Open Sound Control

[raboof]

OSC seems to be all the new rage these days, so I've checked it out for a bit.

My impression so far is that it is /very/ general, and that this is both a blessing and a curse. It seems to be intended as a `replacement for MIDI, and much more beyond that'. However, as far as I can see no standardized way to represent musical data (like 'play a C4 note here') has been defined yet - which means 2 applications both supporting OSC still won't understand each other unless they have agreed on certain conventions `on top of OSC.

What are your experiences/insights?

[ntnunk]

I did a little reading on the OSC site, and I have to say it seems more like one of those experimental things that people do in college as a proof-of-concept or something similar. It's interesting, and there are some things about it that would definitely be useful (the multi-device targets via pattern matching seems particularly useful to me at first introduction) but it seems like it's not ready for the real world yet. It actually feels more like a transport protocol to me than anything else, like you'd need another layer on top of OSC to do anything useful. It certainly seems like it needs more specifics before Korg, Yamaha, et al will be ready to ship hardware that supports it or even before soft-synth developers could release code that worked more-or-less flawlessly with sequencing software via OSC.

Do you have a specific problem or task that you're trying to solve/accomplish?

[brummer]

from what I understand OSC is more a protocol for comunication between a GUI and a audio tread. When comunicate via OSC, they can run as independend application. For example DSSI use OSC for comunicate between GUI and audio tread.

[raboof]

it seems more like one of those experimental things that people do in college as a proof-of-concept or something similar. It's interesting, and there are some things about it that would definitely be useful (the multi-device targets via pattern matching seems particularly useful to me at first introduction) but it seems like it's not ready for the real world yet.

Certainly - but if the concept is solid, I think the open/free world is a great environment for early adoption, bringing it into the real world for feedback-from-the-trenches.

It actually feels more like a transport protocol to me than anything else, like you'd need another layer on top of OSC to do anything useful. It certainly seems like it needs more specifics

My thoughts exactly.

Do you have a specific problem or task that you're trying to solve/accomplish?

Not really - this is one of those projects where the journey is more important than the destination. I'm brainstorming about a kind of notation editor, which of course should have 'live preview' functionality. Talking to a synth though MIDI (either jack or alsa) is an obvious candidate, but it's fun to explore new up-and-coming alternatives .

[Claudio]

I remember some time back in the late 90s, almost around the turn of the millennium, about some solutions for replacing MIDI. It was even covered in a Keyboard Magazine article around that time IIRC. It was called ZIPI (and yes, pronounced as you think it would be so) and it was supposed to be a significant upgrade to MIDI. However, there were some drawbacks to it and it just never caught on, mainly because of the success of Firewire.

http://en.wikipedia.org/wiki/ZIPI

As mentioned in this Wikipedia entry, OSC seems to be the evolution of the concept of ZIPI.

[studio32]

check out: http://www.linuxjournal.com/content/introduction-osc

[ntnunk]

Certainly - but if the concept is solid, I think the open/free world is a great environment for early adoption, bringing it into the real world for feedback-from-the-trenches.

Very true.

Not really - this is one of those projects where the journey is more important than the destination. I'm brainstorming about a kind of notation editor, which of course should have 'live preview' functionality. Talking to a synth though MIDI (either jack or alsa) is an obvious candidate, but it's fun to explore new up-and-coming alternatives .

Interesting. Perhaps an approach would be to separate the MIDI functionality to it's own abstraction layer. That way you could use OSC to communicate between the main program (the notation editor or whatever) and the MIDI abstraction layer, which would then talk to the sequencer or sound module. The MIDI abstraction layer would function as a OSC/MIDI interpreter (either simplex or duplex) and it would be modularized such that it could be dropped in favor of OSC talking directly to the sound module should that ever become feasible.

Sounds like fun!

[badmuthahubbard]

Well, if OSC seems unusable, maybe you don't need an improvement over MIDI (It was perfectly adequate for mid-80's mainstream pop ). As for me...
One of the areas MIDI fails in is polyphonic microtonality. You can only have one pitch-bend per channel, and if you want four notes in just tuning, say, with frequency ratio 4:5:6:7 (which you might hear from a barbershop quartet), you've just used 4 of your 16 MIDI channels on one instrument.

OSC seems perfectly simple to me. If you want a certain value to control a certain parameter, you set a path for it in the receiving app/hardware, and set the same path in the sending app/hardware to send that value. Done. It's perhaps not unlike continuous controllers in MIDI, but an unlimited amount, and you can name them something more relevant than 178. And they're not limited to 128 values, nor to integers. If that's too hard, maybe you don't need it.
I haven't seen too many hardware controllers that had a very useful implementation of OSC. The most celebrated one, I think called Lemur or something, was all buttons, no sliders or knobs. But for inter-process communication, OSC has been very useful for me.

-Chuckk

[raboof]

Well, if OSC seems unusable, maybe you don't need an improvement over MIDI (It was perfectly adequate for mid-80's mainstream pop ).

If it'd seem unusable to me, I'd ignore it rather than ask how it can be best used .

If you want a certain value to control a certain parameter, you set a path for it in the receiving app/hardware, and set the same path in the sending app/hardware to send that value. Done.

Right. The issue we're talking about is that we haven't found standardization (yet) for transmitting MIDI-like data ('play a C4 now') over OSC. You need to do work to make sure the sending app/hardware and the receiving app/hardware understand each other. With MIDI, that's laid out in the spec, so roughly any MIDI-app can talk to any MIDI-app. You don't seem to have that right now with OSC - though of course it's by no means impossible.

I didn't mean to say OSC is unusable, just that it's promising but seems to be missing some pieces of the puzzle MIDI did/does have.

I haven't seen too many hardware controllers that had a very useful implementation of OSC. The most celebrated one, I think called Lemur or something, was all buttons, no sliders or knobs.

You're probably thinking of the monome, not the lemur.

[studio32]

You're probably thinking of the monome, not the lemur.

Are people in the linuxaudio world using such a monome?