Generative Music and Distribution

I have been working on some iOS audio programs lately and the topic of generative music came up for me. Some 10 or 11 years ago, I remember thinking about the issue that generative music–or music that was not fixed in performance–did not have a means of content delivery. At that time, I imagined some form of meta-music player that would have plugins that could read Csound, PD, and other kinds of projects, yet could be queued up in a playlist much like you would find in something like WinAMP or VLC Player. The idea was that composers could then distribute their works in a standardized format, and consumers could then queue it up to listen to in the same manner they might load up an MP3 or CD.

Back in the mid 90’s, Brian Eno had done some generative music work with the Koan Music Player, releasing an album that would be performed differently each time by the player. The Koan Player though never really caught on, and since then the format and program have died off. What a shame it is to me that musical work can be tied to the life of a commercial program.

These days things like Bjork’s Biophilia for iOS start to renew an interest in generative music. However, the distribution scheme here is a custom application for a closed platform (iOS). I wonder too if the kind of thing that happened with Eno’s album and Koan might not happen once again with custom formats and distribution means.

What I would love to see is a generic means of distribution for both fixed and non-fixed works. Such a system would allow meta-data to describe what plugins would be required, as well as what hardware requirements would be necessary. So, for example, if your work required an 8-speaker octagonal cube or a video camera feed, that could be marked up in the distribution meta-data. That way, the distribution format would then be usable not only for consumers at home, but also as a means for concert delivery or generic installation setups.

Thinking about this all again, I think the idea still has merit and that developing a standardized distribution package could expand the audience for such work as well as create a platform that promotes longevity of work.


  1. Hello. I am surprised to see that “Bloom” is not even mentioned. I think that “Bloom” defines by itself the concept of an app-that-is-a-composition, the perfect way to distribute a generative piece of music.

    1. I did not mention Bloom as I was not aware of it. I took a look at it just now, and it seems to have the same problem as Biophilia would have, in that it works for a limited platform for now, but it’s difficult to know how it would do for the future. By being a custom application, it certainly offers many possibilities, but since it’s only for iPhone and iPod Touch, it is limited in scope. I would like to see something more like MP3 or FLAC, something of a format that could be opened on multiple platforms.

  2. I have long come to the conclusion that it can not be done, not without a decent software synthesizer written in a platform-independent language (Java perhaps). This is the reason i was bothering you with questions about a Java synthesizer, a couple of years ago if you recall. I have not managed to program anything particularly interesting, but i am quite convinced that this is the way to go.

    On the other hand, iDevices have proven to be extremely successful, so if the primary goal is to reach a big enough audience, programming the thing in Objective-C should not be disregarded as an option, despite my personal preference for more “open” standards.


  3. Well, there’s been a fair amount of Csound work I’ve been involved with for getting it working on iOS and Android. The iOS build works, and I’m in the process of getting Android working. This should open up cross-platform music projects that work across the board: desktop, tablet, phone, etc. We’re also looking into browser-based Csound now as well. I think it will be quite interesting in a few weeks! 🙂

  4. I think the answer of delivery actually lies with single board computers. I am working on an album of generative music, using PD, that I plan to load onto a modified Raspberry Pi (or a similar device) The songs would be fixed compositions, but would be different with every play, based on decisions the computer makes in real time during playback. The user interface would only consist of buttons for play, stop, and a simple system for song selection. This package would then become the final product; a mobile playback device specifically configured to play my album. Future albums would be as simple as swapping out the SD card.

    That’s the goal, anyways!

Leave a comment

Your email address will not be published. Required fields are marked *