Re: Appoggiature too fast

[mskala]

On Thu, 12 Sep 2024, Werner LEMBERG wrote:

>
> > I'm still hoping for a response to this, if only to confirm whether
> > others see the change in behaviour and that it's not something
> > unique to my installation.  [...]
>
> I don't have time to investigate this further, sorry (and I'm rather
> poor with MIDI).  Can you do some git bisecting to find out the
> problematic commit?

---cut---
7a31661dd33a69aece5f8c986f03b770768eafad is the first bad commit
commit 7a31661dd33a69aece5f8c986f03b770768eafad
Author: Dan Eble <nine.fierce.ballads@gmail.com>
Date:   Sun May 28 09:11:46 2023 -0400

    \initialContextFrom music

    \initialContextFrom enters the initial context of its argument and
    ignores the rest of it.  This is useful for prepending music while
    preserving the influence of the original music on the context.

        {
          \initialContextFrom \originalMusic
          \prependedMusic
          \originalMusic
          \appendedMusic
        }

    Using this in def-grace-function makes the various grace-note commands
    operate consistently though they use slightly different start music.

    Closes #6617.

 input/regression/grace-start-context.ly            | 14 ++++
 input/regression/initial-context-from-tags.ly      | 19 +++++
 input/regression/other/display-lily-tests.ly       |  2 +
 input/regression/other/grace-start-and-length.ly   |  4 +
 .../repeat-start-and-length-no-alternatives.ly     |  6 +-
 input/regression/other/start-and-length.ly         |  4 +
 lily/initial-context-music-iterator.cc             | 93
++++++++++++++++++++++
 ly/music-functions-init.ly                         | 21 +++++
 scm/define-music-display-methods.scm               | 13 ++-
 scm/define-music-types.scm                         | 11 +++
 scm/music-functions.scm                            | 16 ++--
 11 files changed, 191 insertions(+), 12 deletions(-)
 create mode 100644 input/regression/grace-start-context.ly
 create mode 100644 input/regression/initial-context-from-tags.ly
 create mode 100644 lily/initial-context-music-iterator.cc
---cut---

The comment that it "makes the various grace-note commands operate
consistently" suggests that this was a change made at least to some
extent on purpose; Dan may have thought \grace and \appoggiatura
should create the same relationship between notated and played duration.
But that's not what the documentation says they do, not what Lilypond
did for a long time previous to this commit, nor what I think would
be best.

Issue 6617 is "Various grace commands create contexts inconsistently":
   https://gitlab.com/lilypond/lilypond/-/issues/6617

It's not clear to me from the description of that issue what the issue
actually is.  There is one sentence describing the grace music data
structure but no other text.  It doesn't mention MIDI or played timing -
grace notes before or after start of following note, and duration of the
grace notes - so I'm not sure the effect on duration was intentional.

As for what historical practice of playing appoggiature may have been, see
   https://www.britannica.com/art/appoggiatura
   
https://kimchangpianostudio.com/2019/12/02/ornamentation-in-baroque-era-part-1/
   
https://carillon-scarlet-67gs.squarespace.com/blog/baroque-music-the-appoggiatura
   
https://www.8notes.com/school/lessons/all/the-appoggiatura-a-guide-to-ornamentation.asp

Those are not necessarily authoritative or primary sources - they are what
I could find with a quick Web search, but the variation among them is
typical of what I've read in more scholarly and offline sources.

It seems clear that - as I said - there's a lot of variation across time,
individual performers, and musical contexts.  In some periods there was a
distinction between "long" appoggiatura and "short" appoggiatura, but some
sources describe "short" appoggiatura as exactly the same thing as
acciaccatura, reserving the unqualified term "appoggiatura" as preferably
used for only the "long" form; and the "long" form would be a variable
fraction of the following note, often at least half, although the third
and fourth links on that list above actually have it at two thirds of the
following note when the following note is dotted (that is, the *entire*
length of the notehead minus dot, twice the former Lilypond practice).

Given that appoggiature are typically notated using eighth notes and
typically include at most two of those, 9/40 of the notated duration is
going to be at most a little less than 1/16, and unlikely to be long
enough if the following note is an eighth or longer.

Lilypond obviously can't be all things to all people here, but I think
that the way Lilypond worked prior to the above commit was better for most
users, and certainly for me, than the way it seems to work now.  I don't
think that a foolish consistency between \grace and \appoggiatura is to be
sought; they are different musical concepts with different notation and
there's no strong reason they should have the same duration effects.

-- 
Matthew Skala
mskala@ansuz.sooke.bc.ca                 People before tribes.
https://ansuz.sooke.bc.ca/