Individual Binaural Synthesis of Virtual Acoustic Scenes

虚拟声学场景的个体双耳合成

• 批准号: 402811912
• 负责人: Professorin Dr.-Ing. Janina Fels
• 金额: --
• 依托单位: Lehrstuhl und Institut für Hörtechnik und Akustik (IHTA)
• 依托单位国家: 德国
• 项目类别: Research Grants
• 财政年份: 2018
• 资助国家: 德国
• 起止时间: 2017-12-31 至 2022-12-31
• 项目状态: 已结题
• 来源: https://gepris.dfg.de/gepris/projekt/402811912?language=en
• 关键词: Individual; Binaural; Synthesis; Virtual; Acoustic

In recent years, Virtual Reality (VR) has become a major impact in many areas. With respect to the technical development of VR, the focus has traditionally been on the visual component. However, the auditory component is essential for a convincing application. There are various methods available for headphone-based rendering of acoustic virtual reality. Model-based binaural synthesis uses Head-Related Transfer Functions (HRTFs) for the auralization of individual virtual sound sources. In data-based binaural synthesis, microphone arrays are used in combination with audio signal processing and HRTFs. Generic HRTFs of artificial heads can cause problems with both approaches in terms of in-the head localization, front-back confusions and timbre. There are numerous methods for the fast measurement of individual HRTFs and algorithms for the individualization of generic HRTFs. For the former, however, there is a lack of systematic validation of measurement uncertainties due to the apparatus and to the influence of the subject's movement. The same applies to the validation of algorithms for the individualization of generic HRTFs. The effects of measurement uncertainties and approximations in the algorithms of model- and data-based binaural synthesis on the authenticity and plausibility of acoustic scenes are the central goal of this project. Here the essential question is to what extent an improvement of a virtual acoustic scene can be achieved with the help of individualized or individually measured HRTFs in comparison to generic HRTFs. In addition to instrumental measurements, empirical studies applying listening experiments are used for the evaluation.

近年来，虚拟现实（VR）已经成为许多领域的主要影响。关于VR的技术发展，传统上关注的是视觉组件。然而，听觉组件对于令人信服的应用程序至关重要。存在可用于声学虚拟现实的基于耳机的渲染的各种方法。基于模型的双耳合成使用头部相关传递函数（HRTF）来可听化各个虚拟声源。在基于数据的双耳合成中，麦克风阵列与音频信号处理和HRTF组合使用。人工头部的通用HRTF可能会导致这两种方法在头部定位，前后混淆和音色方面的问题。存在用于快速测量个体HRTF的许多方法和用于通用HRTF的个体化的算法。然而，对于前者，由于设备和受试者运动的影响，缺乏对测量不确定性的系统验证。这同样适用于通用HRTF个体化算法的验证。该项目的中心目标是，在基于模型和数据的双耳合成算法中，测量不确定性和近似值对声学场景的真实性和可扩展性的影响。这里的基本问题是在多大程度上可以实现的虚拟声学场景的改善与个性化或单独测量的HRTF相比，通用的HRTF的帮助。除了仪器测量，应用听力实验的实证研究用于评估。