Once the five themes and Other were composed in both solo piano and 16-member ensemble and solo piano forms, the SONOR ensemble at the Department of Music of the University of California, San Diego, and the pianist Jean-Marie Cottet of the Paris Conservatory were scheduled to make recordings. The more complex ensemble recordings were done in Warren Studio A at UCSD over a three-day period. The recording protocols were devised in order to provide opportunities to examine aspects of the process that might prove useful at later stages. My composing itself was, in fact, affected by the posited requirement that the thematic materials should be quickly grasped and economically mastered by the musicians. I used tempi, barring decisions and other organizational as well as notational features to achieve these ends. Rapid tempi allowed for rhythmic relationships to be established with fewer beams. Careful thinking about the metrical consistency and placement of events within metrical contexts aimed to maximize the immediate precision of the ensemble.
On the first day, the musicians assembled, having been asked to study their parts beforehand. After very brief administrative comments by conductor Harvey Sollberger, he began to rehearse Theme 1. Progress was rapid, and this element was soon recorded. The first recordings of each theme were made without any input from the composer and no interpretative comments from the conductor other than clarification of explicit relationships. I then went into the rehearsal space and described what I would like the musicians to give more specific attention to. A particularly dramatic effect arose from one of my explanations and interpretative metaphors in relation to Theme 2. The ensemble was having difficulty sharpening the dynamic differences as much as I wished, and I used that metaphor of "distance". I asked them to think of very soft events as far away, and very loud ones as proximate. This had a remarkable impact, improving results immediately.
After all five themes had been recorded (Other did not have an ensemble version, as discussed above), that evening and the next day were spent editing and evaluating the first day's results. In the interest of maximizing the psychologists' access to different component layers of the ensemble music, we planned to use the initial, edited, whole-group recordings as reference guides, to be played back into the musicians' ears as they re-recorded their parts again the third day, family by family.
After some initial patching up of Theme 5 the third day of the San Diego sessions, it was possible to record several of the themes in familial layers: woodwinds, brass, percussion, and strings. In addition to these horizontal "slices", the core elements of all themes, and all of the subsections of some were recorded as separate units, so that, both with the computer transformations and in psychological experiments, it would not be necessary to artificially create beginnings and endings editorially. Each instrument was individually close-miked and placed on its own track, so that balances could later be adjusted. In addition, baffles were placed between each of the instrumental sections in order to maximize separation, and reduce cross-talk between the sections.
Recordings of the piano versions of the themes were easily accomplished with the exception of Other. The technical difficulties mentioned above proved too formidable at anything close to the tempo I had originally intended. Yet, the overall effect of the section was so persuasive that I decided it should be recorded at a slower speed, that I would face the consequences of this change at a later time (see RECONSIDERATIONS, below).
Other portions of this publication will detail the purposes to which the psychologists put these initial recordings. More than a year passed from the time they were made until I began actual work on the full score. During that period, there was extensive testing that focused upon the degree to which the themes were heard as distinctive and independent from one another. But there was another crucial purpose in making both the ensemble and the piano recordings (The latter were made in the Espace de Projection at Ircam.) They later served (see THE COMPUTER PART below) as the raw materials on which the computer aspect of this work was based. In some instances, the total, massed ensemble recordings were used, but in others (e.g., D10) it proved valuable to have the recordings of individual sectional layers with which to work. The processes used in digital transformations will be discussed below.