Acoustic texture rendering for extended sources in complex scenes

摘要

Extended stochastic sources, like falling rain or a flowing waterway, providean immersive ambience in virtual environments. In complex scenes, therendered sound should vary naturally with listener position, differing notonly in overall loudness but also in texture, to capture the indistinct murmurof a faraway brook versus the bright babbling of one up close. Modeling anambient sound as a collection of random events such as individual raindropimpacts or water bubble oscillations, this variation can be seen as a changein the statistical distribution of events heard by the listener: the arrival rateof nearby, louder events relative to more distant or occluded, quieter ones.Reverberation and edge diffraction from scene geometry multiply and mixevents more extensively compared to an empty scene and introduce salientspatial variation in texture. We formalize the notion of acoustic texture byintroducing the event loudness density (ELD), which relates the rapidity ofreceived events to their loudness. To model spatial variation in texture, theELD is made a function of listener location in the scene. We show that thisELD field can be extracted from a single wave simulation for each extended source and rendered flexibly using a granular synthesis pipeline, with grainsderived procedurally or from recordings. Our system yields believable, realtimechanges in acoustic texture as the listener moves, driven by soundpropagation in the scene.