飞行模拟器中的虚拟声音建模与实时渲染技术

Virtual Sound Modeling and Real-time Rendering in Aircraft Simulator

虚拟声音的建模与渲染是飞行模拟器虚拟环境仿真中的关键技术,结合物理、心理、建筑、室内声学以及数字音响艺术等多门学科,提出适用于分布式飞行模拟器虚拟仿真系统的离线综合声音建模与动态均衡渲染算法,经过数字压限器、扩展器以及激励器等环节,将大规模音频数据根据虚拟权重进行离线重构;运用五段等响折线法,并结合心理声学构建了基于大规模音源的动态压缩比算法;结合座舱虚拟环境的等效吸声系数算法和混响时间算法;综合功率谱密度函数与特性共振峰,总结出与一系列状态量对应的幅频特性,应用前馈(BP)网络算法对数字均衡器进行实时调参,实现了发动机推力噪声的动态数字音频实时渲染;并基于DirectSound的音频特效程序接口开发了一套基于高层体系结构(HLA)的分布式飞行模拟器实时音效系统,仿真结果验证了声音建模与渲染算法的有效性。

Virtual sound modeling and rendering are key technologies in the virtual environment of an aircraft simulator. Based on physical acoustics, psychoacoustics, architectural acoustics, room acoustics and digital audio technology, a synthetic underline audio modeling method and an optimal dynamic audio rendering algo- rithm are presented for a distributed aircraft simulation system. Large-scale sound data are remodeled by the digital compressor, expander and aural exciter. A 5-segment equal loudness level contour is proposed and an advanced compressor ratio algorithm is realized based on psychoacoustic parameters. An equivalent absorption coefficient and reverberation time estimating algorithm are also presented based on virtual capsule environment information. Frequency spectrum and formant data samples are summarized according to different state variables and a back propagation （BP） network algorithm is realized by several equalizer parameters adjusted in realtime thrust noise simulation. A real-time virtual sound system is adopted in a certain aircraft simulator based on high level architecture （HLA） which makes the aural rendering more accurate and vivid.