CARL System published on 15/10/2011 da Alex di Nunzio
The CARL System is a computer music project developed in the early eighties at Computer Audio Research Laboratory (CARL) at the University of California at San Diego (UCSD). It was essentially a software package dedicated to real time computer music composition and performance. Inside the CARL system, Cmusic is considered the most important software.
Background history - In 1979, Richard Moore (who had worked with Max Mathews for the development of Music V and the GROOVE System) moves from the Bell Laboratories and went to the University of California at San Diego (UCSD), where, within the Center for Music Experiment (CME, which later became the Center for Research in Computing and the Arts - CRCA), founded with Gareth Loy the Computer Audio Research Laboratory (CARL). Within this new center, more specifically dedicated to computer music, Moore and Loy realize the project that became known as CARL System, also known as CARL Software Distribution, a software package dedicated to composing music through the computer . The project, developed through funding from the Ford and Rockefeller Foundations, the National Endowment for the Arts and the UCSD, also participated Julius Smith and Rusty Wright.1
Goals - According to the wishes of the founders, the CARL System would have to standardize the management policy of the CARL Laboratory. This meant make tool flexible enough to be applicable in the field of composition, teaching and research. As regards the composition and performance of computer music, the CARL System was designed to meet different needs. The first goal that put the researchers was the composition and performance in real time. Alongside this there was the need to avoid a work environment that suffer the limitations arising from the rapid obsolescence of the hardware. While sacrificing the computational speed, then, the researchers chose to focus the project on the software system, so that was easily implemented on different machines. In this regard, then, the CARL System was designed under Unix and C programming language, thus ensuring more portable and, therefore, greater longevity. The CARL System, overall, is emerging as a tool of a fairly general structure, likely to be applied in different contexts of computer music, while being able to meet very specific needs.
Teaching and Research - In addition to questions concerning the composition and performance, there are also targets of slightly different nature, such as teaching and research. In the first case, the CARL System was expected to facilitate the study of the techniques of sound processing, synthesis, performance and computer programming in general. The decision to adopt the C programming language, using macros, should facilitate the approach to computer music for those that do not have a long experience. The research, however, was addressed issues related to the psychoacoustic and design of real-time systems.
The system - It is said that real time was the first goal set for the CARL System. Since no computer in those years was able to ensure adequate computational speed, the whole system was designed to work using three different computers, each dedicated to a specific function. CARL, in fact, first of all was provided by a computer that hosts the software for synthesis, analysis and processing of digital audio materials. The latter was supported by a dedicated computer to real-time data control and computation and, finally, a digital audio processor (DAP), dedicated to the synthesis and signal processing. The whole software package that made up the CARL system was installed on a DEC VAX 11/780 with Unix operating system. The real-time computation and computer control, however, was a DEC PDP 11/55. This, finally, was used to control the digital processor which, in turn, was equipped with several controllers, such as knobs and sliders, 4-channel DAC and 2-channel ADC made by Digital Sound Corporation. The project was completed through three different stages of implementation. During the final stage was implemented also the digital synthesizer FRMbox also made by Richard Moore.2
The Software - As we mentioned earlier, the most important aspect of the CARL system was not so much its physical structure but rather the set of software that made it up. In those years many software for computer music suffered, negatively, the rapid development of technological equipment and the formation of reference communities too small to ensure longevity and continuity. To counter these two aspects, then, the CARL system was developed in C language The CARL software package allows different work processes such as synthesis, processing and management of sound files, the movement of sound in space, the analysis and resynthesis. Over the years, the CARL System, also became a useful tool for the development of new software. In general you can count almost one hundred programs in the early years of the project.3 In addition to Richard Moore, who made Cmusic, other figures have helped to flesh out the entire software package: it should be noted the work of Mark Dolson focused on the Phase Vocoder, Julius Smith and Gareth Loy who made the Player and Procom software. The latter, together with the Unix pipeline, ensure communication between the various processes undertaken.4 It should be noted that from the beginning the CARL System was distributed for free along with the source code, just as he had done years before with Music V by Max Mathews. It was a choice which, together with the adoption of C, would ensure the emergence of a broad community.
Cscore, Cmusic and csound - Among the many programs of the CARL System, Cmusic is undoubtedly considered of greater importance for the number of features that characterize them. Besides Cmusic, we find Cscore, not a real software rather than a preprocessor designed to return scores in a format compatible with Cmusic. One more thing needs to be clarified about the program csound. Over the years we have supplied a terminological confusion because the csound named in the CARL System has nothing to do with the software developed by Barry Vercoe at the mid-eighties (Csound). In the case of the CARL System, csound, developed by Gareth Loy, indicates a program already in use since 1979, helping to manage sound files, designed to record, listen, and many other managerial functions.5
Diffusion - The CARL System has seen a good spread across several centers dedicated to computer music, both the U.S. and Europe. First, its use should be reported at IRCAM in Paris where, moreover, he had occasion to use it even Trevor Wishart, who, years later, realized the Composers' Desktop Project. The CARL System, then, was implemented with the PCS Cadmus of Monaco6 and, in Italy, are also used in the Laboratorio di Musica Informatica(LIM) in Milan as well as in several English computer music centers.
BICSF - At IRCAM, finally, were developed some innovations to the CARL System. It should be mentioned the so-called BICSF stands for Berkeley-Carl-Ircam Sound File. This is a new version of the program for the management of sound files that had hitherto been the responsibility of csound by Gareth Loy. As it was for his predecessor, BICSF allowed to carry out numerous functions including the management of the library and the conversion between different audio formats.7