On My Compositional Process

From the outset, my compositional process has been algorithmic. Even before using actual computers, I planned my compositions "top down", by first deciding upon and then computing compositional materials in numerical form. Based on a design, I then developed the resulting "alpha score" in two different ways, depending on the musical medium chosen:

instrumental, vocal, and chamber music: the numerical score is interpreted for acoustic instruments and/or human voices
electronic music: the numerical score is realized by way of electronic instruments and orchestras of my own design

In both cases, a large margin of freedom existed, in two different forms:

In acoustic composition, I was left with a large margin of interpretative freedom as to how to use the specific instruments or voices chosen, as well as the timing of events.
In electronic composition, I was sounding the score through instrument groups, or "orchestras", of my own choosing; these orchestras could also be overlayed (mixed) with each other, either harmonically or contrapuntally.

The algorithmic score thus had two beneficial effects: while it provided me with a compositional "blue print", is also left me with a large measure of freedom for using my intuition. Moreover, I myself decided what was to be left to the computer to decide, and what the constraints should be that I would follow.

In my later electronic compositions, after 2000, I used computer scores very freely, often mixing them. As a result, they were acting more as "sound teasers" than syntactic strictures. In this way, the initially strictly followed syntax of scores was relaxed under the influence of purely auditory intuitions of mine. "Algorithmic composition" gave way to composition following the rules of the composer's ear, not any computer program. This was an expression of having internalized the logic initially embodied "objectively", by a computer program, which had become entirely my own.

The slide set that can give deeper insight into my compositional process can be viewed here in the pdf entitled Koenig's Project One.