Finished musical scores from the keyboard: An expansion of the composer's creativity

Music copying (the writing of musical scores) does not lend itself easily to automation. Like language, music notation has evolved over a long period of time, resulting in notational conventions which are complicated and often illogical. Many of the symbols required are quite elaborate, and some can vary in size and proportion. For a musician to be able to read the music quickly and correctly in a performance situation, the notation must follow a very precise set of aesthetic rules. For these reasons music is almost always hand copied, even today. Composers typically draw up their own scores and then send them to a copyist to have the parts extracted. The parts usually contain mistakes, and the scores are often not as legible as they should be. Furthermore, musicians often compose under extreme pressure, allowing little time for proofreading and revision. The following is a description of a computer system which offers an alternative to this expensive and time consuming process by automatically transcribing keyboard performances into legible musical scores. In designing an automated music printing system, a number of complex problems arise. The user should not be expected to be a computer expert. Input to such a system should be simple and natural and not require unusual skills. The system should be able to produce valid music with minimal user interaction, and therefore must handle unassisted many complex aspects of written music. For the system to do everything completely automatically would require artificial intelligence capabilities which could not be implemented in a small system affordable to the musician. Instead, comprehensive, flexible editing must be provided to allow the user to step in where the machine fails to copy correctly. Finally the system must provide quick, clean output on durable paper so that the music can be reproduced easily and will withstand the demands of stage performance.