Re: PATCH: Arrowed accidentals for microtone notation

[Joseph Wakeling]

Dear Maximilian and others,

A few comments on the proposed patches for microtonal arrow notation.
I've already noted these on the Google Code Issue dedicated to this topic:
http://code.google.com/p/lilypond/issues/detail?id=694

... but obviously the -devel list is the place for serious discussion.

My own starting point on this was the quarter-tone arrow example given
in the NEWS section of LP 2.12:
http://lilypond.org/doc/v2.12/Documentation/topdocs/cb/lily-a8064505.ly

... and I began a thread on the -user mailing list discussing my
concerns and attempting to develop a solution:
http://lists.gnu.org/archive/html/lilypond-user/2009-04/msg00139.html

Briefly, the concern is this: that arrowed accidentals for quarter-tones
(and indeed other microtonal intervals) introduce some new issues
related to enharmonic equivalence.  The note a quarter-sharp above A
natural can be notated either A-natural-arrowup or A-sharp-arrowdown,
and it may be desirable to use both within the same piece of music.
Similar issues apply to quarter-flats.

The existing solution in the NEWS section does not address this,
permitting only one unique accidental for A-quarter-sharp.  I don't
think Maximilian's patches do either -- from what I can see they don't
include natural-up or -down accidentals.

I attempted to develop a solution to this with a slight cheat, by
defining a very slight difference between quarter-sharp (+999/4000
instead of +1/4), which would be notated as a natural sign with an arrow
up, and 'sharp-quarter-flat' (+1001/4000), which would be a sharp sign
with an arrow down.  A similar procedure was applied for quarter-flats.
 See:
http://lists.gnu.org/archive/html/lilypond-user/2009-04/msg00649.html

The problem with this approach was that it generated significant errors
in transposition.  It could not deal with the case when transposition
suggested a pitch alteration of, say, +4999/4000, to say nothing of the
case when pitches would be transposed by quarter-tones.  (I have to say
that the latter is a very tricky case anyway.)

Beyond the need for natural-up and natural-down as well as sharp-up and
sharp-down, my main concern for Maximilian's patches is that they will
also run into transposition issues of this kind.  Has this been tested for?

A few more general comments.  In response to Graham Percival's remarks,

> Basically, I'm concerned about mixing the presentation and
> definition (content) of pitches this closely.  But I admit that
> I'm not very familiar with this area, so I'll bow out of the
> discussion now and let others comment.

I think the key problem is that up- and down-arrows can mean different
things in different musical contexts.  The typical use is to indicate
quarter-tones, but other composers have used them to indicate
eighth-tones (e.g. Ferneyhough in several works, where arrows are
applied not just to standard accidentals but also to the standard
quarter-tone accidentals that Lilypond already supports).  There are
surely still other uses for arrow notation that I'm not familiar with.

So with that in mind, it seems to me that the optimal approach is to
have an up/down notation as the standard, and then let users redefine
the pitch modifications (i.e. the values of SHARP-MICRO-UP etc.) as
needed.  This would mean the user could get the notation to mean what
they want with a minimal level of fiddling.  Probably it would be good
if the default values were quarter-tones as this is the most common
interpretation.

Note that even with that implementation, users wanting a more
content-oriented approach could define their own pitch notations (as I
did in my patches) suitable to their approach.  Lilypond could still
include some switches, as you suggest, to define notations for various
'standard' approaches.

Hope this email can lead to some productive development on Maximilian's
patches as I am keen to see more extensive microtonal possibilities
included in Lilypond.

Best wishes,

    -- Joe