ABSTRACT

We describe several experiments involving the detection of a change in the spectral shape of a complex auditory signal, what we call profile-analysis. All of the experiments are discrimination tasks involving a broadband “standard” spectrum and some alteration of that spectrum produced by adding a “signal” to the standard. For all of the experiments described here, we used a standard composed of a set of equal-amplitude sinusoidal components. The spectrum of the standard was, therefore, essentially flat. In different experiments, various waveforms were added to this standard to create changes in its spectral shape, and the ability to detect such changes was measured. In the first experiments, we describe how the relative phase among the components of the standard waveform influences the detection of a signal. The results are very simple. Phase seems to play no important role. The detection of a change in spectral shape appears to depend only on changes in the power spectrum of the signal and is independent of the temporal waveform. Next, we describe how the detection of an increment in a single component depends on the frequency of that component. These results provide the basic data to evaluate complex changes in the whole spectrum, such as a sinusoidal ripple in the amplitudes of the components over the entire spectrum. Our data indicate that there is a sizable discrepancy between the ability to detect changes occurring over the entire spectrum and the ability to detect changes in single components.