Models of room acoustical perception

房间声感知模型

• 批准号: 192385161
• 负责人: Professor Dr. Steven van de Par
• 金额: --
• 依托单位: Abteilung Medizinische Physik
• 依托单位国家: 德国
• 项目类别: Research Units
• 财政年份: 2011
• 资助国家: 德国
• 起止时间: 2010-12-31 至 2017-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/192385161?language=en
• 关键词: Models; room; acoustical; perception

Perceptual models will be developed that predict the perception of room acoustical attributes such as spaciousness, apparent source width, listener envelopment, reverberance, warmth, brilliance, clarity, transparency, and loudness. Whereas most existing approaches are based on signal statistics such as e.g., InterAural Cross Correlation (lACC), the proposed work will be based on an auditory front-end that creates a signal representation that is 'internally' available to the listener. Using such an internal representation allows to make a distinction between signal properties that are perceptually relevant and properties that are inaudible. This should enable a more accurate prediction of the perception of room acoustics. Variations in room acoustical simulations will be used to create room acoustical variations of different degrees that can be evaluated by human listeners and tested against model predictions to adapt and improve the perceptual models. Here the aim will both be to accurately predict discriminability of room acoustical simulations that are close to one another (as a means to test the basic validity of the internal representations in the model) and to predict perceived room acoustical attributes such as envelopment, reverberance, etc. The role of head orientation in the perception of room acoustical perceptual attributes will be investigated by listening tests and in the perceptual models developed.

将开发感知模型来预测房间声学属性的感知，例如宽敞度、视源宽度、听众包围感、混响、温暖、亮度、清晰度、透明度和响度。虽然大多数现有方法基于信号统计，例如耳间互相关（lACC），但所提出的工作将基于听觉前端，该前端创建听者“内部”可用的信号表示。使用这种内部表示可以区分感知相关的信号属性和听不见的信号属性。这应该能够更准确地预测室内声学感知。室内声学模拟的变化将用于创建不同程度的室内声学变化，这些变化可以由人类听众进行评估，并根据模型预测进行测试，以适应和改进感知模型。这里的目标是准确预测彼此接近的房间声学模拟的可辨别性（作为测试模型中内部表示的基本有效性的一种手段），并预测感知的房间声学属性，例如包围感、混响等。头部方向在房间声学感知属性感知中的作用将通过听力测试和开发的感知模型来研究。