Buell T N, Hafter E R
Department of Psychology, University of California, Berkeley 94720.
J Acoust Soc Am. 1991 Oct;90(4 Pt 1):1894-900. doi: 10.1121/1.401668.
Perceptual grouping of the frequency components from a source into a single auditory object is needed when localizing a complex sound in an environment where other sounds are also present. Two acoustic regularities that might allow for such grouping are a harmonic relation among the components and a commonality of their spatial positions. The utility of these cues was examined in a forced choice psychophysical task by measuring sensitivity to interaural differences of time (IDT) for low-frequency stimuli presented via earphones. In the first experiment, stimuli were composed of either one, two, or three frequencies. A signal detection analysis used to predict the effects of combining information across frequencies found summation to be optimal, regardless of the harmonicity of the complex. A second experiment presented two-frequency complexes in which one tone, the target, contained the IDT to be detected while the other, the distractor, was constant across all three intervals of the forced choice. For inharmonic complexes, performance for the target-distractor combinations was equivalent to that found for targets presented alone, suggesting segregation of the targets and distractors into separate auditory objects. However, for harmonic target-distractor combinations, performance was diminished. A signal detection analysis of these data supports the idea that for purposes of lateralization, the interaural information in the targets and distractors was combined into a variance-weighted value, even though it meant a lowering of performance. Thus it seems that for the grouping of complex acoustic stimuli in space, harmonic structure is more important than commonality of spatial position.
当在存在其他声音的环境中对复杂声音进行定位时，需要将来自一个声源的频率成分感知分组为单个听觉对象。可能允许这种分组的两种声学规律是成分之间的谐波关系及其空间位置的共性。通过测量耳机呈现的低频刺激的耳间时间差（IDT）敏感性，在强制选择心理物理学任务中检验了这些线索的效用。在第一个实验中，刺激由一个、两个或三个频率组成。用于预测跨频率组合信息效果的信号检测分析发现，无论复合体的谐波性如何，求和都是最优的。第二个实验呈现了双频复合体，其中一个音调（目标）包含要检测的IDT，而另一个音调（干扰项）在强制选择的所有三个间隔中都是恒定的。对于非谐波复合体，目标-干扰项组合的表现与单独呈现目标时的表现相当，这表明目标和干扰项被分离为单独的听觉对象。然而，对于谐波目标-干扰项组合，表现会降低。对这些数据的信号检测分析支持这样一种观点，即出于定位目的，目标和干扰项中的耳间信息被组合成一个方差加权值，尽管这意味着表现会降低。因此，对于空间中复杂声学刺激的分组，谐波结构似乎比空间位置的共性更重要。
