Azimuth – page 4

Back Next

B. The Process

The process utilized for the composition of Azimuth stems, in part, from a plan for a workshop on Algorithmic Composition developed in partnership with Otto Laske. This workshop was meant to disseminate Otto’s process of “score based sampling” involving Kyma and either Cmask or Project One. The thought was that not every potential venue had a Kyma system available. Therefore, since they are universally available, I developed a three-step system, utilizing Csound in conjunction with Cmask designed to emulate Otto’s approach.  

The basic idea was to give the participants four samples, each approximately six seconds in length. They were then to mix these samples in various combinations, using Cmask to generate the scores, to render samples of approximately 15 seconds in length. The next step was to combine these 15 second long samples in various ways, again utilizing Cmask and adding a few effects, to make 90 second long samples. The final step was to mix any or all of these samples into a composition.  

While at first this approach seemed very simple, the power and expressiveness were soon revealed. It became obvious that this was a very subtle and yet dynamic approach to algorithmic composition. By the time work began on Azimuth, several pieces had been composed, each of which began with this basic approach, expanding it and taking  it to new and greater depths. The possibilities associated with this very basic premise are endless.

The extension of this approach for Azimuth was to create a schema to organize the process of rendering the samples. Instead of four samples the intention was to begin with sixteen; four groups of four. Also, recent reading about how Koenig had formalized a mixing technique for three samples in the composition Essays sparked an interest in utilizing a variation of this approach for Azimuth. Of particular interest was the idea of morphing the samples from one to the other. Ultimately, the extension of this idea was to create four samples, considering them the base sample of each of four groups. Group One was comprised of base sample one, base sample one morphing to base sample two, base sample one morphing to base sample three and base sample one morphing to base sample four, see Figure 1. It became evident as the planning continued that each sample group adequately represented the potential for a different movement. It was surprising that, even at this stage, the form of the piece was beginning to solidify. This initial “out of time” work, to use the vernacular of Laske and Xenakis was seriously and directly impacting the “in time” composition. After the base samples were established, as with the workshop, a set of 15 second long samples and finally a set of 90 second long samples were to be created. As the final step, all samples would be imported to and subsequently mixed in a DAW environment utilizing multi-track editor/recording software such as Digital Performer or Sonar

*The sample numbers in the gray cells are links to an mp3 of the sample.

Step 0 – Create 1 – 3 second samples (16 samples created)


Sample names

 = Step-Mix-Group-Sample (example 1111)


Group 1







Base Sample #1






Morph group 1 base to group 2 base





Morph group 1 base to group 3 base





Morph group 1 base to group 4 base




Group 2







Base Sample #2






Morph group 2 base to group 1 base





Morph group 2 base to group 3 base





Morph group 2 base to group 4 base




Group 3







Base Sample #3






Morph group 3 base to group 1 base





Morph group 3 base to group 2 base





Morph group 3 base to group 4 base




Group 4







Base Sample #4






Morph group 4 base to group 1 base





Morph group 4 base to group 2 base





Morph group 4 base to group 3 base



 Figure 1: This first section of the Schema exemplifies the original sample designations for the four groups.

Go To Page:


Back Next

Comments are closed.