LATENCY, mostly TOTAL LATENCY

[Fill]

Without making assumptions; have you;
1. Connected directly to the modem with an ethernet cable. 2. Made sure you have no other programs running including web browsers, Mail or Notes (force quit on Mac to check if there is anything still running). 3. Made sure nothing, and no one is connected to your router via wifi while you are using JK. Have unplugged any unnecessary peripherals from the modem and computer.
Also; Makes no difference to ping, but; I got a massive improvement in broadband speed by getting in touch with the service provider and complaining about the slow speed.They sent an engineer who put in an updated socket. He also re set the router which cn help a lot as I think it has a cache memory which can slow it down. I 'm not certain but I doubt there is much you can do about the ping speed. I think you can do an online check to compare service providers which might give you some info on whether you have the fastest available.

[StuartR]

I'm just getting started with Jamkazam and have terrible total latency, as observed by others. I've been thinking of switching ISPs to the same one as my neighbor down the block and was wondering if I could use ping to compare my network latency to his. To take that further, several cohorts are also new and troubleshooting latency issues, and i would like to compare pings with them as well.

So, my question is: Would it be a useful comparison if different computers all ping jamkazam.com and report their results back to me? Or, is there perhaps a different host at Jamkazam that would be more appropriate (considering that the machine answering to "jamkazam.com" might be strictly a webserver and relatively slow)?

I do realize that pings also include latency introduced by home routers, but I think the comparison would be interesting nonetheless.

(If there's another thread that answers this question, please, somebody, point me to it.)

Thanks,
Nate

Ok, first thing to know is that PING isn't the best tool to measure latency. Since it uses a dedicated protocol (ICMP) it doesn't correctly reflect the latency of IP-based packets. As an alternative, use something like the TESTMY.NET latency test. 

But I wouldn't even bother right now as the JK team is getting close to rolling out the second phase of their network acceleration service. This will be a game changer for some of us, depending on where we are located. Hang in there for a few more weeks to see what happens.

[mchron1222]

If my latency from my provider is 25ms, can it be reduced in order to use Jamkazam, or is it pointless to even try?

i am hooked up directly to my modem, not using wifi.  This latency came from a speedtest, my provider is xfinity

Mike C

[Dimitri Muskens]

If my latency from my provider is 25ms, can it be reduced in order to use Jamkazam, or is it pointless to even try?

i am hooked up directly to my modem, not using wifi.  This latency came from a speedtest, my provider is xfinity

Mike C

>>>
Just assuming you mean ping time with "my latency"?
Who/what are you pinging that gives this result?
Did you try multiple addresses? And really 'generic' ones like some google or MS servers, apart from the obvious (your isp)

If it could be reduced is between your personal situation (location/connection/network eqpt) and your ISP (modem-/router-/firmware/package deal)
Jamkazam itself will not reduce your inet speed or improve inet quality. 25ms does seem rather high/slow for just inet connection ping.

[Nate Stelton]

It seems that choosing the right ISP may be an important factor regarding total latency. I haven't figured out how to determine this yet. This PCMag ISP latency comparison article be helpful: https://www.pcmag.com/news/the-best-gami...s-for-2020. Also, from StuartR's recommendation, here is a Test My Latency page: https://testmy.net/latency?testALL=1.

However, to StuartR's point, most latency tests use ping, which sends/receives packets via ICMP, whereas I believe Jamkazam uses UDP/IP (true?), so pings results may be misleading. There are ping-like programs that measure latency at the UDP layer, such as hping, but I haven't tried one yet.

I thought I saw somewhere that Jamkazam used to have a test server you could use to find out if your network path was causing poor latency or jitter, but that they took it offline.

I would think the true test would be to send a UDP ping to whatever Jamkazam server(s) does the actual work, and finding an IP address for that was the main point of my query.

[Dimitri Muskens]

It seems that choosing the right ISP may be an important factor regarding total latency. I haven't figured out how to determine this yet. This PCMag ISP latency comparison article be helpful: https://www.pcmag.com/news/the-best-gami...s-for-2020. Also, from StuartR's recommendation, here is a Test My Latency page: https://testmy.net/latency?testALL=1.

However, to StuartR's point, most latency tests use ping, which sends/receives packets via ICMP, whereas I believe Jamkazam uses UDP/IP (true?), so pings results may be misleading. There are ping-like programs that measure latency at the UDP layer, such as hping, but I haven't tried one yet.

I thought I saw somewhere that Jamkazam used to have a test server you could use to find out if your network path was causing poor latency or jitter, but that they took it offline.

I would think the true test would be to send a UDP ping to whatever Jamkazam server(s) does the actual work, and finding an IP address for that was the main point of my query.

>>>
Pinging Jamkazam servers be it ICMP or per a UDP ping of any kind is not going to give you any valuable information at all. The actual jamming is done via a P2P connection between the session peers. Your audio (video) never touches the Jamkazam server(s) during a session.

if the ping time is of major importance to you, finding out your session peers IP-address(es) and pinging that would make a lot more sense. But, wouldn't it be much easier to just get online and try a jam, see what's what and take it from there ... .

