Lutfi R A
Department of Communicative Disorders, University of Wisconsin, Madison.
J Acoust Soc Am. 1990 May;87(5):2141-8. doi: 10.1121/1.399182.
A series of experiments investigated listeners' ability simultaneously to process information across different acoustic dimensions. On each trial, the listener heard a pair of brief n-tone sequences (n = 1 to 12). The frequency, intensity, and duration of each tone in the sequence varied randomly from trial to trial. On average, the values of these three parameters were greater for one sequence, the target, than the other, the nontarget. The listener's task was to identify the target on each trial. For an ideal observer in this task, d' performance grows as the square root n. Obtained d' grew at a rate slightly less than the square root of n. Close to cube root of n growth was observed when the average difference occurred in only one of the three tone parameter values within a block of trials. Although performance fell short of ideal, optimum weights were consistently given to each tone and each parameter. The results are consistent with a model in which performance depends predominantly on the information content of the sounds regardless of how the information is "packaged" in the stimulus. Transmitted information is estimated to be 0.9-2.0 bits within a single acoustic dimension, 2.1-3.0 bits when distributed across dimensions.
一系列实验研究了听众同时跨不同声学维度处理信息的能力。在每次试验中，听众会听到一对简短的n音序列（n = 1至12）。序列中每个音调的频率、强度和持续时间在每次试验中随机变化。平均而言，这三个参数的值对于一个序列（目标序列）比另一个序列（非目标序列）更大。听众的任务是在每次试验中识别出目标序列。对于此任务中的理想观察者，d' 表现随n的平方根增长。得到的d' 增长速率略低于n的平方根。当平均差异仅出现在一组试验中的三个音调参数值之一时，观察到接近n的立方根增长。尽管表现未达到理想状态，但始终会对每个音调和每个参数赋予最佳权重。结果与一个模型一致，在该模型中，表现主要取决于声音的信息内容，而不管信息在刺激中是如何“打包”的。据估计，在单个声学维度内传输的信息为0.9 - 2.0比特，当分布在多个维度时为2.1 - 3.0比特。
