CMJ Editor's Notes

The use of modern audio recording, synthesis, and processing technologies in the development of new performance interfaces and performer/instrument situations started early in this century-many cite the use of recorded bird calls in Ottorino Respighi's "Pini di Roma" (1924) as the first use of recorded music together with live instruments. Unfortunately, the typical paradigm for the mixture of old and new sound sources in performance remains essentially the one Respighi used--tape and live performer.

The use of modern digital tools for the construction of "active" (or even "intelligent") performance instruments has been underway since the appearance of the first analog music synthesizers in the 1950's, although most systems still oriented towards the creation of non-real-time tape music, or use performance interfaces taken from traditional instruments. In the 1970's, analog synthesizer users were generally either thrilled or distressed by the fact that the instruments built by Buchla and Associates had no traditional piano-style keyboards (as did most other systems of the day).

Several questions arise when reflecting on these developments; the most important of them relate to the physical (and cognitive) user interfaces used in traditional and new musical instruments, and to the possibilities of new performer/instrument relationships.

Many observers have commented on the extremely keyboard-centric nature of the MIDI interface, and the very traditional performance interfaces used on most commercial MIDI synthesizers. "Computer Music Journal" presented two topical issues dedicated to new performance interfaces in 1990 (issues 14:1 and 14:2), but few if any of these new instruments have found any broad application, and most remain research prototypes. Groups such as the STEIM Foundation in Amsterdam have dedicated themselves to the development and delivery of novel performance interfaces, but with little success outside of very specialized (i.e., small) user communities.

Why is this? Is the argument of the large "installed based" of expertise with traditional performance instruments going to impede the extension of instruments beyond what was available at Brahms's time? (I hope not.) Are the interfaces of traditional instruments really the result of good ergonomic design and an evolution driven by the needs of musical expression? (I would say "certainly not!") Is there any possibility of developing new instruments with the hope that, if they are good, they will find some larger audience? Is it going to be easier to make new instruments whose interfaces are derivations or extensions of traditional ones (as in the various extended keyboards under development or in use today), or to invent entirely new performance interfaces that are not based on the limitations of mechanical tools designed for the performance of dodecaphonic music? (I would vote for the latter.) What are the reasons for the poor reception of such novel interfaces as the "Video Harp" or the "Hands"-are they economical, technological, or aesthetical in nature? Is anybody really building new instruments whose design is driven by musical considerations?

The primary question I have related to the second theme mentioned above is what points or transitions are relevant to musicians along the continuum between completely passive instruments and autonomous (algorithmic or knowledge-based) performing machines? What do the various buzz words used in this area (e.g., "hyperinstruments," "intelligent instruments," "interactive composition systems") really mean, if-as I doubt-they mean anything at all (outside the marketing departments of companies or the public relations departments of institutes of technology)? Where does one draw the line between an instrument with a stored procedural response to pre-specified input (my understanding of "hyperinstruments"), and instruments which generate material based on user interaction with a model of their compositional algorithms (my understanding of "interactive composition systems")? Are there meaningful definitions of "composition" and "performance" whereby the new instruments and performance situations make sense (and the term "interactive composition" is made an oxymoron)? How should we classify a system such as Robert Rowe's "Cypher," in which the computer software actually models a "listener" component that responds to the characteristics (at some level) of the live performer's playing? If we had a meaningful taxonomy of these systems, how could it help us to develop more flexible human-instrument interfaces and more interesting computer-based musical instruments?

The trend towards electronic instruments with more complex responses to performer stimuli is evident, but it is not obvious that the developers of these devices are basing their designs on any useful insights into the evolution and current state of the interfaces of mechanical instruments, or on any meaningful models of the realm of possible performance paradigms and performer/instrument relationships.

	Vol. 16 Issue 3 Editor's Notes	CMJ Library >
	xx
	x The use of modern audio recording, synthesis, and processing technologies in the development of new performance interfaces and performer/instrument situations started early in this century-many cite the use of recorded bird calls in Ottorino Respighi's "Pini di Roma" (1924) as the first use of recorded music together with live instruments. Unfortunately, the typical paradigm for the mixture of old and new sound sources in performance remains essentially the one Respighi used--tape and live performer. The use of modern digital tools for the construction of "active" (or even "intelligent") performance instruments has been underway since the appearance of the first analog music synthesizers in the 1950's, although most systems still oriented towards the creation of non-real-time tape music, or use performance interfaces taken from traditional instruments. In the 1970's, analog synthesizer users were generally either thrilled or distressed by the fact that the instruments built by Buchla and Associates had no traditional piano-style keyboards (as did most other systems of the day). Several questions arise when reflecting on these developments; the most important of them relate to the physical (and cognitive) user interfaces used in traditional and new musical instruments, and to the possibilities of new performer/instrument relationships. Many observers have commented on the extremely keyboard-centric nature of the MIDI interface, and the very traditional performance interfaces used on most commercial MIDI synthesizers. "Computer Music Journal" presented two topical issues dedicated to new performance interfaces in 1990 (issues 14:1 and 14:2), but few if any of these new instruments have found any broad application, and most remain research prototypes. Groups such as the STEIM Foundation in Amsterdam have dedicated themselves to the development and delivery of novel performance interfaces, but with little success outside of very specialized (i.e., small) user communities. Why is this? Is the argument of the large "installed based" of expertise with traditional performance instruments going to impede the extension of instruments beyond what was available at Brahms's time? (I hope not.) Are the interfaces of traditional instruments really the result of good ergonomic design and an evolution driven by the needs of musical expression? (I would say "certainly not!") Is there any possibility of developing new instruments with the hope that, if they are good, they will find some larger audience? Is it going to be easier to make new instruments whose interfaces are derivations or extensions of traditional ones (as in the various extended keyboards under development or in use today), or to invent entirely new performance interfaces that are not based on the limitations of mechanical tools designed for the performance of dodecaphonic music? (I would vote for the latter.) What are the reasons for the poor reception of such novel interfaces as the "Video Harp" or the "Hands"-are they economical, technological, or aesthetical in nature? Is anybody really building new instruments whose design is driven by musical considerations? The primary question I have related to the second theme mentioned above is what points or transitions are relevant to musicians along the continuum between completely passive instruments and autonomous (algorithmic or knowledge-based) performing machines? What do the various buzz words used in this area (e.g., "hyperinstruments," "intelligent instruments," "interactive composition systems") really mean, if-as I doubt-they mean anything at all (outside the marketing departments of companies or the public relations departments of institutes of technology)? Where does one draw the line between an instrument with a stored procedural response to pre-specified input (my understanding of "hyperinstruments"), and instruments which generate material based on user interaction with a model of their compositional algorithms (my understanding of "interactive composition systems")? Are there meaningful definitions of "composition" and "performance" whereby the new instruments and performance situations make sense (and the term "interactive composition" is made an oxymoron)? How should we classify a system such as Robert Rowe's "Cypher," in which the computer software actually models a "listener" component that responds to the characteristics (at some level) of the live performer's playing? If we had a meaningful taxonomy of these systems, how could it help us to develop more flexible human-instrument interfaces and more interesting computer-based musical instruments? The trend towards electronic instruments with more complex responses to performer stimuli is evident, but it is not obvious that the developers of these devices are basing their designs on any useful insights into the evolution and current state of the interfaces of mechanical instruments, or on any meaningful models of the realm of possible performance paradigms and performer/instrument relationships.