Effects of relative phase and frequency spacing on the detection of three-component amplitude modulation

These experiments explored the effect of relative modulator phase on the detection of a three-component modulator applied to a 4,000-Hz sinusoidal carrier with a level of 70 dB SPL. The central modulator component had a frequency of 50 Hz, and the two other components had frequencies of 50+/-5, 10, 25, 40, or 45 Hz. Thus, the modulator waveform was always periodic. Each modulator component had the same modulation index, m. The relative phases of the components were chosen to give a variety of modulation waveforms differing in the ratio of maximum to minimum value (max-min) and in crest factor. In experiment 1, modulation detection thresholds were measured by varying m, using an adaptive two-interval forced-choice procedure. Thresholds were found to be independent of relative modulator phase and of the frequency spacing of the components. In experiment 2, detectability (d') of the modulation was measured for several fixed values of m. Detectability was found to be independent of relative modulator phase and of the frequency spacing of the components. The results are not consistent with the idea that modulation detection thresholds are determined by the max-min value or crest factor of the envelope. The results are consistent with a model which assumes that the stimuli are subjected to a nonlinearity, and thresholds are determined by the root-mean-square value (or the mean square value) of the ac component of the envelope, following this nonlinearity. The nonlinearity may partly reflect compression on the basilar membrane, but other nonlinearities may be involved. This model can also explain some aspects of earlier results on the sensitivity to relative modulator phase [E. A. Strickland and N. F. Viemeister, J. Acoust. Soc. Am. 99, 3638-3646 (1996)].