Round-robin writing

[Otto Halmén]

Not to be confused with round-robin sampling.

This is an idea I've had in mind for a while. Granted, it has it's issues. Namely, the scoring template has to consist of a DAW, libs and plugins that everyone can agree upon and get their hands on easily. But it's not entirely impossible in theory.

What do you think?

An example would be REAPER (or even LMMS) + VSCO 2: Community Edition + Hibiki Reverb. Each participant gets something like 30-60 seconds of music to lay down. The order of participation could be drawn or agreed upon, or each participant could get to pick who gets the next turn until the list reaches the end.

Just throwing it out there.

[Mattias Westlund]

Cool idea, and the resulting music could be really interesting!

The main hurdle I see is that there's a lot of personal preference involved in working with DAW's and samples. I doubt that many people would be willing to spend the time to learn a completely new DAW in order to do a project like this, fun as it may sound. Also, we all have different ways of working and setting up our templates. I for one rarely use keyswitches while others use them all the time. And so on.

So it kind of depends... if there was a clear and well thought-out plan in terms of setting up the base template, where everyone interested could have their say, then maybe it could work. For me, switching to another DAW is kind of a deal breaker though. Especially if it's LMMS

[Mattias Westlund]

There's also this: how do you make sure that the project template loads the correct samples on the target system? How many free samplers support relative paths? And what about moving between platforms? Mac -> PC or vice versa might not work at all.

It's easy enough to try though I suppose...

[Otto Halmén]

Hmmm...

One option would be to simply agree upon the sounds used. Let everyone set them up in their own DAWs and use bounces to pass the ball. This solves the issue with different DAW users, but completely shuts out those who are more visually oriented in their composing. It's a bit hard to read a plain audio file with your eyes.

Another one would be to pass MIDI stems. Still, there could be hurdles. While keyswitching and separate articulation tracks might work side-by-side in a single project, things like Logic's Articulation ID probably won't translate smoothly, should someone insist on using that.

An obvious option would be to do all the writing in a notation program like MuseScore, and just have someone do the mockup on their own. That, however, shuts out a lot of people. Reading notation (let alone composing through notation) isn't exactly universal, especially among all self-taught composers.

This whole idea is definitely unripe. I like the thought of getting out of the comfort zone, though. You'd have to trust someone else to carry on with something you started, and limit yourself to a predefined DAW and template.

Just have to figure things out a bit, I guess.

[Mattias Westlund]

(09-02-2016, 06:54 PM)Otto Halmén Wrote: This whole idea is definitely unripe. I like the thought of getting out of the comfort zone, though. You'd have to trust someone else to carry on with something you started, and limit yourself to a predefined DAW and template.

Just have to figure things out a bit, I guess.

Yeah, you're going to have to put the thinking cap on. I'm not saying it can't be done -- actually I hope it can be, because it sounds like a lot of fun! -- but spontaneuously, I see a lot of potential pitfalls. From my perspective, pure midi or audio stems are useless. You wouldn't be able to put together something worth listening to from that. And it seems like a lot of work.

For the sake of argument, the most universally available tools for the purpose are... what?

DAW: REAPER? Tracktion 5 (completely free and capable but unorthodox)?

Sampler: Sforzando? TX16Wx?

Library: VSCO2? SSO?

Reverb: Hibiki? Sanford Reverb? (less important though, we could always render the final track with a single agreed-upon verb and let everyone work with what they have)

[kmlandre]

Maybe a combination of dry, rendered tracks and midi data?

That way, for us note nerds, we could import the MIDI data into our notation software of choice to ferret out the themes and such, but still preserve the "sonic intentions" of the original.

-- Kurt

[Samulis]

This sounds like a great idea, I think you guys should go forward with it.

A few years back I participated in a similar collaboration, in this case there was a common key, tempo, and number of measures each person got. We each rendered a piece and left the reverb tail on the end to decay out. Unfortunately I can't find the piece now, but I know it's out there somewhere.

The big downside to the approach was a major lack of continuity, as some people would do electronic stuff and then there'd be an orchestral bit or something, etc.

Perhaps the best approach would be to do some sort of 'free-for-all' collaborative composition. I've done a few of these with Simon Dalzell (Ivy Audio) and a few other folks using VSCO 2 in Ohm Studio (free)-
https://www.youtube.com/watch?v=9uB0a_jmI0Y

In that case, we did the division horizontally, so we each picked a few instruments. It was pretty crazy, but we had fun. It would be very possible to do the same thing vertically too- load up a "template" in Ohm Studio and give people an ~8 measure range to write in.

[Michael Willis]

I like the idea of agreeing on samples, plugins, and such, and then to start out, collaborators just sent exported midi files back and forth. Later when the structure is better defined and it's time to work on finishing touches, the tracks can maybe be shared as audio as well.

If VSCO is used, midi cc events can control articulations (using the *_full.sfz files), so there is a fair amount of information that can be encoded as just midi, which affords the possibility for each collaborator to use a preferred DAW.

Mattias, could you share more detail about your objection to collaborating via midi? To me it seems like a lot of what the DAW stores is just midi.

I also suggest doing horizontal slices as well as vertical - rather than feeling like the first person must fully score 30 seconds, and then have the next attach 30 more seconds, what if one just starts with a melody, the next person fills out the harmony underneath, the third extends it with a variation, and so on?

Another possibility would be to split it up completely horizontally - one person owns the string sections, another person owns brass, and a third owns woodwinds. I'm not sure I like that one as much, I would rather be free to contribute to any section, but I'm just brainstorming at this point.

EDIT: Looks like I was thinking of the No Budget Orchestra, rather than VSCO. NBO has the *_full.sfz files that allow you to switch articulations with midi CC events.

[Mattias Westlund]

(10-03-2016, 03:00 AM)Michael Willis Wrote: Mattias, could you share more detail about your objection to collaborating via midi? To me it seems like a lot of what the DAW stores is just midi.

TBH I don't remember exactly how my thinking went, but from the top of my head the main problem with exchanging midi clips is that DAW's handle midi importing and exporting very differently and the results can be unpredictable. User A exports his part from DAW X and sends it over to User B. When User B imports the file into DAW Y, he gets forty tracks named MIDI_01, MIDI_02, MIDI_03 and so on, with now idea what is what or what goes where.

Admittedly I haven't done a lot of importing/exporting midi the latest decade though, and maybe DAW's today handle it better than in the past.

[Michael Willis]

Maybe it would be worth doing some experiments about how different DAWs handle imports and exports of midi. It seems like I've had reasonable results from Ardour, but I'll have to try it again to see what happens, for example, if I have one track with multiple midi channels. It would certainly be a big pain if those imported as multiple tracks with useless names for somebody else.

Just curious, what ways have other people used to switch articulations on a single track? I use a combination of midi CC events for soundfonts that support that, and using multiple channels for those that don't. If this is going to work out at all, we would have to come to some agreement about such things.