ABSTRACT:
I've thought of and implemented a way to enter a polyphonic piece
voice by voice, i.e. all voices are entered and edited virtually
independently from each other. This does not use a preprocessor; it
is "raw" TeX code which is \input into the score source.
If you're not interested in this at all, e.g. if you're happy with
some of the available MusiXTeX preprocessors, I recommend to stop
reading here.
Currently, there seem to be mainly two ways of using MusiXTEX:
Entering the score as "plain" MusiXTeX code, which gives you the full
power of typesetting virtually everything, even beyond the scope of
MusiXTeX itself. The other way is using some preprocessor, which may
dramatically simplify the task of translating written music to
computer readable coding.
I admit that I haven't succeeded in getting familiar with M-Tx or PMX.
I've tried both several times, but my desire to manually interfere
PMX's layout decisions has always led me to make heavy use of inline
TeX commands. This gobbled up one main advantage of those preprocessor
input languages, namely the readability of the block-wise input
structure. (PLEASE, authors of and contributors to preprocessors,
don't feel offended by this statement of mine. This isn't meant to be
criticism of your work at all - it's simply my personal working style
which makes me an outsider of the preprocessors' user community.)
Thus, as a fan of MusiXTeX's power mentioned above, I thought of still
another alternative to the tedious "plain" MusiXTeX coding. What bored
me most is the fact that music has to be entered chord-wise or - let's
rather say - column-wise. For polyphonic pieces, this implies that you
can't consider the rhythmic structure of a piece voice by voice (as
you would do, for example, by entering a piece by voice-wise real-time
MIDI playing); instead, for each "column", you have to have in mind
the rhythm sequence of all voices simultaneously. This makes music
entry and error correction for polyphonic voicings rather an effort.
Therefore, I wanted to address the following goals:
- providing voice-wise music entry
- retaining full control over all typesetting capabilities of MusiXTeX
Things that I've NOT intended to alter:
- requirement of MusiXTeX knowledge
- all layout decisions (line breaks, ...) made directly by
MusiXTeX/musixflx
Experiments with those topics resulted in a set of TeX macros which
can be \input into a MusiXTeX score source. The main feature is that
all voices are entered separately from each other, each voice as a
sequence of note, pause etc. commands. A brief example:
\setvoice21{+4 \qa g+3/8 \qup c}
Which means: Voice 1 of instrument 2 consists of a "\qa g" with 1/4
note time value followed by a "\qup c" with 3/8 notes time
value. Things to be explained:
- The "+" sign is a key character indicating the beginning of a new
"note event" of the voice.
- Why explicitly specifying an event's time value although this time
value can already be retrieved from the note command, e.g. "\qup"?
There are several reasons. First, the relation between printed note
and "logical" time value isn't always one-to-one (consider heads of
beamed notes, for example). Second, and more important: You can
define own macros, e.g. for chords, and use them within the
\setvoice definition.
After having set up all voices with \setvoice, they are automatically
merged together to build the actual score. Bars are set automatically
once you have specified a bar's time value.
An important aspect which the above example reveals: The \setvoice
feature does NOT reduce the amount of typing work; instead, it imposes
a "serial" structure on the native "vertical" input style of plain
MusiXTeX. This shall simplify entering the music by setting the focus
on the voice-wise structure of a polyphonic piece. Everything else
remains the same compared to plain MusiXTeX.
So, why am I writing this e-mail? I'd like to know whether some of you
are also interested in this kind of MusiXTeX add-on. If so, please
give me some feedback. If not - and I won't be unhappy about it at all
- safely forget about the whole thing.
Regards,
Rainer
