Very nice paper Volker. This reminds me I didn't fired my server up for months. Oooops.
I'm typically the kind of people you describe in your paper :( : had issues I couldn't diagnose so I gave up, thinking it was a too bad quality network. Thank you for making me to think about the CPUfreq governor, both on my laptop and even the server :).
Just an idea from a networks newb: did you though using multicast for the server to feed back clients? But maybe all ISPs won't allow it?
Bye bye
Fabrice
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
This reminds me I didn't fired my server up for months. Oooops.
It would be great to see a new server in the server list :-).
had issues I couldn't diagnose so I gave up, thinking it was a too bad quality network.
If you need help you can always ask in this forum if you like.
Just an idea from a networks newb: did you though using multicast for the server to feed back clients?
What benefit do you get from the multicast transmission? Usually the number of clients connected to the server is like 2 to 10. In my opinion there is no need for a multicase. A point-to-point connection should work just fine if the number of clients is that low.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Well, the idea of multicast went to my mind when in jamulus GUI I saw the bandwidth (Audio Stream Rate) from my client to the server being ~306kb/s (including udp headers?) when even set to "low quality" and "mono". Beleiving the stream to each client is the same and having only a solid 800kb/s DSL upload, I infered I could only host 2 clients.
At the moment I do not publish my server I set up for friends to jamulus.dyndns.org because I have security concerns. Maybe we could speak about that privately? I'd be glad to see it in your central server :-)
Whatever, thank you again for your gentle job.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I tried to publish my server few minutes ago, but I think I hit the bug#12 (https://sourceforge.net/p/llcon/bugs/12/). News about it?
I'll try again tonight 8-23pm when I can check away from home...
hmmm, I just tried with my smartphone tether, but strangely all usual servers were shown... but none out of the 3 (among them mine) french ones...the ones I have the better ping to. ???
Last edit: Gronaz 2015-03-09
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I didn't it world really work over the tether, but at least see my server published like I see the others. I surely have a config error. I'll see that later. If I want my server visible in jamulus.dyndns.org, I must specify -e "jamulus.dyndns.org" in my server launch script isn't it?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
You state in the paper -
"To play synchronized with the others, one
must only listen to the audio mix coming back
from the server. All local direct monitor sounds
must be muted"
Actually isn't better to listen to the audio mix of everyone else coming back, but mute your own instrument coming back and listen to yourself playing locally (with no delay)? I find that even with a small delay, it is really hard to play when you hear yourself lagging.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If you mute your instrument, you are not playing in sync with the others anymore. That may work if the overall delay is <25 ms. But if you have higher delays, you will notice that you will play the song slower and slower with time if you try to sync on what the other musician plays. I have experienced that myself.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Actually I found the opposite. If I try to play hearing myself come back from the server with lag, I start playing slower and slower, whereas if I hear myself locally in real time, I stay with rest of them.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
That is very interesting. The question is how the other musicians feel if you play layed back by your overall delay. Since this is the case if you synchronize on the signal coming back from the server but you have muted your own signal from the server.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello Mr. Fischer,
I'm working on a app named Jamfinder see jamfinder.info and have the idea to offer a a plugin where the musician's can connect over internet and jam together.
Now i see that you have written something that look verry nice und work.
I like to ask you, should it be possible to use your tool for my idea. It might be like an online chess with players and visitors ?
It would be very nice if you can contact me for a discussion.
gd@jamfinder.info
Kind regards Gert Dorn
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Do you mean a browser plugin? I am pretty sure that will not work because you need a low latency audio interface. For different operating systems you have different interfaces (like ASIO in Windows and Jack under Linux). The audio interface must be set up correctly otherwise the performance of Jamulus would be bad. The average internet user will not have his computer setup for real time audio.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hello all, newbie here. I've been experimenting and evangelizing Jamulus for at least a week, and finally convinced some band mates to try it.
Regarding monitoring, I've been muting myself in Jamulus, and listening to a direct monitor of my instrument (eDrums) mixed with the other musicians audio coming from Jamulus (using an external mixer). For this to work, I have to focus on playing a consistant tempo with myself, and expect to hear the rest of the band behind the beat (and hopefully they don't gradually drift further and further behind).
If I were to mute my direct monitoring, I'd still be hearing my self directly, but the delayed Jamulus mix would be louder. I'd have to get used to playing a note before it was supposed to sound, and be able to adjust. Is that the method most people are using?
I had a first Jamulus rehearsal with 4 bandmates a couple nights ago, and while it sounded sloppy, it was better than we expected. More tweaking is in order, and just getting used to the latency (ugh).
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
[Sorry it took so long to reply. Forgot my password, and it wasn't sending me a reset link.]
I've been averaging 30-40ms overall.
I did have an equipment issue during that last rehearsal: set up an external mic (muted while playing) and my eDrums to play through the line-in on my computer, but forgot to disable the computer's onboard mic, which was picking the actua lacoustic sound of drumsticks on mesh drumheads - pretty cheesy and off-putting. Didn't realize it until we were calling it quits and saying goodnight.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have written a paper about Jamulus and online band rehearsals. The paper can be downloaded from:
http://llcon.sourceforge.net/PerformingBandRehearsalsontheInternetWithJamulus.pdf
Very nice paper Volker. This reminds me I didn't fired my server up for months. Oooops.
