If and when I do a design, I'll be sure to post it. It would be as simple as possible since I don't want the signal to go through more circuits than necessary.
The metadata idea is a good one. Dolby already has an "E" variation on DD which has a metadata bitstream which could be used for this kind of waveform steering. It's used in broadcast mostly at this point.
I very rarely have more than one other person in my HT other than myself at a time, so the sweet spot is not an issue.