Software that eliminates latency issues between virtual platforms

[Tone Deaf Tony]

Hi there
Virtual lessons and Zoom meetings are a part of our everyday lives now but lag and latency issues prevent actual playing in real time. Someone recently mentioned software that eliminates any lag and he said he was playing duo with someone on the other side of the planet. Is this really true and what is it and how does it work? Do both parties need to be software equipped?
Thanks for your help!

[Eugle]

Hi! Sorry for bumping but I'm looking for the same solution. I read that JamKazam might be a good option as it aims to enable real-time collaboration by minimizing latency between participants. But it would be great to learn from the experience of others. What software do you recommend?















Enable GingerCannot connect to Ginger Check your internet connection
or reload the browserDisable in this text fieldRephraseRephrase current sentenceEdit in Ginger×

[Tone Deaf Tony]

I haven't tried any yet and people who've had good experiences haven't gotten back to me as to what they use. I'll surely look into JamKazam.
Yeah, it'd be nice to find something that works reliably. Game changer.

[sgosnell]

It's not possible to violate the laws of physics. If anyone could come up with something that removed the lag they could make a LOT of money, and already would have. But the reality is that it takes time for a signal to travel around the world, even it traveled at the speed of light. But the signal travels through an unpredictable, and often changing, number of relays in random data centers, in both directions. In short, this unicorn does not actually exist.

[starjasmine]

I don't recall what the platform was but there has been a jam platform that has no noticeable latency for years now. The participants have to be <500 miles apart physically and they need to have low-latency internet connections. Basically any decent DSL connection will do.

No violations of the laws of physics required. The server that is feeding streams to the various participants and accepting each jammer's audio out syncs up the in and out streams such that each jammer can participate in real time with what they hear.

Zoom, Teams, FaceTime etc. don't even try to do this level of latency reduction. If there's 500 ms or even a one-second delay in a zoom call you'd never know it. But trying to jam in real time on FaceTime is a non-starter (I tried it once with someone 2000 miles away and the delay was way too much.)

[Jimmy blue note]

I've worked with people who've been happy with some kind of software. They did notice that on very fast tunes, it's not perfect, but for medium tempo music, it's really quite amazing.
For teaching, it'd be a godsend.
When I was at Berklee, they'd have virtual videos on the big screens between Spain, or Japan and Boston. They'd have master classes with students half way around the world and in those situations, I was amazed that I couldn't really tell the lag, comments and even playing passages together could be done in front of audiences at both ends.
I'd really like to know just what it takes these days to have that capability in my own set up.

[RJVB]

It's not possible to violate the laws of physics..

No one is talking about violation - and we are indeed talking physics where "elimination" doesn't carry the same sense as it would in mathematics. So no one is talking about reducing latency to 0, just to levels where it is no longer an issue.

How do you think surgeons can carry out long-distance operations, or drone operators missions halfway across the world?

During the corona circus years I stumbled upon what I recall as a German project promising ultra-low latencies for remote "jamming". Initially they did have a page that set out the hardware requirements which were a little bit inhibitive for most home users.
I've tried to find them several times since, somehow always forgetting to bookmark them, but I think it was Make Music Together Online | JackTrip Labs | JackTrip Labs . They still have some best performance recommendations up on their download page.

[sgosnell]

Small lags, a second or less, don't really affect drone operations, nor even surgical procedures. But playing a tune where time feel is involved is a different story.

[RJVB]

Until something goes sideways and your reaction time is more than doubled all of a sudden (1s is the average reaction time when you're moderately familiar with a task).
OTOH and IIRC, we can get very good at coping with a fixed delay in repetitive tasks (and rythm in [dansable] music is more or less exactly that for the most part).

[benhatchins]

It's not possible to violate the laws of physics. If anyone could come up with something that removed the lag they could make a LOT of money, and already would have. But the reality is that it takes time for a signal to travel around the world, even it traveled at the speed of light. But the signal travels through an unpredictable, and often changing, number of relays in random data centers, in both directions. In short, this unicorn does not actually exist.

