Disseminating Otto Laske’s

Algorithmic Compositional Methods

 By Michael Rhoades

Otto Laske’s approach to algorithmic composition is the culmination of many years of research, collaboration and development. Through the benefit of his experience, young composers of electroacoustic music are challenged and encouraged to go beyond it, continuing and extending the development of his dynamic approach to future generations. 

One of the issues composers  of algorithmic composition have to deal with is how to instill a sense of musicality into their compositions. Algorithmic compositions can sometimes seem to sound rather lifeless and random. While they may be interesting on an intellectual level, the other half of what makes a “good” composition, the musicality  is sometimes lacking. Laske’s processes tend to naturally engender a musical approach to algorithmic composition as well.

Generally, a score is massaged and complexified by means of several iterations of manipulation, through the creation of subscores, at various stages of the compositional process. Instead of generating one score by algorithmic compositional means, several layers of subscores are created, the results of which are collected to be mixed in a timeline format at the final stage. This enables the composer to create families or classes of sounds, establishing a relational quality to the various scores created and therefore to the various sections and layers of the overall composition. 

In order to disseminate Laske’s  approach to algorithmic composition, a combination of Csound, Cmask and a spreadsheet program were the determined to be the tools of choice. These applications are universally available and provide a robust feature set for creating and implementing a model approach. The following criteria were determined to be guiding factors on which a workshop of Laske’s approach would be based:

• The process would be simplified enabling students to compose a complete piece in a one week long workshop setting.

• The overall approach would provide a base process from which further exploration and development could genesis, thus empowering the student to continue developing his or her own process.

• Although a predefined approach to the algorithmic compositional process would be directed in the course, many variations of the steps presented should be possible and likely to ensue.

• The course would be easily expanded into a longer time frame making it possible for it to be used in a university curriculum.

• The focus would be to disseminate Laske’s approach, not to teach an in-depth study of the applications utilized.

• The workshop structure would be designed emphasizing content over form. 

The basic process is plotted in three steps:

Step One: The students are provided with many samples from which they are to select four sets of four. Each set or “class” of samples should be related by a common element such as timbre, percussive quality, etc. Samples are approximately five seconds long and of comparable amplitude. After four classes are chosen, the students begin to mix samples within each class to create several new samples, approximately ten to fifteen seconds in length. This is accomplished by utilizing a Csound .orc file that is provided and a base Cmask file. The files will be manipulated and altered to create Csound .sco files, which in turn will be used to render and therefore create the new samples.

Step Two: The students study and categorize the samples created in the first step and rename them as soundin.x in an order of preference, keeping notes with information about their characteristics. The samples from one class are then combined with samples from another class, depending on the student’s preference, to create samples of approximately one minute in length. After several samples are satisfactorily created in this configuration, a different combination of two classes is utilized, creating samples which extend the original classes. This is facilitated by the use of a second Csound .orc and a second base Cmask file. The second .orc file is more complex and versatile than the one used in the first step. It introduces several modifiers (effects) that may be included in the sample creation as well as offering a selection of several panning algorithms.

The scores created by Cmask are further manipulated by the use of a spreadsheet program. Transposition, retrograde and general altering of the score is possible through the use of formulas in the various pfields. Vertical, as well as horizontal, editing of the data adds a distinct layer of manipulation ability not available with a standard horizontal only text editor. Also, as Laske mentioned earlier, Cmask cannot create variants of the same score.  However, a spreadsheet program makes easy work of such tasks and thus affords the ability to create related but different scores from the same base score.

Again, the new samples are classified, organized, studied and renamed to make ready for the next and final step.

Step Three: At this point, the student should be familiar with the body of samples that has been created and should have a basic understanding of how they fit together to form a cohesive composition. The final Csound .orc file is simply a sample playback opcode, with a few twists, allowing the composer to insert any of the samples thus far worked with into a timeline. This includes the original four sets of four samples and the samples created in steps one and two.

The composer is provided with several sample manipulation tools to facilitate final mixing of the work. At this point in the process, this could be seen as another form of score manipulation. Specific sections or entire samples can be inserted into the timeline and connected in series, overlapped or layered. Further, the ability to play samples back at different frequencies, forward or backward as well as several other manipulating features are available. Through critical listening, in conjunction with following the playback of the piece in the score, the final assembly or mix of the piece is accomplished. A global reverb instrument is also offered in this step to help give a sense of continuity to the work. This is especially helpful for smoothing the transition between disparities in sample content.

Implementation of this approach to Laske’s process is yielding remarkable results. Utilizing the computer as the composer’s alter ego to implement a top down approach to algorithmic composition, along with a process of multi-layer subscore manipulation, allows the composer to achieve compositional results that are unique and innovative while maintaining a musical aesthetic. As such, it is a viable starting point for continued development in the area of algorithmic composition. This workshop was designed to offer students the opportunity to begin to integrate algorithmic composition into their own compositional processes. It exemplifies a vast territory yet to be explored and developed, truly bringing algorithmic composition into the realm of musical composition for the 21st century.