It measures and accounts for latency between devices.
> NTP can usually maintain time to within tens of milliseconds over the public Internet, and can achieve better than one millisecond accuracy in local area networks under ideal conditions
https://en.wikipedia.org/wiki/Network_Time_Protocol#Clock_sy...
The algorithm is more sensitive to variance than latency, but over a LAN variance is usually pretty okay. Wifi can be a lot worse, though, especially if there are a lot of clients.
Thank you kindly!
A fun fact is that that's why a symphony needs a conductor: one corner of the symphony to the other is 17-ish meters, which is 50 ms of latency. They have to lock in to a visual queue because at those distances, you can't synchronize on sound. It's also interesting to note that a symphony very literally sounds different on one side than the other, and not just in relative prominence: at march tempo (120 BPM), if you're sitting on one side of the symphony, some instruments may be a 32nd note ahead of the ones on the other side by the time they reach your ears.
I am trying to envision how syncing DAW's through the network could possibly avoid what easily amounts to more than 20ms of delay back and forth between devices.
Is this just not an issue for digital music production? It's a massive issue for live music and in-ear-monitors on stage.