Announcing talon-organteq for hands-free control of registration

Since I introduced this project a couple weeks ago, I've made considerable progress in developing this into something much more powerful than I had originally envisioned.

It's now a sophisticated and fully customizable registration system. It's ideal for use with voice control, but that's no longer a requirement; you can control it either from the command line in a Python interactive session, with custom mappings on your MIDI keyboard/controller, or with voice control. Or with some combination of all three methods if you like.

Basically, with this system you have a knowledge base and a set of rules that you construct. Some basics are provided for you (such as mapping stops to their tonal families -- flutes, strings, reeds, etc), but there's no limit to how far you can customize and build on it from there.

I believe this system has important implications for:

• New music composition and improvisation: you can achieve more nuanced, varied, and spontaneous registration effects

• Education: nice tool for learning about principles of registration and building one's intuitions around them. For example, you could construct rule files demonstrating French classical and German baroque registration principles, and apply them to different Organteq presets. Or you could work on building up and codifying your own aesthetic principles.

• Personal MIDI organ setups where budget constraints are a limiting factor -- because this system means you no longer need a bunch of hardware switches, or even any at all. You may even be able to use couplers in creative ways such that a two-keyboard setup could give you  a flexibility that feels like having three or four keyboards.

• Organization: you could set up your rule files in any way you like to help you stay organized. For example you could have a file tocatta_and_fugue_bwv565.pl that stores your registrations that you've designed for that specific piece. (Note: yes you can also do this with FXP files natively in Organteq. But the method proposed here allows some additional advantages, such as being able to store notes side-by-side with your stop specifications, e.g. you could have a comment "% for the solo voice in measures 8 - 16" or something to that effect -- where the percent sign denotes a comment.)

Admittedly it's less compelling for performance of historical music when one is going for authenticity, since this is far more powerful and more convenient than anything Bach or Couperin, for example, had available to them. But then again, they also didn't have the ability to go from the Freiberg cathedral to the Royal Chapel Versailles at the push of a button like we can do with Organteq.

----- What's next? -----

I'm working on writing a paper on this, perhaps targeting New Interfaces for Musical Expression or Computer Music Journal. Towards that end I am seeking:
- a collaborator or co-author with expertise on organ registration, composition, and/or musicology
- users willing to give it a shot and share your impressions and/or notes about what you've been able to accomplish with it, especially if you do any composing or improvisation

How can you get started trying out this system?
First of all you have to have some comfort level with the command line and with at least being able to edit Python code. (You don't have to be totally fluent in Python.) Take a look at this script, which shows how to map faders on an Arturia KeyLab MK3 to control crescendo effects. You would have to adapt this code for your particular MIDI controller, but in essence: these crescendo effects, defined in crescendi.pl, are modeled after Modartt's built-in crescendo pedal settings for a couple of specific presets (Baroque Cathedral Freiberg and Neo-Classical Church), but with two differences that highlight the power of this new system:
1. Instead of applying the crescendo effects across all manuals at the same time, you can do just one manual at a time (or at even finer granularity if you like)
2. You can define crescendo effects like this that are extensible to any preset, not just one (as is done here).

Anyone interested in getting set up with this and trying it out, or in collaborating in some way on the paper or on a workshop demonstration or something, please feel free to reach me by PM.