The effect of the Negative Level Effect

Hartmann and Rakerd [JASA 94, 2083–2092 (1993)] measured listeners’ ability to identify the source location (front, overhead, or behind) of short (25-μs clicks) and long (880-ms noise bursts) broadband stimuli as a function of intensity. Error rates increased markedly with intensity for clicks—the Negative Level Effect (NLE)—but not for noise bursts. In addition to direct practical applications (e.g., stimulus choice in auditory displays), the result of this simple experiment has inspired a sub-field of its own including psychophysical, physiological, and computational modeling studies. The differing effect of level on vertical-plane localization of short- and long-duration stimuli has been replicated behaviorally in humans and cats and for additional stimulus types. Although degradation of spectral cue representations via saturation of auditory-nerve rate profiles provides the most straightforward explanation of the NLE, attempts to account for NLE-release for long-duration stimuli (via temporal integration, high-threshold fibers, lateral inhibition, or efferent effects) have so far proven inconclusive. Those studies have, however, revealed further complexities of spectral cue processing also requiring explanation. This presentation will survey the research inspired by Hartmann and Rakerd’s experiment and describe why the mechanism of NLE-release observed at long durations is still an open question.