These types of features (spatial audio, HRTFs, beamforming, etc.) are actually pretty straightforward and well known techniques, and like machine learning, we're just now getting to a point where we have the hardware to properly implement them.