Music analysis and, in particular music criticism, two fundamental tools of historical research and musicology in the broadest sense have become extremely difficult and problematical disciplines in the last few decades. The actual practice of analysis and criticism has by no means come to a halt, but the amount of 'truth' which was once thought obtainable from them or, at least, which was believed to have been won in the actual moment of undertaking them has diminished. We have begun to question the validity of certain elementary notions current in musicological parlance, for instance those connected with the concept of phraseology;' the harmonic concepts which have been the pillars of our musical thinking are now being discussed and reexamined from different points of view,2 and even the very idea of 'talking about music' must be revised with regard to its epistemological foundations.3 From all this it is clear that no scholar today can write about music with the same confidence and equanimity as his predecessors. His responsibilities are greater, and the fields he has chosen to research require kinds of knowledge that had previously been thought neither necessary nor relevant. These difficulties arise from at least two factors. First, there is the obvious problem that the analysis (or description) of a musical text always presupposes, sometimes implicitly or even quite unconsciously, some underlying theoretical basis: even in the most elementary analytical discussions for example when numbers of bars are counted, or corresponding phrases are distinguished reference is inevitably made to a set of theoretical ideas4 sometimes even without the writer being directly acquainted with it. What is more, the latent ideas which support an analytical or descriptive statement are

The concept of musical grammar

Material published on the subject of Acquisition, Representation and Reconstruction of printed music by computer is reviewed. The paper is divided into three sections. 1. Acquisition covers music data entry by automatic pattern recognition, directly-connected keyboard, soundtrack analysis, purpose-built hardware and other, compromise, methods. 2. Representation deals with Music Representational Languages (with particular reference to DARMS, MUSTRAN and ALMA), their background, structure, software support, and associated problems concerning score reconstruction. 3. Reconstruction discusses briefly the hardware available for computerized printing of sheet music. An Appendix is included which lists a representative selection of the most significant computer systems for printing music. The Bibliography contains nearly 200 references from diverse publications.

The acquisition, representation and reconstruction of printed music by computer: A review

The results of a detailed study of errors in a musical database are presented. It is shown that the form of analytic processing employed may substantially magnify the effect of database errors so that very small error rates can produce very large errors in the analytic results. An equation is devised to assist in estimating the effect of an overall error-rate on arbitrary analytic measures derived from the database. Considering the potential for erroneous or misleading analytic results, it is recommended that scholars using computer-based methods should habitually calculate error rates associated with their analytic procedures, and ought to present an analysis of errors in tandem with their results in order to validate their interpretations. Four methods of error detection are examined: manual proof-reading, double-entry method, programmed syntactic checking, and programmed heuristics. Significant differences were found in the thoroughness of different detection methods in uncovering all errors of a given type. The double-entry method was found to be superior to all other methods of detection; the humanities scholar's traditional allegiance to manual proof-reading was not supported by this study. Programmed methods of error detection were found to be fallible, but nonetheless useful.

Error Categories, Detection, and Reduction in a Musical Database '

A system that recognizes common rhythmic patterns through template matching is described. The use of template matching gives the user the unusual ability to modify the set of templates used for analysis. This modification effects a tradeoff between the temporal accuracy required of the input and the complexity of the recognizable rhythm patterns that happen to be common in a particular piece of music. The evolving implementation of this algorithm has received heavy use over a six-year period and has proven itself as a practical and reliable input method for fast music transcription. It is concluded that templates demonstrably provide the necessary temporal context for accurate rhythm recognition. >