Great point indeed! Despite not being particularly tech-savvy, I found remi customer service incredibly helpful in explaining their software. Their team took the time to guide me through the features, making it easy for me to understand and manage most aspects by myself. It's empowering to have the tools and knowledge to handle tasks independently, thanks to their clear explanations and support. With Remi's assistance, I feel more confident navigating technology and managing my tasks efficiently.

[Jimmy blue note]

If you're local, or teaching or playing over a short distance jamkazam has been used
The origin signal is still the limiting variable but processing the signal, where a lot of latency occurs, is reduced

[st.bede]

If I remember correctly Reaper had an app called ReaJam… it seemed to work ok. I have not thought about that in years. I have no memory if the latency was bothersome or not. I just remember it working better then I thought it would… what that means I have no idea, because I do not remember what my expectations were.

[b0gy]

I used Ninjam several times and had no noticeable latency.

Cockos Incorporated | NINJAM

[michaelapk]

I've been using online platforms for several years with good results. You can't get around the physics of signal trasnmission but if all players are within ~ 500 miles of each other, all have wired ethernet connections, and are using low latency interfaces with ASIO drivers, the latency is about as low as trying to play with each other across a large stage.

Tempo is an issue. Things break down more at high tempo. It is also very helpful to have someone or something laying down a reliable rhythm. A good drummer or drum track helps. The latency problem magnifies everyone's individual timing issues.

I've used three different platforms: Jamkazam, Jamulus, and Sonobus. Each has a slightly different interface and different tradeoffs. All are multiplatform. Jamulus and Sonobus are both free. Jamkazam has a free membership but that has limits on session duration and some other features. The premium membership is something like 100 per year. None of them have the sound quality of playing live but all of them serve the purpose of being able to play with other musicians. They all can be glitchy and some nights the technical issues can be frustrating but most of the time they work smoothly.

I regularly play jazz with musicians on the east coast from NJ, MA, CT. We often have people join from Atlanta, Chicago and Northern ON and it still hangs together.

There are some very good musicians online and I've found the online community welcoming to those of us that are still very much learning.

[RJVB]

It is also very helpful to have someone or something laying down a reliable rhythm. A good drummer or drum track helps. The latency problem magnifies everyone's individual timing issues.

Hmm, it should be possible for the software to measure the average latency on each connection continuously based on a central clock, and provide every participant with a metronome that's synchronised as closely as feasible (and a BPM that could be based on what the "designated rythm guy" is doing).
I guess you'd call that a feed-forward control loop

[michaelapk]

I think all three of the platforms I use have a metronome function that sends a synchronized click. That works ok but you lose the normal time feel of a good drummer. I doubt that they do any kind of latency correction. I'm not a developer but it isn't possible to shift audio backwards in time so the only corrections that could be accomplished would shift audio forward in time to create alignment and that would increase latency to something at least slightly longer than the longest latency in the ensemble. I don't think that is desirable as there is often one player whose latency is at the edge of what is useable or sometimes even longer. When that happens we usually just mute that person (usually somebody that happens to just drop into a public room).

[RJVB]

I doubt that they do any kind of latency correction.

I was thinking of latency correction applied to the metronome function I mentioned, which would be possible. Indeed, you can't shift streamed audio back in time (you'd need to predict notes that haven't been played yet), but technically it would be possible to increase the latency on the faster connections so that everyone experiences about the same amount of lag. I couldn't tell if there's any advantage to that; probably not if the latency differences are big.

[rpjazzguitar]

Some friends have been using Jacktrip for several years, to play together on-line.

They say it's not perfect, but it's good enough. I think they avoid really high tempos.

I haven't tried it myself. At one point you had to buy some hardware. I don't know if that's still true.

I recall that sound travels at about 1.12 feet per millisecond.

