@dbb I don't think its that the wrong things being measured, rather they don't have the tools and/or measuring protocol to replicate the complexity of human hearing and individual perceptions of what we hear. In regard to 005, I still question whether 005 would sound better with lower jitter and better SINAD? Jitter was major issue with early digital, people weren't aware of this until they learned to measure and apply fixes, lowering jitter was important for improving digital sound quality. Perhaps 005 would be even more relaxed analog like with lowered jitter. If sound quality not improved, does that mean there is some threshold of jitter beyond which it doesn't matter for perceived sound quality?
Its kind of like my experience when I added audiophile switch, adding clock and galvanic isolation from my router should have resulted in lowered noise/better sound, not the case. There is now new thread up in regard to poster getting similar poorer quality sound with newly added Phoenix USB, same issue I had? Is this mismatched clocks or a result of excessive clocking within streaming chain? Perhaps better measurements may in fact lead to poorer sound quality. Or is adding the device resulting in better measurements only exposing defect in other part of system? So much still to be discovered.
@americanspirit Melm condensed issue of general service computer in streaming setup very well. In regard to Nucleus, pretty nice for Roon, only issue with those is you'll need usb decrapifier/streamer/Roon endpoint for absolute best usb connection to 005. To get beyond need for the decrapifier/streamer/endpoint would require much greater expenditure to top line servers with built in optimized usb. In the end, Nucleus with quality linear power supply is about as good as it gets below $10k. The Nucleus alone will be nice upgrade over your present server, you can always get the add ons later.