The perception of waning signals: decruitment in loudness and perceived size.

When a tone or broad-band noise sweeps smoothly from a moderate intensity to a low one, the loudness at the end of the sweep is far less than what would be predicted from its intensity. The accelerated reduction in loudness, which was first reported by Canévet (1986) and confirmed in several later reports, has been called loudness decruitment, and has been tentatively interpreted as the result of some form of adaptation. Since both simple and induced adaptation have distinctive temporal profiles, we undertook a series of studies in which we varied the duration of a tone whose intensity was continuously changing, to see whether the effect of duration on decruitment resembled the effects of duration on adaptation. We discovered that the magnitude of decruitment remained unaffected when the duration of the sweep was reduced far below the durations of 90 to 180 sec that have been used in previous studies. The same effect was observed for durations of around 20 sec, but it declined rapidly to a low level at the lowest duration of 1.0 sec. This temporal pattern is strikingly different from what has been reported for either simple or ipsilaterally induced adaptation, which suggests that neither form of adaptation can account for the entire effect. We also wanted to know whether an analogous phenomenon could exist for a sensory modality other than hearing. In the present study, observers were asked to judge the apparent size of a solid disk on a computer monitor, the disk increased or decreased continuously in area, or appeared as a series of separate areas, either in random order or in ordered progressions. We found that, as in the case of loudness, apparent size decreased more rapidly when the areas decreased continuously than would have been predicted from the actual areas themselves. We also found that some part of the accelerated shrinkage was due to a response bias in the observers' judgments that stemmed from knowledge that every value in a continuously changing series is predictably smaller (or larger, for a growing series). Whether the remaining part of the effect is a sensory phenomenon is an important issue for future research.