With increasingly complex environments, the cost of auralization can quickly become a significant bottleneck for interactive applications, such as video games or simulators. While limitations of the human auditory perception have been successfully leveraged for lossy audio compression, real-time auralization pipelines still implement brute-force processing, independent of the content to process and perceptive capabilites of a human listener. Several recent studies have shown that listeners are unlikely to perceive a complex auditory environment in its entirety. In this paper, we review a number of recent approaches which can be used to organize and simplify virtual auditory scenes, leveraging both the nature of the contents and limitations of human perception. Such approaches are independent of the spatial reproduction techniques and lead to scalable processing performance with minimal degradation in the perceived spatial and tonal quality. They can be applied to a variety of applications and allow for rendering highly complex environments at interactive speed.