Fluctuation strength

Established auditory attributes, e.g. loudness, roughness, or fluctuation strength, figure prominently in commercially available sound-evaluation software. The computation of some of them, however, is based only on scarce empirical findings. This is particularly true for the attribute of 'fluctuation strength,' which supposedly depends on both the modulation rate, and the depth of modulation in frequency (FM) or amplitude (AM). A major flaw of earlier studies is that fluctuation strenght has been investigated only while varying one physical parameter at a time. Thus, a series of experiments was conceptualized in an effort to re-evaluate this psychoacoustical attribute, and to empirically test its presumed uni-dimensionality. In experiment I, replicating the procedure of the original investigators, either modulation rate, or depth of modulation were varied one at a time in sinusoidal FM. In experiment II, to investigate the potential interaction of the variables, both dimensions were varied in a factorial design. Participants were asked to give numerical estimates of the magnitude of the fluctuation strength they perceived. The results of both experiments showed large interindividual variations, and widely differing response patterns, that cast doubt on the concept of fluctuation strength being an elementary auditory percept. While largely agreeing on the effect of modulation depth, subjects judged the effects of the modulation rate in idiosyncratic ways. To investigate, whether the experimental task played a role in obtaining this outcome, or whether it must indeed be concluded that the fluctuation strength of stimuli cannot be considered a psychophysically valid, uni-dimensional attribute, a third experiment was designed. Here, instead of requiring direct numerical estimates, subjects were simply asked to decide which of two stimuli was the more fluctuating. An 'additive-conjoint measurement' analysis revealed that both physical dimensions do not additively contribute to a uni-dimensional percept for most of the subjects. (Center contract)