I'm about to publish a (yet another) version of Debussy's Girl with the Flaxen Hair.
I've arranged this piece for tuba quartet, brass quintet (2 versions), brass septet, brass nonet, the Denver Brass instrumentation of 422.12, Boulder Brass 423.11, large brass ensemble 443.11, and now this version 444.12. Unless I'm forgetting one, that's 9 versions for brass.
After I finished the most recent iteration, in a moment of self reflection (or loathing, if you wish), I wondered "why can't I leave this alone?"
The answer is a little more interesting than boredom or just being out of ideas (which, I will admit is a growing concern).
Here's my premise: When a human being has an original thought, it is unique and pure and cannot EVER be 100% faithfully replicated. The original thought itself is the only pure version of that thought. When a composer creates a musical composition, even the process of composition becomes a series of compromises between his/her original idea and what eventually ends up on paper (or in the air). In this regard, improvisation is the closest a musician ever gets to delivering an original thought unencumbered by their own thought processes. Even that is a compromise between the impulse and execution - there are synapses, and nerve endings, and muscle tissue, and the instrument itself in the way of a completely pure vision. You might argue that those "impediments" are integral to the impulse and I wouldn't fight you over it.
When an arranger or orchestrator works on another human being's original thought, it is by its very nature a compromise, no matter how crafty or artistic. In science, it is axiomatic that even the act of discreet observation inherently changes the subject of that observation (closely akin to the butterfly effect). If you orchestrate for a large orchestra, there might not be that many basic compromises, but you cannot ever make an orchestra sound like a piano. So the stuff I do - now some 500 arrangements and orchestrations for various combinations of brass instruments - are all a compromise. Generally speaking, the smaller the ensemble, the greater the compromise. The larger and more complex the original work (scope and instrumentation), the greater the compromise. In either case, the more instruments I have to work with the fewer basic compromises (notes, range, register) I have to make.
That is a very simplistic statement. Sorry.
Girl with the Flaxen Hair was originally written for solo piano. Setting aside sustain and other possible extended techniques, vertically speaking that means at any one instant there could be as many as 10 notes. If my ensemble is less than 10 voices, those maximum-utilization-of-available-resources-moments (we're talking digits here) will result in a compromise - I'm going to have to leave out a digit.
So when orchestrating Debussy or Rachmaninoff for brass quintet or, gulp, tuba quartet, you are merely suggesting the music. That takes a hell of a lot of skill to suggest effectively - to make good thoughtful compromises and craft it so well that it strongly suggests the original. But no matter how crafty, it's still a long way removed from the original intent of the author.
Doesn't make it bad - just different.
I'm currently in my "large brass ensemble period" - I think it might be due to limited bandwidth (union stuff, kid stuff, life stuff). Making good compromises for smaller ensemble requires concentration and continuity of effort. But I'm also attracted recently to a much larger sound palette. Even these large instrumentation charts have a lot of subset chamber music moments, and that pallette allows for a lot more recreational (at least for me) color experimentation.
And fewer compromises.