Jeff, Except for your "non-audible black holes", everything else you described was developed 3 decades ago by Michael Gerzon. Ambisonics records and uses vector information to reproduce localization, rather than create it using pan-pot techniques.

Think of a drawing of a circle, stored as Bitmap vs vector. If you print large, the Bitmap image will suffer breakdown; you'll see the individual pixels and jagged edges of the circle. Not so with the vector drawing, which is a set of instructions on how to draw a circle, and will scale itself to any size print.

Similarly, Ambisoncis will scale itself to any size playback: 2 speaker to 5.1 speakers to 64 speakers to as many as your heart desires. The information is stored on 4 channels: main/mono, left vs right, front vs back, up vs down. If you just want stereo playback, you just need the first 2 channels. For surround sound, all the speakers locations can be extrapolated from the first 3 channels. If you want height information too, you can use all 4 encoded channels.

The Nimbus label recorded their entire catalogue of classical music in Ambisonics and Meridian processors have an Ambisonics mode for proper decoding.

Best,
Sanjay