>*window=68    AEP/LEC  114b   IBM-AEP Boston June 29, 1987
>BM=58 TM=1 LM=8 RM=76 J=N LS=1
PRINTING MUSIC ON SMALL COMPUTERS
Presented by: Leland C. Smith, Stanford University


      While written languages had already reached a high degree of
development one thousand years ago, it is curious to consider that
musical notation was still in its infancy.  Notation consisted of
little more than a few marks designed to serve as an aid to the
recall of memorized melodies.  Many musical "scores" from those
times remain totally undeciphered.  But then the technology of
musical notation seemed to take a quantum leap forward in the 10th
and 11th centuries - just at the time when Europeans began to add
harmony and counterpoint to their music.

     The idea of a group of musicians making different sounds,
with different rhythms, at the same time, and in an ordered way,
rapidly led to situations where it became very difficult to teach
the various performers by rote.  Musical notation was forced to
become much more precise.  By the twelfth century, most of the
fundamental aspects of notation as we know it today had been
developed.  Now, with a true representation of complex sound
events available for study and contemplation, the creative
musical minds rapidly found ways to put sounds together in yet
more complex ways.  In earlier times a musician had to travel
long distances to become acquainted with different local musical
practices.  With the advent of handwritten musical manuscripts,
the music itself began to do the traveling.  Thus, the
development of standardized musical notation brought about an
even greater acceleration in changes in musical style.

     A significant turning point in western civilization came in
the fifteenth century, when the use of the printing press spread
rapidly across Europe (1440, Gutenberg).  (It seems there is
evidence that printing presses were used in Korea over 100 years
earlier.  However, as with some other early technological
developments in the far east, the society did not seem to be
able - or was not interested enough - to make the crucial
advances that would have led to the continuous use of the
process.)

     The printing of music began by fits and starts.  Some of the
earliest printing was done from carved wood blocks.  The results
were rather primitive - nowhere near to the standards of mediocre
hand copy. While moveable type served admirably for the printing
of European languages, the purely graphic aspects of music
notation caused problems for printers which have persisted to
this day.  A great number of the symbols used in music can be
produced by individual pieces of type, but a number of other
musical symbols are not fixed and require an infinite variety of
sizes and shapes.  Various methods were tried to overcome this
problem.

     Some of the best music printing at this early time was that
produced in 1501, in Venice, by Ottaviano Petrucci, who is
generally considered to be the first real music publisher.  He
apparently printed his pages by making two or three carefully
registered impressions.  The time required to print each
impression must have been very much longer than that required
to do a hand-written musical manuscript.  However, once the setup
and procedure for a page was complete, it is likely that 10 to 20
pages could be done in the time needed for one page done by hand.

     The moveable type, single impression method was by far the
fastest way to produce music but the results were always rather
inelegant, the greatest problem being that the long, continuous
staff lines always had a rough and uneven look.  Despite its
shortcomings, this method persisted well into the nineteenth
century.

     Once these early technologies became widely used, there was
 a  veritable flood of printed music in Western Europe.  Some
 composers such as Josquin des Prez and Orlando di Lasso became
 internationally famous.  In earlier days only the church and
 nobility could afford to have collections of music.  Now printed
 music was available to every bourgeois family.  The art of music
 was democratized.  The various printed collections of motets,
 masses, chansons, etc., became source material for each new
 generation of musicians to build upon.


     As mentioned above, the moveable type method of printing
music was never satisfactory.  While the art of engraving became
well known at about the same time (the mid-fifteenth century) as
printing from type, it took over a hundred years for this
technique to be applied to the printing of music.  The first
engraved musical publications appeared in Rome in 1586.
("Ghirlando di fioretti musicali," a volume of madrigals.)
This process developed rather slowly, and it wasn't until
another hundred years had passed that the engraving of music
began to really replace typesetting.

     Many believe that music printing reached its highest point
in  some of the engravings of the mid-nineteenth century.  While
there is some free hand work in music engraving, for the most
part the various symbols are engraved by means of "punches" which
are steel dies, very precisely cut to the shapes of notes, clefs,
etc.  It was said that apprentices in engraving (beginning at the
age of 14) would spend as much as ten years at work before being
allowed to produce a page of music independently.  However by
1920 a writer commented, concerning the state of music publishing
in Britain, "One circumstance that has stood in the way of
improvememt in the music engraving trade is that the prices for
the work are very low. ... The prospect for youths entering the
trade (is) not ... encouraging."  He went on to say that there
were only about 100 engravers of music in Great Britain in 1920.
Today there may be less than that in all of Europe and  America.

     At the end of the eighteenth century, Alois Senefelder, of
Munich, discovered the means of lithographic printing.  This
method had the great advantage that errors were relatively easy
to correct.  It is said the composer Carl Maria von Weber
interrupted his musical studies in 1799 so that he could learn
the art of lithography.  This process enjoyed continued success
throughout the nineteenth century.  However music printed in this
manner could not match the beauty and consistency of engraved
music.

     Any number of elaborate mechanical schemes for printing music
were developed at the end of the nineteenth century and into the
twentieth century, culminating in the first music typewriters
about 1906.  Music typewriters are still widely used, especially
for the preparation of popular sheet music.  However the great
disadvantage of typewriters is that the symbols can be printed in
only a few fixed sizes, and it is impractical to type in slurs
and the larger items such as clefs.  In fact, about half the work
on a page of music is usually done by a draftsman using templates
to create all the things that cannot be done by the typewriter.