[Nate Stelton]

Thanks much to Dimitri for the insight on how Jamkazam works. Knowing that Jamkazam clients communicate peer-to-peer sheds a lot of light on the whole subject. For example, it perhaps explains why 2 players seem to be able to play together in sync while a third player drifts off the beat.

I'm still thinking that the Jamkazam clients all communicate via UDP because, since UDP is connectionless, it inherently has lower latency than TCP.

Anyway, getting back to my original post, I'm trying to determine if I should switch from my current ISP (AT&T twisted pair) to Metronet, which touts fiber to the house. My neighbor 3 doors down has Metronet, but is not a musician, so trying a jam from his house is not an easy option. Therefore I was trying to come up with another way to test if Metronet will give me better latency, before I go through all the hassle of switching ISPs.

[Dimitri Muskens]

Thanks much to Dimitri for the insight on how Jamkazam works. Knowing that Jamkazam clients communicate peer-to-peer sheds a lot of light on the whole subject. For example, it perhaps explains why 2 players seem to be able to play together in sync while a third player drifts off the beat.

I'm still thinking that the Jamkazam clients all communicate via UDP because, since UDP is connectionless, it inherently has lower latency than TCP.

Anyway, getting back to my original post, I'm trying to determine if I should switch from my current ISP (AT&T twisted pair) to Metronet, which touts fiber to the house. My neighbor 3 doors down has Metronet, but is not a musician, so trying a jam from his house is not an easy option. Therefore I was trying to come up with another way to test if Metronet will give me better latency, before I go through all the hassle of switching ISPs.

>>>
There's absolutely no discussion nor secrets about the connection protocol used by Jamkazam. It is most definitely UDP.

Fibre/fiber/glass is always going to be superior to any copper/coax connection.

As far as testing from a neighbors house; your laptop with JKZ, audio interface, headphones, any instrument and/or microphone (so not even a shopping bag full) and a knock on his door, will tell you everything ... ;-)

[jazzerone]

I'm a novice to UDP, as described here by Dimitri, but have a basic understanding of the difference between UDP and TCP, and how peer-to-peer works. So, my question is this: With 3 players on board, does it matter who initiates the session, or are each of the participants on equal status, so to speak?

I have a 200mbps wired internet connection, with about 8ms latency on my AI. I have regular session with a local player and one in California (1500 miles away). I always initiate the session and invite the others to join. The other local player has no latency issues with the California player. His internet speed is 300mbps, with about the same latency as I have. He experiences virtually no latency issues with the California player; I, on the other hand, see about 45ms total latency and experience the same, which puts us slightly off --- still playable, but noticeable. Yesterday I had a session with 2 California players, and it was nearly impossible. Again, I initiated the session, invited them to join, and they had no issues with each other.

I understand that the distance from me to them, and their proximity to each other comes into play here, but I'm trying to figure out why my local friend doesn't experience this problem with the California player, while I do. We use the same ISP, and live about 15 miles apart. We're running comparable audio interface gear, both on MAC platforms. So, back to my original question: In a UDP setup, does it matter who initiates the session?

Thanks for any insights.

[StuartR]

I'm a novice to UDP, as described here by Dimitri, but have a basic understanding of the difference between UDP and TCP, and how peer-to-peer works. So, my question is this: With 3 players on board, does it matter who initiates the session, or are each of the participants on equal status, so to speak?

I have a 200mbps wired internet connection, with about 8ms latency on my AI. I have regular session with a local player and one in California (1500 miles away). I always initiate the session and invite the others to join. The other local player has no latency issues with the California player. His internet speed is 300mbps, with about the same latency as I have. He experiences virtually no latency issues with the California player; I, on the other hand, see about 45ms total latency and experience the same, which puts us slightly off --- still playable, but noticeable. Yesterday I had a session with 2 California players, and it was nearly impossible. Again, I initiated the session, invited them to join, and they had no issues with each other.

I understand that the distance from me to them, and their proximity to each other comes into play here, but I'm trying to figure out why my local friend doesn't experience this problem with the California player, while I do. We use the same ISP, and live about 15 miles apart. We're running comparable audio interface gear, both on MAC platforms. So, back to my original question: In a UDP setup, does it matter who initiates the session?

Thanks for any insights.

I don't think it matters who initiates a session. Have you verified that your home router isn't congested while you're in a JK session? If it's being used for video streaming or game play for example, your time-critical UDP packets aren't going to make it in and out in a timely manner. Check your routers ability to handle this at dslreports/speedtest.

Also, if you're using MAC gear, be sure you use an audio interface that supports Thunderbolt 3 and has a custom driver for that. Otherwise you'll have you use the MacOS Core Audio subsystem which isn't optimal and may not give you the lowest hardware latency your audio interface is capable of delivering.