So, if you're 30 feet from another musician, you hear him about 27 millisec after he makes a sound. Not that different from when the trumpets play in my big band's rehearsal space. It's workable.

Jacktrip advertises being able to match that and, based on my friends, it seems to be true.

[RJVB]

At one point you had to buy some hardware. I don't know if that's still true.

I do recall that they had a pretty extensive "how-to" up and IIRC it did strongly recommend investing in LAN equipment that most people do not have at home.

[rpjazzguitar]

I do recall that they had a pretty extensive "how-to" up and IIRC it did strongly recommend investing in LAN equipment that most people do not have at home.

I think that the box they were recommending is history. All they recommend now, iirc, is that you go wired to your router. Not wifi. In my setup, that's a Cat5 cable from the router to my computer. I had to buy an adapter to go from old-USB to USB-C. The cable and the adapter, if you need one, are cheap. It's not a smart cable of any kind, just wire and connectors.

[RJVB]

It must have been late 2020 or 2021 when I saw their recommendations, so not exactly ancient history - and there were presumably more laptops around with proper built-in ethernet connections than nowadays. I understand from your post that there's no need even to put the wired connection on a better-than-USB3 interface (which makes sense).

I can't remember now if their protocol supports video but presume it does and it is quite possible that they have been able to make progress on that end (browsers have, too, and so has hardware acceleration of encoding and decoding of efficient codecs like VP9).

But I wonder if they didn't recommend using a Thunderbolt audio interface. AFAIK most USB DACs still use USB2 which is largely sufficient for most audio applications in terms of bandwidths but which IIRC doesn't have the same (constant) latency guarantees that USB3 has.

The Wayback machine might give more answers, but it's rather good news that they now have a How-To that no longer puts lots of potential users off, so why bother

2021 ... also the year I was forced to follow my classes over a videolink. One-on-one and playing along wasn't part of the programme anyway but would have been impossible because the school allowed the use of only "Jitsi Meet" because of how they perceived compliance with the GPRD. Teacher taught from home using her connection and computer so it still beats me how the school could impose their rules but back then Jitsi was barely up to the task of even teaching with its one-way audio transmission design. Any sound detected on either end during a period of "silence" would preempt the mic and cut sound transmission from the other end leaving the other person talking to the air.

[sgosnell]

Two to three years is a long time in computer years. Has been since the personal computer was introduced.

[RJVB]

True but I don't think that really applies here (CPU power isn't the bottleneck, nor the amount of RAM or screen resolution).

I have been seeing talk of 10Gb LAN speeds the past few months but it's perfectly possible that that is only because I my radar was tuned elsewhere - and I don't think that technology is already common enough that it doesn't need to be mentioned as a beneficial thing.

I just checked: they do mention fibre network links and "USB/Thunderbolt" audio interfaces. I wonder if they actually mean USB4 with that...

[sgosnell]

The 10GB is the theoretical bandwidth of high quality cable. Not all cable is capable of carrying that much bandwidth. The newest fiber internet connections only go to 1GB/sec or less. It's theoretically possible to have a 10GB/sec local LAN at home, but LAN speeds are not the same as internet speeds. The basic problem is still that the route between internet connections is unpredictable and variable throughout the connection. The bandwidth is more than high enough for carrying data and video, even at a few hundred MB/sec, since watching YouTube videos is easy enough. Lag, and especially variable lag, is the main issue. Unless I'm totally misunderstanding the technology.

[RJVB]

Yes, lag is the problem but AFAIK bandwidth and delay aren't completely independent in that a system with high transmission times (probably) cannot ensure a very high bandwidth.
(I'm pretty certain one can perfectly play apart together (PAT ) with only 640x360 resolution although it may get tricky to read what chords everyone is playing.)

I'd have sworn that I've seen subscription offers for "fibrespeed" connections of more than 1Gb/s. (Of course one has to be careful to speak of the same units, bits/s and bytes/s aren't exactly the same thing, and kilo, mega etc. also have 2 interpretations.)