Oh no, I know. I'm just limited to wireless right now because I'm renting an old house with massive amounts of insulation. So I had tried to get the Sonos speakers working with a combined sink wirelessly, but it just wasn't able to keep up, leading to intermittent interruptions to the stream. I'm going to play with that wired in a test environment at some point, but I think I'd prefer something like Snapcast over Airplay.
But once I buy a house in the coming months, I'll do some low voltage runs to support the audio network, among other things. I figure I'll probably have a dedicated POE switch so I don't have to worry too much about QoS, probably Mikrotik if Ubiquiti doesn't release some new EdgeSwitch gear.
I'm just not sure if the software is there yet, with Pipewire AES67 support. It was "new" in v1, with I think some PTP patching in the first point release. So I'm trying to see if anyone has cut their teeth on this yet, since it's going to be pretty costly to get gear. I imagine I can just create a combined sink, but I'm not sure if PTPv2 is just automagic within the RTP configuration of Pipewire.
And potentially needing a second server for MPD/Pipewire is something I'm keeping in the back of my head. I'm hoping to run it in a container on the NAS server, probably running Debian (or maybe something more cutting edge if I'm reliant on new Pipewire releases). But RTP and PTP might need something a bit more dedicated to the task. It's not like I'm doing broadcast or some other form of professional audio here, so I'm optimistic that a container will be fine. Just a single 16/44 FLAC decode to combined AES67 sink. And since containers use a shared kernel, I wouldn't need to worry about the clock scheduling issues some hypervisors had with Asterisk and Free Switch in my previous life working on VOIP networks. But I'm also not planning on a ton of cores, 4-8 only in a low voltage CPU, sooooo.... I dunno.
Yeah, I think that's my backup plan is to get some powered speakers and Pi's to run Snapcast. But it adds a lot of complexity, and more power requirements at the speaker. On the other hand, it's more hackable than a speaker running a specific piece of software directly, without any real alternatives like I would get with a Pi. Thanks!