I have an interesting problem to mull over how to do. I have a client that wants a custom multi-room sound system setup throughout his haus. His requirements are:
1: Full control over the system, no cloud based solutions, avoid subscription services, and absolutely no voice control
2: Integration with streaming apps like Spotify and Apple
3: Complete individual zone control
Full system control naturally kicks out Bose systems and the like; and the software was easy enough to find: Home Assistant and Snapcast. Done.
Snapcast has integration capability with most popular streaming apps, also it's app is very functional. Also done.
Complete zone control on an individual basis is where my problem lies. This would be easy if I were limited to two, three, or four zones; but his new home is quite large, and specifies thirteen zones. Creating thirteen individual containers (one for each zone) is challenging when I need to attach a sound output to each of them. My current options are as follows:
1: Construct at least 2 identical servers and fill each server with sound cards, passing through each sound card to its respective container. Unless I can get a cheap PCIe expander out of a used mining setup this is not ideal but offers the distinct benefit that I can know exactly which device goes where.
--OR--
2: Use a rackmount USB hub connected to a single server and use a bunch of USB sound adapters and hope that USB passthrough keeps each port addressed exactly where it needs to be.
Any advice on either option or any alternative would be appreciated. As a sort of suggestion, would it be possible to isolate each output channel of a soundcard and map them to containers or VMs? I know this sounds ridiculous but this is a legitimate use case. Other hamfisted options I have explored involves a bunch of Raspberri Pis in a sort of cluster setup, and that option is still on the table. The primary reason I want to explore this first is because the server provided will likely be expected to be a personal file server for him.
1: Full control over the system, no cloud based solutions, avoid subscription services, and absolutely no voice control
2: Integration with streaming apps like Spotify and Apple
3: Complete individual zone control
Full system control naturally kicks out Bose systems and the like; and the software was easy enough to find: Home Assistant and Snapcast. Done.
Snapcast has integration capability with most popular streaming apps, also it's app is very functional. Also done.
Complete zone control on an individual basis is where my problem lies. This would be easy if I were limited to two, three, or four zones; but his new home is quite large, and specifies thirteen zones. Creating thirteen individual containers (one for each zone) is challenging when I need to attach a sound output to each of them. My current options are as follows:
1: Construct at least 2 identical servers and fill each server with sound cards, passing through each sound card to its respective container. Unless I can get a cheap PCIe expander out of a used mining setup this is not ideal but offers the distinct benefit that I can know exactly which device goes where.
--OR--
2: Use a rackmount USB hub connected to a single server and use a bunch of USB sound adapters and hope that USB passthrough keeps each port addressed exactly where it needs to be.
Any advice on either option or any alternative would be appreciated. As a sort of suggestion, would it be possible to isolate each output channel of a soundcard and map them to containers or VMs? I know this sounds ridiculous but this is a legitimate use case. Other hamfisted options I have explored involves a bunch of Raspberri Pis in a sort of cluster setup, and that option is still on the table. The primary reason I want to explore this first is because the server provided will likely be expected to be a personal file server for him.