence between interferer and target. Experiment 1 replicated their findings but showed, by using low-pass and wideband noise, that its dependence on FM coherence was consistent with the detection of combination tones and of beating between the interferer and target. Cohen and Chen [J. Acoust. Soc. Am. 92, 766-722 (1992)] reported that the detection threshold for an FM signal was higher when it was masked by two components modulated coherently with it than when the masker and signal modulations were incoherent. Experiment 2 showed that, when the masker consisted of eight components, thresholds were largely determined by the modulation coherence between the signal and the masker component closest to it in frequency. Experiment 3 presented evidence that Cohen and Chen's findings were influenced by harmonicity between the masker and signal, even in conditions which attempted to control it. Experiment 4 replicated another of Cohen and Chen's findings, that when the masker modulation was held constant and the signal FM depth varied, some listeners' thresholds reached a maximum when the masker and signal FM depths were equal. By manipulating the frequency ratio between masker and signal, it was shown that this finding, too, could be attributed to harmonicity between the signal and one of the masker components. Finally, experiment 5 replicated Carlyon's (1991) findings at a higher sensation level and with a different pattern of modulation than used previously.