Got hyped up during the pandemic. Built 2 analog bridges and a digital bridge. Never could find a way to get latency down to a useable level, so never used any of them (but could probably sell the pi’s for a profit). Also, other musicians around here are largely computer-phobic, so I couldn’t interest anyone else. We are in Spectrum cable territory for internet, with fiber not available. I had a few musicians ping me or I pinged them to see if I needed to figure out how to run a server here, but the results were always 45ms or longer, and we know adding audio processing to that would only increase that, so it didn’t seem worth pursuing.
I’m in a small Zoom group and a few of us are discussing online jamming, so I’m exploring this again.
I am in Akron Ohio and have the nearest Jacktrip server in Columbus, about 120 miles away. After examining a bunch of Jacktrip servers all over the Eastern US, it turns out the server with the lowest latency for me (most days) is in Chicago, about 380 miles away. I’ve attached a spreadsheet where I have tested different servers at different times and days. It pays to check it out and not assume closest is best. Note that all these latencies are for ping -n 25. After deciding on a “server of the day” for a particular session, I plan to use PingPlotter to plot the latency over 10 minutes and determine jitter. My prospective collaborator is in Atlanta and isn’t available until after the holidays, at which time we will add to this spreadsheet, all the ping times from his location to the various servers. Then we will have to find a compromise and see what our latency tolerance is.
It seems like the latency is dependent on the routing between me and the AWS servers, and apparently the routing is generally better between me and Chicago. A question which I cannot get answered is, is there anything I can request of my ISP in terms of getting better routing? I noticed that I can ping the Columbus ookla SpeedTest server (about the same distance as the Columbus JackTrip server) and I get a 1 ms time, which is a special one-hop routing to make our internet service look better than it is, but this makes me think the routing to AWS could be optimized to a lesser degree if they wanted to. Anyone know the best way to explore this with them. I have attached a picture of the ping time to the Columbus Jacktrip server (left side) and the Columbus SpeedTest server (right side).
Edited: to add, that using a Focusrite Scarlett 2i4 with ASIO driver, my latency reported in the Chicago studio seems to be about 2ms greater than my ping time.