An exemplary audio processing system accesses audio stream data representative of a plurality of virtual sounds that virtually originate within an extended reality world and include real-time captured speech originating from a speaking user experiencing the extended reality world and a prerecorded or synthesized sound originating from a real audio source. The audio processing system also accesses acoustic propagation data representative of characteristics that affect propagation within the extended reality world of the virtual sounds to a listening avatar. As the virtual sounds propagate to the listening avatar, the audio processing system renders, based on the audio stream data and the acoustic propagation data, a composite binaural audio stream that represents simulated reproductions of the virtual sounds. The simulated reproductions of the virtual sounds are customized to account for the characteristics that affect the propagation of the virtual sounds to the listening avatar. Corresponding systems and methods are also disclosed.