In the 1970s and 80s I started developing techniques for expanding the amount of musical information that might go into a short time frame. I soon realized that the resulting work would require a different sort of approach for the observer, focusing on a very short time frame without a necessity to remember past events or to anticipate those in the future. In practice, I found that it made listening to complex patterns an effortless activity. I started referring to this as "present-time listening."

To provide musical material to satisfy this new listening mode, I started developing transformation techniques for use in fragmenting and recombining pitches. In a manner similar to what I imagined the Impressionist painters had used, I wanted to define coherent images from small bits of color and shape. Instead of making harmonies that moved from chord to chord through passing tones and voice-leading rules, I worked on methods that could start with a collection of pitches relating to a stable consonance, then rapidly vary the ordering of the set many times while replacing individual set members, until a new stability would be reached. For me, the music was all in the transitions.

All of the preceding was worked out in the Rube Goldberg environment of that era of electronic music - homemade circuit boards branching out to analog synthesizer modules and half-controlled by early microcomputers. It was a wonderful dead end of hardware acquisition, but a dead end still. By the mid 1990s, I was looking for ways to recreate the workings of the physical electronics but within a more flexible software environment. I tried rebuilding one of my custom circuits in computer software, and found that it would give me an even greater variety of evolving patterns than in my previous work. Sebastian’s Shadow was made at this stage, using a short Bach fugue as the resource for transformation. I had never dealt with historical music in the context of my own work, and it felt like I was going backwards and forwards at the same time.

The original fugue, known to all organists as that from the Fantasia and Fugue Nr. 4, BWV 542, is only about 4 minutes long. I started with a computer-encoded version of this and first elongated, layered, and segmented the entire fugue, using temporal and proportional formulas that I had developed for the piece. Each layer was then subjected to the black box modeled after my old hardware. All the resulting transformations resided as streams of encoded note information – really just representations of the pitches. These were converted to actual sounds using simple oscillators, with only slight variation of brightness and sonic motion tied to individual pitch choices. I just wanted to let the patterns speak as clearly as possible.

The title contains a double reference: The first, obviously, to Bach, and the second to my brother’s dog, who will always be remembered by our family.