[LAU] Choirless - a new low-latency AI remote music collaboration platform

Giso Grimm gg3137 at vegri.net
Tue May 11 17:18:08 CEST 2021


Dear Linux Audio Users,

On 11.05.21 16:45, Ketil Thorgersen wrote:
> Hi group
> I guess the reason why the IBM solution came up in the group is that it claims to be built upon open source. Otherwise it seems very similar to https://easyvirtualchoir.com . None of these seem to be synchronous and as such are more similar to an online DAW like SoundTrap or BandLab than Jamulus which is an online low-latency open source jamming tool.
> 
> I use Jamulus on a regular basis and it works really great as long as you are in the same area and everyone is wired to the internet.  I play in a jazzband and even the drummer finds the latency tolerable (at least if we play below 220 bpm)
> My wife is a choir conductor and they rehearse with Jamulus. 
> 
> There are other projects out there such as https://sonobus.net (tested but did not really work well at the time so we went back to Jamulus) and Jacktrip https://github.com/jacktrip/jacktrip

Just to complete the list I would like to mention the "ovbox" system:

Documentation:

https://github.com/gisogrimm/ovbox

Source Code:

https://github.com/gisogrimm/ov-client

It is using zita-njbridge from Fons Adriaensen for packaging and
resampling of audio, but is wrapped into a more end-user compatible
layer, which manages firewall hole punching and session management. It
uses (optionally interactive multi-headtracked) 3D-audio.

In the end all solutions reach more or less the same latencies. Those
solutions which do the complete rendering locally (no server-side
mixing) can achieve lower latencies, as the jitter compensation buffer
is required only once (and jitter does not add up linearly). This is on
the cost of higher bandwidth usage and more signal processing
requirements locally.

In a Raspberry Pi 4 based solution with ovbox it is possible to bring 10
people together in a session. Within Europe we achieve
microphone-to-headphone delays of 28 to 40 ms given all ends have a good
internet connection, and around 50 ms (with some jitter-related
dropouts) via mobile networks. Unfortunately in Germany only very few
have access to good internet. Europe-US-connection was 100ms.

Best,

Giso

> 
> All the best!
> Ketil
> 
> Linux-audio-user på vegne av Chris Caudle <linux-audio-user-bounces at lists.linuxaudio.org på vegne av 6807.chris at pop.powweb.com> skrev følgende den 11.05.2021, 16:24:
> 
>     On Tue, May 11, 2021 9:04 am, Robert Jonsson wrote:
>     > Let me just chime in that there are others, including at least one
>     > open source solution.
> 
>     I just saw that in the upcoming Spring Audio Engineering Society
>     conference there is a presentation titled "A complete guide to Networked
>     Music Performance using free and open-source software"
> 
>     The presentation is Friday 28 May at 1:15PM-2:00PM CEST.
>     I believe this link should go to the page for that presentation:
>     https://aesshowspring2021.sched.com/event/j83Y?iframe=no
>     I am not sure if you have to be a member of the AES to attend or not. 
>     Probably not, but there may be a cost difference for registration for
>     members vs. non-members.
> 
>     The description of the presentation is:
> 
>     This workshop will review a thorough representation of the currently
>     active open source and freely available software projects that allow for
>     networked music performance. Even prior to the pandemic,  the development
>     of tools that meet the requirements of performing music together online
>     were growing in number; since the last year, we have seen new tools and
>     existing ones have become more sophisticated and powerful, as a
>     unprecedented level of attention is paid to their use and application by
>     those musicians and music groups who traditionally work together offline.
> 
>     Networked music performance has three particular demands of the technology
>     it uses: first, the audio stream must be suitably high-quality; secondly,
>     it must minimise interference, for example avoiding echo cancellation
>     algorithms and unnecessary processing; thirdly it must be low-latency,
>     where what defines low-latency exactly depends on the intention and
>     resources of the musicians. Given these factors, the variety of
>     applications and services available that can be used for network
>     performance each have their own specific approach, usually born from the
>     original intended purpose of the developer. Certain tools may, for
>     example, focus on low-latency while others on group usability, and others
>     on solutions where participants do not have access to broadband or 4G.
> 
>     The outcome of this workshop will be that viewers who are interested in
>     performing together over the internet will receive a comprehensive review
>     of software tools, with the aim to select those that suit their specific
>     needs. In addition, the tools discuss are free-to-use and, in some cases,
>     modify. The availability of such tools leads to an expansive array of
>     music possibilities that extends the core of music practice itself.
> 
>     -- 
>     Chris Caudle
>     _______________________________________________
>     Linux-audio-user mailing list
>     Linux-audio-user at lists.linuxaudio.org
>     https://lists.linuxaudio.org/listinfo/linux-audio-user
> _______________________________________________
> Linux-audio-user mailing list
> Linux-audio-user at lists.linuxaudio.org
> https://lists.linuxaudio.org/listinfo/linux-audio-user
> 


More information about the Linux-audio-user mailing list