Re: Melisma with manual syllable durations

[David Wright]

On Wed 22 Feb 2017 at 16:21:27 (+0000), David Sumbler wrote:
> On Mon, 2017-02-20 at 19:49 +0100, David Kastrup wrote:
> > David Sumbler <address@hidden> writes:
> > 
> > > 
> > > On Sun, 2017-02-19 at 14:49 +0000, David Sumbler wrote:
> > > 
> > > Thank you for the responses to my question.
> > > 
> > > Unfortunately, I now realise that the question was not clearly
> > > expressed.  It was not intended to be "is there a way to get
> > > melisma in
> > > upper and lower voices", but "is there a way to get melisma marks
> > > when
> > > using manual syllable durations?"
> > > 
> > > As I said in my original post, I had tried various other ways of
> > > setting the second set of words, including the obvious one of
> > > creating
> > > a \new Voice at the relevant point of the vocal staff and using
> > > \lyricsto.  Unfortunately there was a problem with this, which was
> > > that
> > > the words were set one note late, try as I might to get around the
> > > problem.  So instead of spending possibly hours trying to get to
> > > the
> > > root of this, I decided to try manual durations, which appeared to
> > > work
> > > perfectly - until I ran into this melisma problem in the second of
> > > the
> > > relevant songs.  (There is no melisma in the first of the 2 songs.)
> > > 
> > > In view of the fact that manual durations are the first to be shown
> > > in
> > > the Notation Reference, it seems surprising if such a standard part
> > > of
> > > musical notation is not supported.
> > > 
> > > So - can anyone tell me if melisma markings be produced when using
> > > manual durations with lyrics?  And if so, how?
> > Again: set associatedVoice accordingly.  Why ask for advice if you
> > are
> > going to ignore it?
> 
> I can understand the evident irritation that my response caused, and I
> apologise.  The fact is, though, that I did not ignore the advice
> given; but I misconstrued what was being suggested and thought that I
> was being asked to use some form of \lyricsto with automatic syllable
> placement.  So again, let me apologise.
> 
> Having now tried David K's method, I can confirm that it works when
> used in my whole score (with sections of the music and the lyrics in
> various included files).
> 
> I thought it was odd, though, that I needed to name an
> "associatedVoice" in order to get the melisma marking to work, when all
> the timing information for placing the syllables is already there in
> the lyric input.
> 
> So I experimented.  I find that, not only does it not matter which
> voice or stave I name as "associatedVoice", but it does not even matter
> whether such a voice exists!  I can put associatedVoice = "Fred" (and I
> assure you I have no items named Fred in the entire piece) and it all
> works fine: the notes are correctly placed (as they always were), and
> the melisma magically appears (as it used not to do).

Re: Fred

My guess would be that LP looks for a context with that name and,
finding none, uses the last/latest one, however that's defined.
In the case of \addlyrics (my posted suggestion to you), I would
assume it does the same sort of thing, perhaps "looking" for a
context called "" or just looking "up", so to speak.

Re: does [it] not matter which voice or stave

That doesn't seem to tally with my experience. I don't want to waste a
lot of time designing diagnostic cases, but if I take a 'tenor' part¹
and mangle the name of the associated voice, the tenor words (set in
the middle of this score) suddenly switch to the alto voice above it
instead of the tenor, demonstrated by the shift in the Aflat on "boy"²
caused by the soprano's note. The extender line on "love" also shrinks
to some default length. Up until where I mangle the name, everything
is set correctly to the tenor part.

That behaviour is with \lyricsto. In the case of explicit durations,
the lyric extenders disappear and the words on melismas change from
left-aligned to centred on its first note. With 2.18, that last point
is different. It complains about having no valid associated voice to
use, and the lyrics are all left-aligned anyway and stay that way³.

> In other words, for no apparent reason just adding an associatedVoice
> (which need not actually exist) causes the melisma to work.  It has
> nothing to do with any slurred or tied notes that are in the associated
> Voice.  I tried putting a melisma after a word that was on a minim with
> no ties, slurs or extra notes, and sure enough Lilypond prints the
> melisma so long as associatedVoice is set to something (anything!)

I don't know what "putting a melisma" means.

¹ I've used SATB terminology to save explanation.
² "There" and "boy" are deliberately offset with respect to their notes.
  Barbershoppers might add an extender after "boy" to prevent the
  'altos' breathing there.
³ I can't find the change in default alignment of lyrics in the
  changes document for 2.18→2.20.


Cheers,
David.

good.png
Description: PNG image

bad.png
Description: PNG image