Microsoft Algorithmic Composition Patent

Mark Coniglio (
Thu, 30 Jul 1998 15:00:51 -0400

Dear List,

I thought that you all might be interested in this email that is making the
rounds of the electronic music community. It deals with a patent that
Microsoft has been given on a kind of interactive, algorithmic music
composition scheme. Even though this is specifically music and not
specifically dance, the idea of granting corporations patents on
compositional techniques in any artform is a grim notion, IMHO, and a
subject that will affect us all eventually.

The letter gives some background on the patent, and a response by
electronic music composer Laurie Spiegal that is scheduled to be published
in Computer Music Journal.



From: Doug Keislar <>

Microsoft was recently (May 19, 1998) granted a United States patent
(#5,753,843) for a "System and Process for Composing Musical Sections"
(including 47 claims), this patent clearly pertaining to the algorithmic
generation of music. Computer Music Journal invited Laurie Spiegel to
write a guest editorial (below) about the implications of such patents.
also invite follow-up letters to the editor in response to this editorial.
The best of these letters will be published, along with the editorial, in
the Winter 1998 issue of the Journal.

To submit a letter to the editor, please send email to
with a subject line of "Patenting Compositional Algorithms". As requested
by the administrator of the ICMA mailing list, please do not post replies
to the ICMA mailing list, as it is not a discussion group. I will serve
a moderator, collecting the replies and posting the more interesting ones
in a message to the ICMA mailing list.

The abstract of the Microsoft patent describes:

"A system and process for comprising [sic] a musical section in response
to a user's interaction with a multimedia presentation. The system
includes a composition engine, performance engine, and arbitrator. The
arbitrator provides an interface with an application program running a
multimedia presentation. The arbitrator receives parameters from the
application program indicative of a user's interaction and the type of
music the application program requests in response to the interaction. The
parameters are passed to the composition engine which composes a musical
section having a chord progression and other data therein. The musical
section and a style provided by the arbitrator are used by the performance
engine to generate music sequence data for driving a musical instrument.
The performance of the musical sequence data by the musical instrument
occurs substantially contemporaneously with the user's interaction which
caused the musical section composition. Because the composition engine
processes which vary the composition of musical sections, the user events
which initiate composition of a musical section and which occur at the
place within a multimedia presentation, still vary the performance at each
user event."

The claims includes the following text:

"A system for composing music in response to a user's interaction with a
multimedia presentation comprising: an application program interface for
receiving parameters identifying a style, a shape, and a personality for
music that conform to said user's interaction with said multimedia
presentation; and a composition engine for composing a musical section
corresponding to said parameters so that a user perceives the performance
of the musical section to be related to said user's interaction with said
multimedia presentation."

Copies of this patent can be obtained for US$ 13 from MicroPatent;
(800)648-6787; World Wide Web

--Doug Keislar


Editorial by Laurie Spiegel

There have been patents for musical inventions, such as piano action
for many years without apparent detriment. However, throughout the 20th
century, the designing of artistic processes and creative techniques has
increasingly come to be considered an integral part of an artist's
work, rather than being seen as the province of a separate tool-building
specialist. Entire fields such as algorithmic composition, interactive
multimedia, and literary "process art" have become established in recent
decades, based on the premise that the designing and implementation of
specific creative processes and artistic techniques constitute artistic
creation, every bit as much as do the data that such processes generate.

If specific compositional techniques are now to be privately owned, must
each composer, especially composers of computer-based interactive process
pieces, now stop to do patent searches routinely as part of their work or
else risk being taken to court? With the floodgates now open for a gold
rush of corporate claims to very specific compositional techniques, how
we composers supposed to preserve our sense of freedom, our exhilaration
exploring, and our deep psychological immersion in following musical ideas
wherever they may lead us, while knowing that we cannot ever be sure
anymore of the simple legality of any new refinement we may make in our
process-based works?

Are we composers going to end up having to pay a royalty to the owner of
each technique we use, when the royalties we receive from the music we
compose using these techniques typically would not even pay for the
paperwork of just keeping track of them? Must the education of every
composer who wants to make process-based music now include courses in what
techniques are exclusively owned by whom and for how long, what fees or
methods will decriminalize their use, and the penalties to expect for
unauthorized use? Will process-based composing or composer construction of
interactive algorithmic tools now become so legally complex (and possibly
dangerous) that such approaches will simply die out? When we want to use
specific compositional technique, how do we keep the question of who owns
it from interfering with our personal sense of creative freedom?

As a method of motivating new development for compositional techniques,
would the use of patents have worked in historical contexts? To
an instance, what if someone had patented the replaying of a musical theme
at a time delay to itself, early in what was probably the last great era
process-based composing: the era of Bach? Would Bach have been able to
and also willing to pay, royalties (or--perish the thought--legal defense
fees) to use or build new techniques based on the patented imitative
contrapuntal processes his works required? Or would composing the way he
did have made him a criminal, as Galileo and others came to be considered
criminals for their scientific work? And could any patent-holding tool
designer ever possess sufficient understanding of the working requirements
of composers of Bach's artistic caliber to be able to create procedural
tools that were adequately fine-tuned for every possible such composer's
unique musical approach? Or, in another hypothetical example, could any
tool builder whose products had been fine tuned for Haydn's methods have
anticipated the ways in which Beethoven would need to break out beyond
their scope, or why this would be important, or how to accommodate the
change in advance?

Then why does our own society assume that the definition, implementation,
and provision of any such creative technique should be done
non-competitively, non-pluralistically, under the complete control of any
single corporate monopoly? Why should the monopoly we call "patent"
artistic methods, tools, or techniques within its domain? Simply because
these can now be constructed within the medium of computer software, and
because the law now allows the patenting of such processes if implemented
as computer software?

The only arguably successful scenario I can envision that takes as a
premise the existence of an 18th-century patent on time-delayed repetition
techniques, and that would still allow us the Well Tempered Clavier, the
Musical Offering, and the Art of the Fugue, is one in which Bach got
himself hired as Company X's official court (well, company) composer in a
corporate reprise of the aristocratic private patronage system. But in all
likelihood Telemann would have gotten the job instead.

We specialists created the field of computer music collectively but from
the intersections of our very own personal visions and desires. This
our lifework, is increasingly influenced by differently motivated
The legal departments of large corporations, or those concerned with the
price of corporate shares trading in international markets, might govern
whether or not the next Art of the Fugue will ever be made. The current
situation is neither without precedent nor easily resolved, but we do want
to preserve what we value in our art and its potential to evolve.

Douglas Keislar
Editor, Computer Music Journal (MIT Press)

email: or tel: 1-510-486-0174
fax: (510) 486-0868
Computer Music Journal, 2550 9th Street #207 B, Berkeley, CA 94710, USA

Mark Coniglio, Artistic Co-Director |
Troika Ranch |