RE: "Ghost" trills

[Mark Stephen Mrotek]

David,

It compiled correctly under 2.20.

Mark

-----Original Message-----
From: lilypond-user [mailto:lilypond-user-bounces+carsonmark=ca.rr.com@gnu.org] 
On Behalf Of David Sumbler
Sent: Thursday, August 6, 2020 10:12 AM
To: lilypond-user@gnu.org
Subject: "Ghost" trills

Having implemented some of the methods shown in Vaughan's marvellous tutorial 
at https://lists.gnu.org/archive/html/lilypond-user/2020-06/msg00114.html
I have been making rapid progress in setting the orchestral piece which is my 
latest project.

But suddenly I came across a problem which, try as I might, I can't
figure out or solve.   I have managed to work around it.  Despite that
I should very much like to understand what is going on here so that I will be 
better able to deal with similar or related problems in future.
Also the work-around makes my input code considerably more complex, and may not 
work well if line breaks get moved around.

The problem is that, after using trillspanners, sometimes the trill reappears 
later on after a change of staff setup.  The code below reproduces this 
problem.  I have also attached a pdf file containing the output.

In this example, it is quite easy to get rid of the unwanted trillspanners in 
the second line by making small, apparently random, alterations in the input.  
For instance, if the 4th bar in the upper instrument is changed to 2 minims, or 
the same bar in the lower instrument is changed to a semibreve, the rogue 
trillspanners disappear.  However, it is not true  to say that the 
trillspanners always disappear whenever the rhythms in the two instruments are 
exactly the same, and of course it wouldn't be a useful solution to the problem 
even if it were true.  In fact, in my actual score the 2 instruments do have 
identical rhythms, but I still get these unwanted trills in the next system.

My work-around, incidentally, involves creating another pair of staves with a 
different VerticalAxisGroup.remove-layer value, and changing \instScorePlan to 
use the first pair up to and including the trills, and the replacement pair 
thereafter.  Not very elegant, and if there were many trill passages it would 
quickly become unwieldy in the extreme.

Can anyone figure out why these spare trills are appearing, and how one can get 
rid of them reasonably simply?
 
David

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\version "2.21.2"

instOne = \relative {
  a'1 | a | a |
  a1 \startTrillSpan | a1 \stopTrillSpan |
  a | a | a | a | a | a | a |
}

instTwo = \relative {
  f'1 | f | f |
  f2 \startTrillSpan f2 \startTrillSpan | f1 \stopTrillSpan |
  f | f | f | f | f | f | f |
}

targetStaff = #(define-scheme-function (ctx) (string?)
  #{
    \set Staff.keepAliveInterfaces = #'()
    \context Staff = #ctx { \unset Staff.keepAliveInterfaces }
  #})

instScorePlan = {
  \targetStaff clarinetI s1*4
  \targetStaff bothClarinets s1*3
  \break s1*5 \break s1*5
}

\score {
    \new StaffGroup \with {
    \consists "Keep_alive_together_engraver" }
  <<
    \new Staff = bothClarinets \with {
      \RemoveAllEmptyStaves
      \override VerticalAxisGroup.remove-layer = 40 }
    << \partCombine \instOne \instTwo \instScorePlan >>
    %% Clarinets on 2 staves:
    \new Staff = clarinetI \with {
      \RemoveAllEmptyStaves
      \override VerticalAxisGroup.remove-layer = 20 }
    << \instOne \instScorePlan >>
    \new Staff = clarinetII \with {
      \RemoveAllEmptyStaves
      \override VerticalAxisGroup.remove-layer = 20 }
     << \instTwo \instScorePlan >>
  >>
}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%