Mosart Report

This report constitutes the first deliverable of the “Composition Tools” Task (T6) of the MOSART Research Network Project and the Final Report on a Qualitative Analysis of Composers as Work conducted at the University of Sheffield in collaboration with Mosart. In it we document work towards defining a set of requirements for enhanced support for electroacoustic music composers. Electroacoustic music composition tools and systems selectively attempt to provide composers with services they require for music generation, e.g., for accessing, generating, organising and manipulating audio (and other) objects, which constitute the composition. However, a primary aim of composition software is also to create conditions in which composers can be creative in the use of these services. The evolution of composition software with respect to the former requirements has been dynamic and innovative, stimulated by continuing development of new programming and audio synthesis paradigms and techniques. However, we believe that the aim of providing a fertile environment for creativity has been largely ignored. Specifically, we believe there to be a need to establish a research base for enhancement of support within composition software for creativity. It is this aspect that is addressed by this report. In the first part of the report (chapters 2 and 3), we address the two questions: do research results exist already that can provide an effective research base, and if not, how should such research be conducted? Accordingly, chapter 2 provides a critical review of previous studies of composition and software to support composers, and the research methods used. In the light of this review, it then considers which research methodology will be most appropriate for further research into music generation systems. It concludes that there is need for studies towards a research base for enhancement of electrocoustic music composition software, such that creativity is better supported. Further, given the preliminary nature of such research, the methods used should be qualitative, with the aim of identifying the sensitising parameters. This preliminary study should involve observation of composers at work in naturalistic situations. A second phase is then envisaged which will involve both qualitative and quantitative methods used in tandem. A future aspiration is to devise software solutions to improve support for creativity, in which engineering research methodology will become appropriate. The second part of the report documents our experimental study of composers, in which we have applied the research methodology proposed. Chapters 3 and 4 respectively document and analyse a series of three experiments, conducted with progressive refinements to our research methodology. In each, composers were observed working in naturalistic setting over protracted periods of time. The main notions that emerged from the analysis of the rich data collected in this way are reiterated in Chapter 5, the third part of the report, which concerns the software requirements motivated by this study. These concern: the value of the unexpected; the value of diverse, unfamiliar and idiosyncratic tools; the need for support for composers both as users and programmers; the key role of new tools; the need for personal and shared know-how; the need for physical and visuo-spatial control of audio; support to allow users to multitask; and the support for holistic approaches to composition. The Editor Dr Barry Eaglestone, University of sheffield university Page 2of 49 T6 (1 Deliverable) Requirements Specification Version 0.1 (31 May. 02) requirements which address these are then set out from the perspectives of system architectures, object and process management. Data and knowledge management, and composer-computer interfaces. The report concludes with a summary and identification future work toward elaborating the theories in this document and validating them through prototyping and evaluating composition software. Editor Dr Barry Eaglestone, University of sheffield university Page 3of 49 T6 (1 Deliverable) Requirements Specification Version 0.1 (31 May. 02)