Ripple density resolution assessments in discrimination tasks with rippled and non-rippled reference signals

Abstract

Rippled-spectrum signals (RSS) are used to evaluate the resolution of the spectro-temporal structures of sounds. Resolution measurements are based on tasks that discriminate between test and reference signals. Therefore, estimates of rippled-pattern resolution may depend on both the test and reference signal types. In the present study, ripple density resolution was measured using different reference signals. The test signals were RSS with several ripple phase reversals during the signal, and the reference signals were either RSS with a constant ripple phase or were non-rippled. The spectra were centered at 2 kHz and had an ERB of 1 oct and an SPL of 70 dB. A three-alternative forced-choice procedure was combined with an adaptive stimulation procedure. With rippled reference signals, the mean ripple density resolution was 8.9 ripples/oct. With non-rippled reference signals, the mean resolution was 26.1 ripples/oct. Different contributions of excitation-pattern and temporal-processing mechanisms were hypothesized for resolution estimates obtained with rippled and non-rippled reference signals: the excitation-pattern mechanism was expected to be more effective for discriminating between rippled signals that differ in their ripple-phase patterns, whereas the temporal-processing mechanism was expected to be more effective for discriminating between rippled and non-rippled signals.