Mechanical Music Digest  Archives
You Are Not Logged In Login/Get New Account
Please Log In. Accounts are free!
Logged In users are granted additional features including a more current version of the Archives and a simplified process for submitting articles.
Home Archives Calendar Gallery Store Links Info
MMD > Archives > August 2002 > 2002.08.30 > 04Prev  Next


Musical Theme Search Using Incipits
By Robbie Rhodes

What's That Song I Just Played?
  by Robbie Rhodes (An Absentminded Performer)

In his article, "A Ragtime Data Base (RAG TIMES, March 1990),
researcher Ed Berlin discusses a computer data base to serve as an
"inventory of material" to aid research, and he worries around about
what types of data should and shouldn't be included.

I think that, given the right tools, the people that use a system
eventually discover for themselves what is and isn't important; the
demands of the market will then lead software developers to create
better and better tools.

But Ed is certainly on the right path, and I think including an incipit
(the first few notes of a theme) is a marvelous idea of his.  If the
data base is concerned with pop songs the incipit, or thematic index,
for a given song would just have a few bars each of the verse and
chorus.  Rags would have three or more themes in the incipits.  Like
Ed says, now you can access works by theme.

Just think what that means: you could tell the computer to find the
melody that's stumped you!  Oh, how many times I've needed this
assistance when I perform a rag and then discover I've forgotten the
name of it.  (If I was wise I would always carry with me the thematic
index from the German edition of Scott Joplin -- it's indexed just like
a Mozart or Schubert folio!)

Well, suppose that we had a data base file for all the rags ever
written, good and bad, and that we have a data base management program
on our home computer to search the "Rags" file.  The program would be
much like any other data base system, but with an important added
feature: besides searching for word strings (i.e., seeking a comparison
with a word or string of words you just entered), the program could
search for theme strings.  That's as handy as having a room full of
ragtime fans when I don't know the tune's name.

The accompanying printout shows a hypothetical example of a melody
search, in which I ask the computer to find a 4-bar melody that I heard
somewhere long ago.  I have a hunch that it's part of an old blues
tune, so I load my data base file called "Blues Song Copyrights".

How do I enter the melody into the computer?  By playing it on a simple
electronic keyboard/synthesizer connected to the computer via a MIDI
adapter.  The computer loudspeaker makes a metronome click at the tempo
I selected, and I just start playing.  When I stop, the single-note
melody is displayed on the screen and I check it for correctness, both
by viewing the "manuscript" display and by "playing back" the notes
over the MIDI wire to the little synthesizer.

I also tell the search program to be liberal about syncopation, not
knowing if the melody should be notated in straight eighth-note rhythm,
triplets or dotted-eighths and sixteenths.

After a few moments (or minutes, maybe) of searching the computer
thinks it found a reasonable match, and displays the incipits of the
themes for the song named "The Florida Blues".  Well, I wasn't too far
off, was I?  I entered the search melody in the key of B-flat, but
fortunately the computer ignored the absolute key and concentrated on
finding the relative note pattern of the melody (filed in the key of C).

Notice that the correct melody has eight notes in the measure, whereas
I had entered a seven-note measure.  The computer decided that I should
view a match that was 7/8 correct, and when I looked at it I said
"Eureka" and quit the search.

Well, that was pretty neat discovering what song I'd been humming, and
of course I want to find a record of the song, so I tell the computer
to search for "The Florida Blues" in other data files containing
phonograph record titles and piano roll titles.  It finds 5 phono
records and 2 piano rolls -- must've been a big hit tune!

My hypothetical computer system example is based on searching a set
of multiple, independent files.  The primary file would be the Music
Copyright data base, taken from the U.S. Copyright Office file cards
and depositions.  It would have the theme incipits, song title,
composer, and copyright dales, and any or all of these fields could
be searched for.

Secondary files would be used to find additional information about
the song, once you've determined the correct title.  The famous
"Discography" by Brian Rust would be invaluable for finding phono
recordings of the song,  and analogous "Rollographies" have been
compiled by piano roll collectors like Mike Montgomery and Richard
Riley.

Other auxiliary files could handle the ever-growing minutiae that
Ed Berlin is concerned with: there would be a file of "Variant Titles",
and of course, the "Index of First Lines" for song words.  We need a
big, and expandable, file for "Sheet Music and Covers", which all the
sheet music collectors would use and contribute to.

The computer program to perform a thematic search has yet to be
published, but it's not an impossible task: I believe that the Center
for Computer Research in Music and Acoustics (CCRMA) at Stanford
University has done research in the area of musical theme recognition.
When music collectors and researchers express enough interest then
someone, either a big software house or a garage hacker, will publish
the tool that's needed.

The big files of reference data will come, eventually, from dedicated
researchers like Brian Rust and Richard Riley who have already entered
the data into a computer and are willing to share it.  (Costs are
presently unknown.) The "Music Copyrights" data base should be published
by Library of Congress, if we can ever get our government to fund the
monstrous task of entering the data.  And the rest will come from you
and me, entering into our home computers the data from our collections
of sheet music, and then swapping copies with friends.  See how it can
grow?

To the ambitious ragtimer that wrote Ed Berlin about a ragtime data
base, I advise, "Get busy typing!"  Anybody's data base format can be
converted to any other format that comes around; the important thing,
and the first thing, is to get it into the computer.  Put in as little
or as much data/trivia as you wish, and then share it with your
friends.  They'll let you know what's missing.

And the realty ambitious ragtimers, equipped with a keyboard and
synthesizer and MIDI music programs, can get started entering the
incipits for thematic indexes of their favorite songs.

I'm going to begin with the "unknown" rags I enjoy playing -- then
I won't have to ask the audience, "What's that song I just played
for you?"

Robbie Rhodes
March 27, 1990

(This article was previously published in the newsletters
"The Rag Times" and "Remember That Song".)

 [ The complete article with manuscript incipits and MIDI sounds
 [ is at http://mmd.foxtail.com/Sounds/index.html  -- Robbie


(Message sent Sat 31 Aug 2002, 01:08:31 GMT, from time zone GMT-0700.)

Key Words in Subject:  Incipits, Musical, Search, Theme, Using

Home    Archives    Calendar    Gallery    Store    Links    Info   


Enter text below to search the MMD Website with Google



CONTACT FORM: Click HERE to write to the editor, or to post a message about Mechanical Musical Instruments to the MMD

Unless otherwise noted, all opinions are those of the individual authors and may not represent those of the editors. Compilation copyright 1995-2024 by Jody Kravitz.

Please read our Republication Policy before copying information from or creating links to this web site.

Click HERE to contact the webmaster regarding problems with the website.

Please support publication of the MMD by donating online

Please Support Publication of the MMD with your Generous Donation

Pay via PayPal

No PayPal account required

                                     
Translate This Page