Hey all,
I've been wrestling with JK over multiple sessions during the past few weeks. Last week I was doing a test with a friend in San Francisco (I'm based in the UK), and despite the fact we both had fast wired internet connections running, the session latency had us way up around 150ms which was obviously unusable. We've tried at different times of day & night with the same results. We also had a session where we added a friend in Arizona. He was able to sync pretty well with the guy in San Francisco.
So, my question is, does the latency naturally increase with physical distance? If so, can I assume that a transatlantic session is not a viable option at this point in time?
Cheers,
Olsen
The ultimate limit is the speed of electrical signals in cables which, at best, could be the speed of light (300,000,000m/s). So in a millisecond light goes 300km. London to San Francisco is about 8600km, so the best possible latency is about 28ms, giving a round trip delay of 46ms. In practice, every intermediate router or other network device introduces a bit of delay. Also, light goes slower in any medium except a vacuum. I may be a bit out of date with my figures, but the actual latency for just the transatlantic cable from UK to East Coast USA is between 40 and 80 ms round-trip delay (depending which cable your packets are routed over).
Just as a follow-up to this, does this mean that JK doesn't actually eliminate or compensate for latency ... it simply tries to minimise the existing latency as much as possible?
That is correct. As stated above there is no way to eliminate the latency that is caused by the natural scientific properties (the physics of electrical signals travelling through communications cables). JK simply attempts to minimize the portions of the total round-trip latency that can be controlled. "Compensate for latency" is only possible in scenarios not requiring synchronization in real time, like when 2 people are trying to play in sync. Audio recording is a scenario where recording input latency can be compensated for, by delaying the prerecorded tracks by the same amount as the measured latency. Even this "local" latency can be distracting, hence many audio interfaces have "direct monitoring" so that the performer being recorded can hear themselves without the roundtrip latency going into the DAW, processed, and back out to the monitors/headphones.
Keep in mind that you as a performer routinely tolerate a small amount of latency when jamming with others, because the speed of sound is much slower than the speed of light (340 meters/sec for sound, 300 million meters/sec for light in a vacuum, >150 million meters/sec for electrical signals); roughly 1 foot of separation between musicians is about 1 ms of latency. If you are a 5-piece band rehearsing in a small basement, you can become a very tightly synced group. But move to a large empty hall/arena and spread very far apart, things would sound a mess, and not just due to the reverbs and echoes, but due to the different delays of the other musician's sound that you hear, caused by the speed of sound delay between each other. Hope this helps
