Spectral Profile Cues and Synthetic Listening

描述（由申请人提供）：近年来，越来越多的心理声学证据表明，落在信号频率区域之外的刺激成分可以对该信号的处理产生重大影响。在剖面分析和调制掩蔽释放（CMR）中，效果是降低阈值：侧翼刺激成分的存在倾向于提高灵敏度。在其他现象中，如跨信道掩蔽，侧翼分量的作用是提高阈值。当前项目的目标是开发一个跨通道处理模型，以捕获促进和干扰效应。该模型的前提是，听觉对象的最生态相关特征是其频谱内容作为时间的函数，并且与合成听力相关的刺激条件（例如，跨频率的共同发作）可以根据频谱轮廓来理解，而不管这种处理类型是否是最佳的。此外，我们还提出了一种心理声学现象，可以用短期频谱线索来理解由振幅调制刺激成分（如CMR）进行的跨通道处理。实验将使用标准的心理声学方法和人类观察者来测试这种方法的定性预测，使用与促进和干扰相关的范式。结果将使用标准参数统计技术进行分析，主要是重复测量方差分析。将对这些数据以及已发表的研究进行建模，以便对涉及跨频率掩蔽和掩蔽释放的一系列现象进行定量预测。

DESCRIPTION (provided by applicant): In recent years there has been a growing body of psychoacoustical evidence demonstrating that stimulus components falling outside of the frequency region of a signal can have a significant effect on the processing of that signal. In profile analysis and in comodulation masking release (CMR) that effect is to decrease thresholds: the presence of flanking stimulus components tends to improve sensitivity. In other phenomena, such as across channel masking, the effect of flanking components is to elevate thresholds. The goal of the current project is to develop a model of across-channel processing that captures both facilitation and interference effects. This model is based on the premise that the most ecologically relevant characteristic of an auditory object is its spectral content as a function of time, and that stimulus conditions associated with synthetic listening (e.g., common onset across frequency) can be understood in terms of the spectral profile, regardless of whether this type of processing is optimal. Further, it is proposed that a psychoacoustical phenomenon demonstrating across-channel processing with amplitude modulated stimulus components, such as CMR, can be understood in terms of short-term spectral cues. Experiments will use standard psychoacoustical methods with human observers to test qualitative predictions of this approach, using paradigms associated with both facilitation and interference. Results will be analyzed using standard parametric statistical techniques, primarily repeated measures ANOVA. Data from these as well as published studies will be modeled in order to obtain quantitative predictions for a range of phenomena involving across frequency masking and masking release.