Template matching for rhythmic analysis of music keyboard input

For this presentation I am much indebted to Deta Davis, who very kindly lent me a prepublication copy of her forthcoming bibliography on computer applications in music. With its several thousand entries, it promises to become the standard reference in its field. On the assumption that publication activities reflect the relative importance of trends and currents, my first two figures are based on quantitative data extracted from the Davis bibliography. The remaining figures combine information from several sources, the most important of which is the Davis book. Concerning recent developments I have received valuable advice from my McGill colleague Bruce Pennycook. Figure 1 shows for a number of areas how the last decade compares to the pre-1977 period. White bars show how much literature in each area falls before and after 1977. They present a clear trend: publication has increased in almost all areas; in fact, it has increased by a factor of about nine if one considers that the late period covers only about a third as much time as the early period. The black bars show the relative output of all areas within each period. From both comparisons it is clear that the areas of aesthetics and musicology, including analysis, have at best held steady, while the other areas have grown very rapidly. Figure 2 looks at some of the factors contributing to the increase. This graph is based on numbers of entries for the years 1977 through 1984 in Davis's chapters on sound generation hardware and software on microand mini-computers, and on articles about MIDI. The micro-computer curve rises early; the advent of MIDI in 1983 pushes both the hardware and software curves sharply upward. The remainder of this presentation is devoted to some of the trends in computer applications during the SMT decade. As the largest and most influential, the composition area would merit its own separate presentation. Since developments in computer composition are readily available in publications of the Computer Music Association, I am restricting myself here to the areas of music theory and analysis. Figures 3 through 5 contain a number of author names distributed over our decade. The lag in library information processing makes entries sparse in 1986 and 1987; also, despite the abundance of names on the charts, omissions are even more abundant, and there may be inaccuracies. I am anxious to stress that omissions do not reflect a judgment of quality on my part. Moreover, I shall cover only selections from the charts while the rest will provide a kind of context for your reflections. Figure 3 concerns music input and output. While much ingenuity has been spent on all aspects of music I/O, two particularly challenging areas are those of score input and sound input. The core of the first area is the problem of optical score recognition, which was solved in principle for a small subset of notation by David Prerau in 1968. Problems of practical and financial feasibility, however, have remained. The Helmers, Andronico, and Tojo entries represent important advances, but, so far, Ichiro Fujinaga's current work at McGill promises to become

Computer Applications: Analysis and Modeling

The high-speed digital computer has provided the musician with a powerful tool for music research. In information retrieval, analysis, digital sound synthesis, and music printing, studies have clearly demonstrated the computer's potential for assisting the musician.' At the same time, however, the difficulty of preparing music data for computer processing has been, and continues to be, a major deterrent to large-scale computer-aided music studies. At present, notwithstanding several experiments in which special optical devices have been used in the encoding process,2 the only generally available device for converting music scores to computer-usable code is the keypunch machine. But the keypunch is an awkward device for encoding music: its character set is not compatible with music symbols; it is tedious to use; and proofreading of encoded materials is cumbersome and timeconsuming, even for those whose computer installations provide on-line text editing. As an alternative to both optical encoding, which is highly desirable but developed for only the simplest encoding tasks, and the keypunch machine, which is available but generally unappealing to the musician, we are developing an encoding system at Indiana University that consists of two independent but related encoding devices: a sonic digitizer and an organ keyboard, both attached to the computer, which permit the operator to enter data directly from the score for automatic conversion into computer-usable code without initmate knowledge of the code itself or the cumbersome proofreading normally associated with keypunching. Once data are encoded, the programs output the code into standard card-image format for storage on magnetic tape. This feature allows a tape to be processed at any computer installation capable of handling magnetic tape. For portability, all programs are written in ANSI standard FORTRAN to allow operation on any computer with a FORTRAN compiler and appropriate encoding devices and display equipment. The encoding language, the MUSTRAN code originated by Jerome Wenker,3 was chosen for several reasons. First, MUSTRAN is a largely mnemonic code and is relatively complete. Second, it has an operational translator and syntax analyzer that we have running on our CDC 6600 computer on campus. (Versions are also available for IBM and UNIVAC machines.) Third, Wenker having recently updated his language and its associated translator and syntax analyzer (called MUSTRAN II), the language is maintained and current. And fourth, our Computer Music System (discussed briefly below) uses MUSTRAN for encoding, music printing, and analysis. At the same time, however, we realize that the DARMS (Ford-Columbia) language is also carefully designed. It recently has been written in canonical form4 and will no doubt eventually have its own translator and syntax analyzer. Since DARMS is a complete language, and in view of the current efforts being made in its behalf, we are now developing DARMS-toMUSTRAN and MUSTRAN-to-DARMS conve sion packages that will eventually allow us to encode in one language and conveniently convert

Rosalee Nerheim

Papers

A System for Interactive Encoding of Music Scores Under Computer Control