Why do you think it’ll ever be possible? 
 

This is a user community so there’s not anyone here that’s likely to be able to answer that. 

I second this. I know it’s been posted a lot, but you would think Sonos would have made this possible day one.

You might even think if it was easily done, they would have made it a feature day 1. I’d think it points to it not being easily done. The question in my mind is ‘how hard is it’, essentially ‘can it be done with the current setup?’.

The headphones connect via WiFi and I’m sure the processor of the soundbar can handle two streams, uni or multicast. I think it’s a question of demand.

What makes you so sure?

Makes me sure of what, the processor enough to handle 2 connections?

a sound bar can handle, 2 subs, two rears, and through 3rd party apps, two fronts - since these are shut off during tv swap, I would guess they can handle two head phones, worst case compressed audio.

Uhhhh, not even close.  The Ace are Atmos capable, so all those channels (including two more for Atmos) are going to the headphones, too. 

Yeah, and Dolby atmos trueHD 7.1.4 is 2 to 8mbps stream compressed, at most 2 headphones would be 16 mbps compressed losses 10 - 20mb.

if it can do 1 losses it can do two compressed.

You are just guessing at that, you have no clue what the actual capabilities are.  Besides, the Ace can do lossless Atmos, too.

Guessing at bandwidth needs? When you encode you set the parameters, and there are defined parameters you need to configure to provide a stream.

if you read what I wrote “if it can do 1 lossless it can do two compressed” you would have known I know it can do lossless.

But you do you. My point is a soundbar can handle multiple wireless connection and I think a soundbar can handle the bandwidth as I said in my second post “worst case compressed” I also said i think it’s due to demand not capability. 

How would dynamic head tracking work?
 

Tracking two movements means sending different audio to each Ace. Or, if the audio is shared it’s distracting for the non-tracked listener to hear the sound placement shift because the head-tracked listener has turned around. 

Or the compromise is to disable head tracking?