Part of Temporal Constructs Appendix, which contains details behind the research: dates, quotations, descriptions of instruments, and sources.
General definition
Relation to the practice
The masterclass is the one teaching activity inside the institution so far, and it is evidence for transferability as much as a teaching record. It is also the statement of its argument about instruments that the practice has made in public. Its subject is the shift from acoustic instruments to electric, then electronic, then digital ones, and what is lost and gained at each step. Its demonstration lets a cohort of students use the software of the MIDI era and play it.
Really enjoyed preparing and presenting yesterday’s masterclass “In Conversation With Your Instrument” at Inholland conservatory Haarlem. We discussed how to relate to the concept of artificiality in a performance context with electric, electronic and the most artificial (and human-like) of all, digital technologies.
The room was prepared with a quadraphonic 4 speaker system for M’s four voices, allocating each “player” to a corner of the room.
Not exactly the lightest setup to bring by train, but it performed well, and the two hours passed way faster than I realised. Great to connect with the students, and to discuss topics from the magic of analogue/digital instruments to generative AI.
artist statement
Preparation of the masterclass
The masterclass was prepared in February 2026. Its proposal was written then in Dutch and English, together with the two versions of its technical rider. That rider tells a venue what the performance needs: quadraphonic speakers, projection and DMX lighting among the requirements. Its other texts came from the work. A six-minute trailer text came from the ALTERITY performance script.
Findings from the masterclass
The first part of the masterclass takes constraints as creative features. A backing track replaces one listening experience with another. An audience listens differently when it knows a person is playing in front of them. A tape loop gives the machine independent behaviour but leaves it unable to respond, so only the person answers the machine. The optigan and then digital samplers turn recorded material into something played, and memory brings modularity and programmability, and with them interactivity.
Analogue instruments tend to keep no gap between the sound source and the performer. Digital instruments conceal their operation as they grow complex, and an instrument that cannot be understood cannot be used expressively. The masterclass’s position is that a clear understanding of an instrument’s constraints enables creativity. It holds that the working conditions are unfamiliar within familiar boundaries. Suzanne Ciani opens it, having asked the same question about her studio and answered it with a Buchla system.
The second part follows the change from analogue to digital and the arrival of memory. Vangelis is shown here in two clips, one objecting to what digital instruments had done to playing and one playing the system he built to answer it. Instrument developers programmed character and unpredictability into digital instruments deliberately, and programming began to resemble tuning an instrument before playing it. Presets were replaced by stored compositions, so performance became a matter of deciding when the machine plays. Memory also makes the distinction the masterclass draws between a tool that is used and an instrument that is played, with play as what makes the result reflect the process.
Its third part is the intelligent instrument. Composing writes a score for later, and improvisation happens once. Stored musical material raised the question of playback with variation and of a system answering what it hears. The answers were instruments with behaviour of their own, built to surprise their player, where later music software settled on tools that follow instruction.
Laurie Spiegel is the figure at that transition. She spent the 1970s working with systems that responded in real time, before those systems began to propose. She then built Music Mouse, a clip of which the masterclass showed. M is the example, described by its author as the cockpit of an aeroplane.
Demonstration of M
M ran on a Macintosh SE of 1987 and made no sound of its own, sending instructions instead to a Roland S-330 sampler of 1988, which generated the sound. Its four voices were mapped to four speakers for quadraphonic sound. A light system made its decisions visible, so what the sequencer decided could be seen as well as heard. A wheel gave direct control of dynamics, of density and of brightness, its role set from the keyboard and stored as presets. M took the notes last played and sent them down up to four paths of its own.
Voice interface for M
M’s interface can be driven externally from a piano keyboard, and here it was driven by the voice. A model transcribed the speech to text, and a second model judged whether the input was a human voice and reported its confidence. A natural language system extracted the words and their values before the result was converted into the data M expects. Commands such as starting and stopping the sequencer and muting a single voice then worked as spoken. Both models were open source and ran locally on an ordinary laptop, with no subscription and no data leaving the machine. The same control system later let agents operate M directly, which the record of results counts as an agent-initiated use of the instrument.
Outcome of the masterclass
The masterclass closes on the point the research rests on. Tools offered by large companies can still be taken and made one’s own. They are used in ways their makers did not intend, in the way an instrument is adapted by whoever plays it.
Unforeseen
Here the unforeseen situations, sometimes difficult, sometimes unexpectedly interesting, will be described as the project runs. Many of these unforeseen situations have already occurred and have been recorded. They will be added in the time to come.