Brief History of Machine Improvisation

(above) Performance with George Lewis' Voyager improvisation machine
(above) Demonstration of “Piaf, Holiday, Schwarzkopf: tre donne del 1915” by George Bloch and Hervé Sellin, using ImproteK and Omax-video, which are variants of OMax.
(above) Improvisation of Isaac Schankler and Mimi, using a piano roll interface to visually represent both the current state and future anticipated actions of the improvisation machine.
The creation and performance with improvisation machines date back to the late 1980s. Examples of pioneering works include the following:

     > George Lewis’ “virtual improvising orchestra”
        Voyager (late 1980s)
     > IRCAM’s OMax software created in MaxMSP
        (late 1990s)
     > Benjamin Carey’s semi-autonomous software
        system _derivations (~2016)
     > Guy Hoffman and Gil Weinberg’s Shimon, a
        robot-marimba player (~2011))
     > Alexandre Francois, Elaine Chew, and Dennis
        Thurmond’s Mimi, a multimodal interactive
        music improvisation system (~2007)

Why improvisation? Why machine?

"Because the machines were made by humans too—their character is in there, their senses are there, and when the machines produce these pictures, they become close to our inner pictures."
- Pipilotti Rist (Louisiana Channel 2019).

Creating improvisation machines results in us learning more about our tendencies and biases as [improvising] musicians. It is not asking “whether machines exhibit personality or identity, but how personalities and identities become articulated through sonic behavior” (Lewis 2000).

By allowing the machine to create a representation of our identities as improvisers, we can achieve a distance to ourselves that is productive for the analysis and discovery about our own playing.