Computational acoustics for virtual environments

将空间声音融入现实模拟（如沉浸式虚拟环境）似乎是显而易见的。事实上，这样做是有益的，原因有很多。空间声音线索可以添加更好的“存在感”或“沉浸感”，可以弥补糟糕的视觉线索（图像），并导致改进的对象定位。话虽如此，在声音在环境中传播时，准确地重现声音与物体/表面之间的潜在大量互动，同时考虑到人类听众，这是非常困难的，超出了我们目前的分析和计算能力，除了简单的环境。也缺乏对一般空间音响系统的有效性和准确性（例如，模型对所模拟的实际现象的忠实程度）进行详细评价的既定方法。许多评估涉及将系统计算的各种属性/统计数据与相应的理论/预测值进行比较。然而，人类听者是任何空间声音应用的最终用户，因此，为了确定这些系统的有效性，任何评估方法都必须包括人类听者，以解释人类的听觉感知。本研究计划旨在通过以下研究开发方法和技术，从而实现更有效、更逼真的动态、交互式、多模态虚拟环境模拟：1)准确性与计算时间之间的权衡；Ii)是否可以利用人类感知来限制计算资源，以及iii)通过使用不同的算法和技术，将基于人类的效果与声学建模相结合，以促进高效的听众声场。培养高素质人才是拟议研究计划的重要组成部分，并将包括培养学生。在我的指导下，学生将有机会使用最先进的设备进行独特的前沿研究，同时与声学专家进行互动和建立联系。

Incorporating spatial sound in realistic simulations such as immersive virtual environments seems obvious. In fact, doing so can be beneficial for a variety of reasons. Spatial sound cues can add a better sense of "presence" or "immersion", can compensate for poor visual cues (graphics), and lead to improved object localization. That being said, accurately recreating the potentially large number of interactions between a sound and the objects/surfaces it may encounter as the sound propagates through the environment while accounting for the human listener is extremely difficult and beyond our current analytical and computational reach, except perhaps for simple environments. There is also a lack of established methodology for the detailed evaluation of the effectiveness and accuracy (e.g., how faithful the model is to the actual phenomena being modeled) of spatial sound systems in general. Many evaluations involve making comparisons between various properties/statistics computed by the system and the corresponding theoretical/predicted values. However, the human listener is the ultimate user of any spatial sound application and therefore, to determine the effectiveness of such systems, any method of evaluation must also include human listeners in order to account for human auditory perception. This research program seeks to develop methods and techniques that will lead to more efficient, realistic simulations of dynamic, interactive, multi-modal virtual environments through the investigation of: i) the trade-off between accuracy vs. computational time; ii) whether human perception may be exploited to limit computational resources, and iii) the potential merger of human-based effects with acoustical modeling to facilitate efficient listener sound fields through the use of different algorithms and techniques. The training of highly qualified personnel is an essential component of the proposed research program and will include the training of students. While under my supervision, students will have the opportunity to work on unique, cutting edge research using state of the art facilities, while interacting and building ties with experts in acoustics.