I'm typically the kind of people you describe in your paper :( : had issues I couldn't diagnose so I gave up, thinking it was a too bad quality network. Thank you for making me to think about the CPUfreq governor, both on my laptop and even the server :).
Just an idea from a networks newb: did you though using multicast for the server to feed back clients? But maybe all ISPs won't allow it?
Bye bye
Fabrice
It would be great to see a new server in the server list :-).
If you need help you can always ask in this forum if you like.
What benefit do you get from the multicast transmission? Usually the number of clients connected to the server is like 2 to 10. In my opinion there is no need for a multicase. A point-to-point connection should work just fine if the number of clients is that low.
Well, the idea of multicast went to my mind when in jamulus GUI I saw the bandwidth (Audio Stream Rate) from my client to the server being ~306kb/s (including udp headers?) when even set to "low quality" and "mono". Beleiving the stream to each client is the same and having only a solid 800kb/s DSL upload, I infered I could only host 2 clients.
At the moment I do not publish my server I set up for friends to jamulus.dyndns.org because I have security concerns. Maybe we could speak about that privately? I'd be glad to see it in your central server :-)
Whatever, thank you again for your gentle job.
Yes, this includes all headers.
If you connect from inside the NAT, then you can play at least with 3 clients on your server.
I tried to publish my server few minutes ago, but I think I hit the bug#12 (https://sourceforge.net/p/llcon/bugs/12/). News about it?
I'll try again tonight 8-23pm when I can check away from home...
hmmm, I just tried with my smartphone tether, but strangely all usual servers were shown... but none out of the 3 (among them mine) french ones...the ones I have the better ping to. ???
Last edit: Gronaz 2015-03-09
No news yet.
Jamming over a smartphone tether will not work since the latencies over 3G or LTE is much too high. You need to have a cable modem, DSL or fiber.
I didn't it world really work over the tether, but at least see my server published like I see the others. I surely have a config error. I'll see that later. If I want my server visible in jamulus.dyndns.org, I must specify -e "jamulus.dyndns.org" in my server launch script isn't it?
Yes, correct.
Hi,
You state in the paper -
"To play synchronized with the others, one
must only listen to the audio mix coming back
from the server. All local direct monitor sounds
must be muted"
Actually isn't better to listen to the audio mix of everyone else coming back, but mute your own instrument coming back and listen to yourself playing locally (with no delay)? I find that even with a small delay, it is really hard to play when you hear yourself lagging.
If you mute your instrument, you are not playing in sync with the others anymore. That may work if the overall delay is <25 ms. But if you have higher delays, you will notice that you will play the song slower and slower with time if you try to sync on what the other musician plays. I have experienced that myself.
Actually I found the opposite. If I try to play hearing myself come back from the server with lag, I start playing slower and slower, whereas if I hear myself locally in real time, I stay with rest of them.
That is very interesting. The question is how the other musicians feel if you play layed back by your overall delay. Since this is the case if you synchronize on the signal coming back from the server but you have muted your own signal from the server.
Hello Mr. Fischer,
I'm working on a app named Jamfinder see jamfinder.info and have the idea to offer a a plugin where the musician's can connect over internet and jam together.
Now i see that you have written something that look verry nice und work.
I like to ask you, should it be possible to use your tool for my idea. It might be like an online chess with players and visitors ?
It would be very nice if you can contact me for a discussion.
gd@jamfinder.info
Kind regards Gert Dorn
Do you mean a browser plugin? I am pretty sure that will not work because you need a low latency audio interface. For different operating systems you have different interfaces (like ASIO in Windows and Jack under Linux). The audio interface must be set up correctly otherwise the performance of Jamulus would be bad. The average internet user will not have his computer setup for real time audio.
Hello all, newbie here. I've been experimenting and evangelizing Jamulus for at least a week, and finally convinced some band mates to try it.
Regarding monitoring, I've been muting myself in Jamulus, and listening to a direct monitor of my instrument (eDrums) mixed with the other musicians audio coming from Jamulus (using an external mixer). For this to work, I have to focus on playing a consistant tempo with myself, and expect to hear the rest of the band behind the beat (and hopefully they don't gradually drift further and further behind).
If I were to mute my direct monitoring, I'd still be hearing my self directly, but the delayed Jamulus mix would be louder. I'd have to get used to playing a note before it was supposed to sound, and be able to adjust. Is that the method most people are using?
I had a first Jamulus rehearsal with 4 bandmates a couple nights ago, and while it sounded sloppy, it was better than we expected. More tweaking is in order, and just getting used to the latency (ugh).
What is the Overall Delay shown in the Jamulus Settings dialog?
[Sorry it took so long to reply. Forgot my password, and it wasn't sending me a reset link.]
I've been averaging 30-40ms overall.
I did have an equipment issue during that last rehearsal: set up an external mic (muted while playing) and my eDrums to play through the line-in on my computer, but forgot to disable the computer's onboard mic, which was picking the actua lacoustic sound of drumsticks on mesh drumheads - pretty cheesy and off-putting. Didn't realize it until we were calling it quits and saying goodnight.
These are values I also have. This delay works fine with me using my edrums. I guess you have to get used to it. I do it for about 7 years now.