Effects on Detectability of Frequency Variation During an Observation Interval

In many cases, observers are highly sensitive to variations in a stimulus parameter. In order to ascertain whether this sensitivity to change would increase the detectability of such a stimulus, noise-masked detection was measured for equal-duration, equal-intensity signals, some of which were held at a constant frequency and some of which were varied in frequency during the observation interval. The continuous signals were gated sinusoids of 250, 450, 950, 1750, 2550, and 3550 Hz; the frequency-varying signals were linear FM slides centered at the frequencies of the fixed tones and with a bandwidth of 100 Hz. A second category of frequency-varying signals included “housetop” FM slides that covered the 100-Hz bandwidth in half the duration of the first slides and then linearly returned to the starting frequency in the second half of the signal duration. Thus, they exhibited the same frequency, bandwidth, and intensity, but twice the rate of change in frequency. The noise was wide-band, Gaussian. The single frequency signals were more detectable than the linear FM slides, which were more detectable than the housetop FM slides.