BrainModular

Further to a thread I started over here: https://brainmodular.com/forums/viewtop ... f=3&t=6646
I'd like to suggest a simple interface within BrainModular for working with people over the internet - it would be great if you could give yourself a simple ID and send/receive invitations that would enable streaming of midi/osc/data/audio(?) between distant machines. If BM could handle all the fiddly stuff - ip addresses, dynamic ips, port forwarding, etc it would take a lot of the pain out of the process.

There are a few things out there that facilitate online collaboration - https://jammr.net/ (using Ninjam), Jamtaba... but they are more oriented towards 'jamming' and aren't very well adapted for midi/osc/computer music.

I have found a slightly old Puredata project that explores this idea in a very interesting way. If anyone knows their way around Puredata - check out Netpd https://www.netpd.org/ . It runs on 'vanilla' pd with a few externals.

I'm sure Brain Modular could integrate something like this!

it's a good suggestion but a huge job to implement, several years to do that properly.
Also we'd need to by a big data center to handle audio flows. impossible for us.
senso+++

senso wrote: ↑

28 May 2020, 08:38

it's a good suggestion but a huge job to implement, several years to do that properly.
Also we'd need to by a big data center to handle audio flows. impossible for us.
senso+++

I don't think it's good idea to try and handle audio flows via data center or any cloud solutions, p2p will have least possible latency. I do wonder how Parsec works on that regard though, it seems to establish p2p connections, and uses server just for users to bind their connections to each other. Encoding video takes about 9ms, while decoding 0.2ms or something like that, but audio encoding/decoding should be faster or with no encoding it's just a matter of bandwith and network latency.

maybe it would be possible to collaborate with other projects that try implementation of this? or use opensource solutions like ninjam to integrate in Usine perhaps?

meanwhile realistically it's best to just finalize JACK support, as Artmesh comes around for PC, it will rely on JACK and OSC. so we could simply use that.

senso wrote: ↑

28 May 2020, 08:38

it's a good suggestion but a huge job to implement, several years to do that properly.
Also we'd need to by a big data center to handle audio flows. impossible for us.
senso+++

Sorry, I wasn't very clear - the central idea would be to share OSC data mainly and synchronise mirrored patches in a simple way. The audio element is not so important (and besides there are already a few ways to stream audio between collaborators- Jamtaba, etc) This would mean that the actual data flow would be relatively light and latency wouldn't matter so much.

If you join the netpd jam on a Thursday night you'll get a good idea of what this means in practice - I can't recommend it enough!

I don't doubt that anything like this in Brain Modular could take a long time to implement though...