A Morley Canzonet in ABC

I've put a fair amount of work into
typesetting some of Thomas Morley's "Canzonets for Three voices" so that
the group I sing and play recorders with can perform from unbarred
parts without having to deal with the facsimile. Here's
what the cantus part looks like (size 194773) as Morley
published it.

There are a number of differences between the way Thomas Morley
published his canzonets and the way they are presented in even a
very good modern edition.

  • There were no scores, showing all the parts and the precise
    relationship between them.
  • There were no bar lines in the parts.
  • Musicians were expected to read a large number of clefs, and
    the clef for a given part (or even line within the part) was
    chosen so that there would be no or very few ledger lines
    below or above the staff.

Other than these differences, most of the notation by the late
sixteenth century when Thomas Morley was writing was quite
similar to what is in use now, and a musician trained in modern
notation can fairly easily read a line from a facsimile.
However, for a group of amateur players and singers to actually
play from the facsimile presents a number of problems:

  • They may have difficulty dealing with clefs that are not the
    standard G clef on the second line or F clef on the fourth line.
  • Frequently the printing, or the reproduction or corruption
    of the printing leaves the performer in doubt about the actual
    notation. For instance, is a particular mark a rest, a dot, or
    a smudge? Does a note stem have 1 or 2 flags on it?
  • Where, besides the beginning, can you restart
    the piece?

page presents the results of transcribing the
facsimile in abc.

I've been thinking about trying to do this for years, and have been
concentrating on the TeX-based music publishing because it seems to
not be as completely dependant on the measure as the unit it deals
with as most of the other music printing programs I know about. In
Renaissance music, there was no such thing as a measure, and the unit
people thought in was the beat, which was usually either a half note
or a whole note. If you try to write this music with barlines the way
most modern editions do, you end up needing to have a lot of ties
across bars. This is a problem for the performers because it is
hard to realize that this note that's a quarter note tied to a half
note is really the same kind of note as the dotted half note.

People play better from parts than from scores, partly because they
have to concentrate harder, and partly because the score implies that
the performer is responsible for hitting the first beat of the measure
at the same time as everybody else, which is true for most music later
than about 1700, but not for renaissance music. In the Renaissance,
the performer's responsibility was to feel the same beat as everybody
else, and within this beat to put the melodic line across as well as
possible. Because the harmonic vocabulary is not very large, the
music can work quite well even if some parts are doing a ritard when
other parts aren't, as long as they all cadence together.

Here's what I do in ABC. (Note that this is no longer the way I
actually transcribe, but if you like ABC, or find
lilypond intimidating, it's still a possible way to
do it, so I'm leaving it in here. LEC)

  • First I type in the music with the words, one line at a time,
    leaving the line breaks where they are in the original I am
    typing from. I find I can touch type both words and music,
    and then line them up. I type each part into a separate
    file. Here's the file for the top line of the Morley:
    The part for the top line (cantus).
    Note that I retain as much as possible of Morley's
    spelling, because I find the Elizabethan laissez
    faire attitude fascinating. I do try to
    avoid things which would confuse me when singing, such as
    the use of u where we would use v, or i where we would use j.
  • Then I use midi2abc and playmidi to proofread each part. I
    also use abc2ps to print out a part and compare it visually
    with the original.
    Here's what the printout looks like:
    Cantus part.
    Cantus part. (midi file 2425)
  • Then I run a script (makescore)which
    creates a score. (The score in abc with all three
  • So far this has
    always needed more corrections. I find these iterations are
    less tedious because I have the instructions for what the computer
    does in a Makefile.
  • I use both the score and the midifile for proofreading the
    data entry of the three parts.
  • I decide where the rehearsal
    letters should go, and put them in using the syntax abc says
    is for guitar chords.
  • For the final printout I use the -c option to abc2ps to let abc decide
    where to break the lines. Here's what it
    looks like. (image 67534)
  • I also run another script, http://www.serpentpublications.org/wordpress/wp-content/uploads/2009/10/makeallparts.txt to produce a file with
    all the parts in it, for emailing or posting on the web site.
    Here's the one for See, See, Myne Owne Sweet Jewel
  • There is a snag to this method. abc2midi, the program that
    produces the midi files on this site, has a very 19th century
    view of accidentals, and assumes they are in force until the
    next barline. In the renaissance, accidentals were something
    composers put in when they felt like it, and assumed
    performers would put them in when they felt like it.
    So to get abc2midi to not carry over the accidentals to the
    end of the piece (since you haven't put in any barlines), you
    insert the line "%%MIDI nobarlines" in your ABC file.