Determination of the measurement and simulation uncertainty of room acoustic quantities

由于两个主要缺点，讨论礼堂内的声学条件是困难的。首先，室内声学测量的不确定性仍在研究中，其次，目前对听众如何敏感地感知礼堂声场差异的理解似乎不完整。当单个数字参数（如Cgo）的波动时，这些方面是如何相互关联的很容易看出。如de Vries所述，取决于麦克风位置的LF等)与ISO 3382中引用的明显差异进行了比较。这样的发现表明，听者位置的微小变化已经改变了对礼堂内声场的感知。由于这种变化不明显是普遍的经验，因此需要分析这种错误推理的原因。测量不确定度和感知方面的联合讨论有可能澄清这方面。在这个项目中，两个主题并行处理。一方面，根据《测量不确定度表达指南》（GUM）的框架讨论测量不确定度，将处理尚未详细讨论的影响因素。另一方面，在实际声场的真实呈现的基础上，确定了广泛使用的室内声量的仅显着差异。将加强用于分析听力测试数据的标准统计方法（如受限最大似然策略），以包括物理量（描述刺激）的统计测量不确定性。基于这些进步，确定的差异阈值有可能更好地反映听者的敏感度。

Discussing the acoustic condition in auditoria is difficult due to two major shortcomings. Firstly the uncertainty of room acoustical measurement is still under invesfigation and secondly the present understanding of how sensitive listeners are to perceive differences in sound fields in auditoria appears incomplete. How these aspects are related to each other can easily be seen when the fluctuations of single number parameters (e.g. Cgo. LF, etc.) depending on the microphone position, as described by de Vries, are compared to just noticeable differences quoted in ISO 3382. Such findings would suggest that the perception of the sound fields in an auditorium varies already with minor changes of the listener position. Since it is common experience that such changes are not noticeable, the reason for this false reasoning needs to be analysed. A joint discussion of measurement uncertainties and perceptual aspects has the potential to clarify this aspect. In this project two topics are addressed in parallel. On the one hand the discussion of measurement uncertainty according to the framework of the "Guide to the expression of uncertainty in measurement" (GUM) will treat Influence factors that have not been discussed in detail yet. On the other hand just noticeable difference for widely used room acoustic quantities will be determined on the basis of realistic presentations of real sound fields. Standard statistical methodologies to analyse the data from listening tests (such as constrained maximum-likelihood strategies) will be enhanced to include the statistical measurement uncertainty of the physical quantities (describing the stimuli). Based on these advances the determined difference limens have the potential to better reflect a listener's sensitivity.