How bad is TOSLINK? Well, the connectors are a terrible design...
These days especially, there shouldn't be anywhere close to audible differences. *Measurable* differences are usually due to the individual equipment, not the protocol. Some sources have measured slightly better or worse with one vs. the other, even between the different types of output from the same source. Likewise, there are often easily *measurable* differences between receiving DACs.
Plastic vs. glass fiber for TOSLINK is a non-issue. Heck, you can get plastic TOSLINK cables rated for 32bit/384kHz.
The point about ground loops is a valid one. In my system, I use TOSLINK and a pair of cheap converters to get the analog signal to the dedicated rack for the subwoofers. The sub system and the main system each have their own dedicated subpanels for power. I wanted complete galvanic isolation, figuring that would be a particularly susceptible and bad place to have hum and buzz from a ground loop.
USB should be technically superior because it can be asynchronous. With TOSLINK or S/PDIF the clock is part of the signal, so it's more susceptible to jitter. In practice, the difference is negligible.
Audio data is pretty much a trivial exercise compared to Ethernet/Wi-Fi. Nothing to lose sleep over.
Sources/processing: stack of stuff that if it isn't vintage now, it will be soon!
Amps: stacks and stacks of old iron
Main speakers: big DIYSG
Surrounds: Bose graveyard
Subs: a bunch
Last edited by fill35U; 04-11-2019 at 12:23 PM.