      With the advent of inexpensive photo reproduction methods in
the mid-twentieth century, many publishers have sacrificed visual
beauty in printed music for financial expediency.  Many
"publications" consist of no more than photocopies of the
composer's own shakey handwriting.  A recent "scholarly"
publication of some 200 eighteenth-century symphonies appears,
for the most part, in the uneven hand of numerous students of the
project editors.  Of course this method has the great advantage
that no engravers, or even professional copyists, had to be paid.


     Just at the point where it seemed that the impossibly high
price of artistic integrity would cause music publishing to go
into a decline, a new technology has become available that
promises to bring about a renaissance in the process of music
printing.  This new technology, of course, is the computer.  It
was not until the early 1970s that the graphics potential of the
computer began to be realized.  At that time, at the Stanford
Artificial Intelligence Laboratory, I  began to develop SCORE,
a  computerized music typography system.  It was a stimulating
environment.  While I was struggling to achieve proportional
spacing of complex orchestral counterpoint, sitting next to me,
some of the pioneers in computer aided design, were struggling
equally hard to get their CRT displays to blink at the proper
points to indicate short circuits in the next generation of
computers.  However, it wasn't until the development of the
powerful personal computers of the 1980's that a program of
the  complexity of SCORE could be made available to the music
community at large.

     While the production of a page of music has something in
common with word processing, the musical code system presents
many unique graphics problems.  Horizontal space represents time
and vertical space represents pitch.  However, both of these
dimensions are treated rather loosely in conventional musical
notation.   Most pertinent musical symbols are related to a
particular set of staff lines, and the spacing of these staff
lines determines the actual size of the symbols.  However,
several different spacings of the staff lines might appear
for  music occupying the same time span.  This has led to the
necessity for different spacing criteria for the horizontal and
vertical dimensions.  Thus, the horizontal requires global
spacing; all notes to be heard simultaneously must be lined up.
However, the vertical dimension must be based on the local size
of the scale step, as determined by the spacing of the staff
lines.

     Although most musical symbols maintain a conventional shape
and orientation, some frequently appearing objects have an
infinte variety of presentations.  The cross beams that are used
for combining rhythmic groups can be of any length and can appear
at almost any angle.  Also, there are a great many different
combinations of long and short beams that might be required.
The curved slur lines that indicate musical phrasing can take on
almost any aspect.  A musical score often requires many different
symbols which indicate special ways to play the various
instruments.  Sometimes it may be desirable to print music using
the notation symbols of different cultures or earlier centuries.
For all of these reasons, it can be seen that a comprehensive
music printing system on a computer must be made up of a
general-purpose graphics program, combined with a large library
of musical symbols.

     To facilitate the creation of musical pages, the SCORE input
language includes the basic aspects of musical terminology.
Notes may be entered by their letter names, C, D, E, etc.
Rhythms are entered either by letter (e = eighth, q = quarter)
or by their fractional denominator numbers (8 = eighth, 4 =
quarter), etc.  All rhythms are automatically spaced in musical
proportional manner.  In this program there is no limit on the
complexity of rhythm.

     While the larger problems of page layout are handled very
well by the SCORE program group, it seems unrealistic to try to
place everything on a musical page by automatic means.  It is of
the greatest importance to have a comprehensive editing system,
wherein every detail of a page of music may be easily accessed.
With SCORE, each item on a page has a series of numerical
parameters which control all aspects of its representation.
In the editing process these parameters may be manipulated to
achieve virtually any result desired.  Then, if the available
symbol library does not include the needed object, a separate
drawing program is provided so new objects may be added to the
library.  The SCORE library contains about 200 items at this
time, and the program is constructed so that it can access a
library of unlimited size.

      The importance of SCORE to the future of music publishing is
quite clear, and the usefulness of the program in the university
environment has already been amply demonstrated.  Many of the
students of musical composition at Stanford have learned to
prepare their works using this method.  The student performers
greatly appreciate the professional-level copy, which is a vast
improvement over the uncertain musical handwriting of most young
composers.  Several musicological dissertations have included
examples using this system.  It is not unrealistic to imagine
that music departments will require computer-produced musical
examples in all dissertations, just as typewritten copy is
required in place of handwriting.

     Since the SCORE parameter encoding system includes all the
information that is present on a page of music, such data can
also be used for computerized studies in musical analysis.  One
Stanford dissertation includes hundreds of musical incipits
produced by SCORE.  A separate program, working on the SCORE
database, has produced a cross-reference index of these incipits,
showing relationships of rhythms and musical interval content.

     It is hoped that college music departments will be able to
provide public music printing workstations for their students -
much as they already provide music practice rooms.  At Stanford,
over the past three years, classes have been offered in the use
of the computer for printing music.  It is likely that such
classes will be regularly presented at many institutions.


     (The SCORE program group runs on all IBM-PC computers with
     graphics capability and 640k memory.  It has been licensed
     to Passport Designs, Inc., 625 Miramontes Street, Half Moon
     Bay, CA. 94019.  It is due to be released in the summer of
     1987.)