Integrating physically based sound models in a multimodal rendering architecture