That detail is super suspicious. I don’t remember it from the call that @eric mentioned–but it makes me think about this theory very differently.
Inside Roon, we pay attention to clocks–the system clock, and "real time’–which is based on our private NTP interactions.
We do the private NTP interaction to protect our cloud services from people with bad system clocks. We want to be sure that certain timestamps can be trusted within a known margin, and that they aren’t broken by someone who’s motherboard battery sent them back to 1970, or something like that. So we do our own time sync + manage that internally to Roon in a place that’s safe from that kind of error.
However–we don’t use that information at all for anything related to audio playback. It’s used in a few critical places related to communication with cloud services–that’s it. That’s why the idea–at first–that our NTP exchange was impacting audio playback made no sense.
But–we also don’t use our private NTP-derived clock for log timestamps. Those come straight off of the system clock.
So this is what’s concerning–your log timestamps changed right around the time of the NTP exchange–but the log timestamps are directly derived from the system clock. We don’t change the system clock as a result of our NTP interaction–we just record the skew privately. This means that your system clock moved by a few seconds. Right when we did the sync. But not because we were mucking with the system clock.
This makes me think–maybe Windows is seeing the NTP response packet from our private interaction and acting upon it by changing the system clock without our intent/knowledge. That could potentially have impact on playback, since we use the system clock to perform short-term interpolation as part of the clock recovery algorithm used by RAAT.
If this is what is happening, this is pretty gross behavior for Windows to have. I’m sure we can re-work the playback stuff to prevent it from being affected, but I still don’t want our private communications with an NTP server to have global impact on the system…
We’ll be investigating further. Thanks for the report.