Can a computer read music

�

A.� Not yet, but a groundbreaking project to research the reality of a computer programme that will read music is under way at Leeds University.

�< xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />

Q.� What is involved

A.� Doctor Kia Ng is using a four-year grant of �140,000 supplied by The Arts And Humanities Research Board to advance computer software beyond current limitations, which enable computers to analyse letters and numbers, but not musical scores.

�

Q.� It sounds complicated.

A.� It is. The major problem is the individual ways different composers have of writing music. The five lines and four spaces of a music score are really only a broad basis ' individual musicians have developed their own particular methods of indicating tone, pace and emphasis, and these can vary massively from one to another, anything from broad thick pen strokes to spidery lines. Add to that various homemade shorthand symbols known only to the composer, and the idea of interpretation by another musician becomes fraught with difficulties. Because of the speed at which some film score composers are obliged to work, their music may not even be written on the page in its proper place.

�

Q.� So how can a computer system help

A.� Doctor Ng has developed a computer programme that actually reverses the standard 'human' method of composition. Instead of beginning with the blank page, and analysing the notes, the computer looks for uniformity of note construction, and uses memory-banked theory to supply the most likely of a series of optional notes where notes are missing. The easiest way to understand the concept is to think of it as a musical 'spell-checker' with the computer making a decision based on the likelihood of a note pattern, which it has 'learned' from the piece as a whole. The computer then finishes by placing the score page behind the notes it has created, and the result is a standard score, which can be saved as a midi file, and interpreted either by a synthesiser, or by another computer.

�

Q.� What does this mean in the wider context of recording musical notation

A.� The possibilities are very exciting. With the new software, it will be possible for ancient scores rendered illegible by either paper deteriorisation, or spots and blemishes on paper that has not be stored carefully, to be resurrected and played in the style intended by their original composer. Work which would have taken a musician well versed in a particular composer's work several months to complete in painstaking small steps, can now be over-viewed and completed by a computer in a fraction of that time.

�

Q.� So some 'lost' scores may now be redeemable

A.� In theory that's true, although the software is not yet completed, but the initial signs are encouraging.

Q.� Does the software have any other implications

A.� It could fundamentally shift the methods by which music is composed, and recorded in a way not seen since music was first committed to paper several hundred years ago. The differences in musical notation, for example, medieval musicians used a four line stave, not the five line stave used by modern composers, can be overcome. If that principle is extended, then Asian music, which uses a different system of notation entirely, could be brought under an umbrella of standardisation, which could mean the beginning of 'world music' in the truest sense of the word.

�

Q.� So mis-interpretation of a composer's work could be a thing of the past

A.� It is beginning to look that way. The current cumbersome legal system for deciding if a particular composer's copyright has been infringed could be swept away. The present�system involves lengthy courtroom debate about what constitutes a 'song' or what makes a particular piece of music distinctive and original. All that could be eliminated by a computer programme, which will analyse and compare two pieces in a few seconds, and confirm if one is unacceptably close to the other, by means of a standard ruling, which can be predetermined.

�

Q.� Anything else

A.� The software under construction could complete all the laborious transcription of orchestra parts. Instead of a musician having to work out individual instrument parts, the software could write out each part simultaneously as the score takes shape, saving vast amounts of labour-intensive transcription.

�

Q.� It all sounds a bit Big Brother doesn't it

A.� You have to approach the idea of this software as an aid to composition, and in no way could it ever be seen as a replacement. Musicians will always create music�- machines will never be able to supersede the concept of composers working through musical ideas. No doubt the software will be greeted by some with the same suspicion that greeted computers that could type, wondering if that would see the end of handwriting, or if the Internet would spell the end of books. As then, this new technology will be assimilated into the existing world, and be seen as a tool to be used by the musicians who want to push music forward through as many boundaries as possible.

�

Do you have a question about music Please click here

�Andy Hughes

�

