Help with [Indexes] [Incipits] [Hymnals] [File Formats] [Databases]
[Plans] [What's New] [Help Wanted] [©Copyright] [Contact Me]
An "incipit" is the first part of a tune or text. Here, the incipit is the first few notes, encoded into a series of letters or digits.
There are several ways of encoding incipits in use. Most are not "complete" encodings -- that is, they do not include enough information to show exactly how to sing the tune, because two different tunes might have exactly the same incipit. But it is easy to convert a known tune into incipit form, and then look up that incipit in an alphabetized list, to see if there are any tunes that might match. There will hardly ever be more than two possibilities.
This is similar to the alphabetic portion of the incipit format used in Hymn Tune Names by McCutchan (but see below.) Each note is expressed as one letter, chosen from the initial letter of the names of the notes in the solfeggio scale (with moveable syllables): do, re, mi, fa, sol, la, ti; with "do" representing the root of the major scale. Lines of the tune are separated by a vertical bar. The index is simply sorted in alphabetical order.
Thus the first two lines of the famous tune "Old Hundredth" would thus be
represented as "
ddtlsdrm|mmmrdfmr". The beginning of the
(minor) tune "Leoni" ("The God of Abraham Praise") would be
This notation does not express the length of each note or its relation to the metrical beat. It does not indicate whether a note appears in a high or low octave. It ignores sharps and flats: a chromatic note is represented as the note either above or below it in the diatonic scale, whichever is different from the note before it and, if possible, the note after it. It does not indicate what notes are tied together (and sung to the same syllable.)
If you have trouble finding a tune in this index, you might check for notes that are tied in one verse but sung to two syllables in another verse, and you might consider alternate ways of expressing accidental notes. Sometimes a passing tone might be included in one version of the tune but omitted in another -- try looking both with and without those notes. Also, note that the minor scale in this form begins with "la", not "do."
This incipit is the simpler of the two forms used in the monumental Oxford Hymn Tune Index, as well as the related American Fuging Tunes, 1770-1820. Each note of the scale is indicated by a number from 1 (do) to 7 (ti). As in the solmization code above, timing, melismas, octaves, and accidentals are ignored (the documentation does not describe whether accidentals are moved up or down.) One important difference is that for minor keys, the scale goes from 1 (la) to 7 (sol). The Oxford index consistently gives the first 15 notes of the song, split up into three groups of five numbers, regardless of line breaks. Where 15 notes are not available, this index inserts closing underscores.
Thus "Old Hundredth" appears as
11765 12333 32143". "Leoni" appears as
51234 53456 75234".
Hymn Tune Names extends the do-re-mi notation in several
ways: slashes represent measure bars; postfix dots and dashes represent
longer notes, and font variations represent different octaves (from lowest
to highest, "D", "d", and "d"). Prefixed, superscript sharp, flat,
and natural signs represent accidentals. The whole incipit is preceded by
a number representing the number of beats in a measure. "Old Hundredth"
appears as "
4: d- / d T L S / d- r- / m-". This gives
enough information to actually hum the first few measures of the tune.
A Dictionary of
Musical Themes, by Harold Barlow and Sam Morganstern,
describes each tune by the normal letter names of the notes (A,B,C,D,E,F,G)
as if the melody were transposed to the key of C (or C minor). It includes
accidentals (sharp or flat sign after the note), and is thus more precise
than either of the forms used here. Old Hundredth would appear as
c c b a g c d e".
This was a pioneering effort, begun before computers were available. But its
greater precision prevents automatic conversion from the "base" do-re-mi
format; so the incipit indexes are not included in this form.
A Catalogue of 18th-Century Symphonies by Jan LaRue, a
pioneering attempt to use computers, uses a fixed-scale format. Its incipits
begin with the key letter (A,B,C,D,E,F,G) followed by optional "$" to
indicate flat, and "-" to indicate minor. Then the first notes (using
letters as above, but assuming the key signature, and representing
accidentals with the
suffixes "$", "N", or "#" for flat, natural, or sharp, respectively. A
series of repeated notes is represented by a prefix digit ("3C" is
equivalent to "CCC".) Old Hundredth would appear as
In view of the common practice of transposing hymns,
this is not a useful format.
The "abc" notation, widely used by the folk tune community, is the most elaborate of these formats: it is even complete enough to provide a basis for printing a full score. The notes are represented by normal note letters (a-f), with capital letters representing the lower octave on a staff, and lower case the higher octave. Double-flat, flat, natural, sharp, and double-sharp accidentals are represented by prefixed "__", "_", "=", "^", and "^^". Keys, chords, note duration, and other standard musical notations are supported. But this format is not directly suited to a searchable incipit database unless something like transposition to a standard key is done.
There is a published index of Gregorian chant that encodes each note as the interval (positive or negative) between it and the previous note. For instance, an ascending scale would appear as something like "1,1,1,1,1,1,1" (the first note is of course not represented), and "Old Hundredth" would begin "0,-1,-1,-1,3,1,1,0,0,0,-1,-1,3,-1,-1". This is a useful form for searching, particularly for modal tunes.
The Hymnal 1940 Companion (Episcopal) uses a verbally-described outline, with entries like "Tunes beginning on a [weak/strong] beat, in a [major/minor] key, with first progression [upward/downward] stepwise for [3/4/5/etc.] steps, then skip [up/down] a [third/fourth/etc.]." Repeated notes are ignored. The only advantage of this verbose approach is that, like the Gregorian chant scheme, it works on files where the modality is ambiguous (e.g. predominately pentatonic Genevan psalm tunes). But I still think it would be much easier to look up a modal tune three or four times (assuming different modal possibilities) using any of the alphanumeric encodings already described.