Hook
More breakout videos from this creator.
You have four sona speakers playing the same song in different rooms and you walk between them with no echo, lag or drift. And every speaker is producing the same audio sample at the same microsecond. And the way they pull it off is a network time protocol, similar in spirit to PTP, that every speaker uses for telecom or HFT. And every speaker buffers about 100 milliseconds of audio and constantly corrects for crystal oscillator drift. Tens of parts per million. Sounds like nothing until you multiply that out over hours. So speakers actually constantly exchange timestamped messages, measure network delay and then correct themselves with roughly 100 milliseconds of buffer to give them time to align. So it's really boring network engineering achieving something really beautiful. THE PAYOFF — sample-accurate across every room. The crystal oscillator in every speaker drifts tens of parts per million. Sounds like nothing until you multiply that out over hours. So speakers actually constantly exchange timestamped messages, measure network delay and then correct themselves with roughly 100 milliseconds of buffer to give them time to align.