Auditory detection of hollowness.

The airborne sounds produced by freely vibrating hollow and solid bars were synthesized according to the equations of bar motion from theoretical acoustics, and were presented to listeners over headphones. In a two-interval, forced-choice task, listeners were asked to distinguish between the hollow and solid bar sounds as bar length was varied at random from one presentation to the next. All other physical properties of the bar were held constant across trials. Listener decision strategies for detecting hollowness in iron, aluminum, and wood bars were determined from regression weights describing the relation between the listener's response and the frequency, intensity, and decay modulus of the individual partials comprising these sounds. The obtained weights were compared to those of a hypothetical listener that bases judgments on the acoustic relations intrinsic to hollowness, as determined from the equations for motion. Results indicate that listeners adopt roughly one of two decision strategies, either basing judgments on the appropriate acoustic relations, or basing judgments predominantly on frequency alone. The decision strategy of some listeners also changed from one type to the other with a change in bar material or upon replication of the same condition. The results are interpreted in terms of the vulnerability of the intrinsic acoustic relations to small perturbations in acoustic parameters, as would be associated with listener internal noise. They demonstrate that basic limits of human sensitivity can have a profound effect on the identification of rudimentary source attributes from sound, even in conditions where acoustic variation is largely dictated by physical variation in the source.