#298 Ties do not appear when shortened because of an accidental

[Anonymous]

Originally created by: *anonymous

Originally created by: *anonymous

Originally created by: gpermus@gmail.com

Steve D posted a long email that pertains to ties between single and
multiple voice sections of notation of polyphonic instruments.

He later wrote:

From my point of view, the bug, if there is a bug besides the fact that the tie is not visible because it is so short, is that the tie stops short of an accidental symbol that is not associated with the note being tied, and in addition, that note happens to be the top (or
bottom) note of a chord.

It seems to me that, ideally, when either the top or bottom note of a chord is tied, the tie should extend from notehead to notehead, or from the first notehead to an accidental (if there is one) preceding the second notehead, *if* and only if that accidental belongs to the note being tied instead of another note in close proximity (another note that is perhaps an interval of a second or third from the note being tied). Otherwise, if it is a top or bottom note of a chord, the tie should (again, merely from my current perspective) arc over (or under) the accidental symbol of another note in order to nearly connect to the notehead of the note that is being tied. (if that makes any sense) :-)

I can imagine cases where the current behavior of LilyPond (2.11.13 in this case) would seem appropriate, if the tie is between two notes that are not at the top or bottom of a chord.

[Google Importer]

Originally posted by: hanw...@gmail.com

this is too much text. Need .ly snippet.

Status: Invalid

[Google Importer]

Originally posted by: v.villenave

Yes indeed. IIUC:

\version "2.11.45"
\relative {
  s2. <ees g c>4~ | <ees g c> ~ <ees aes c>4 s2 |
}

The upper tie does not appear, though the accidental is not placed on one of the tied notes. Without the flat, everything is fine.

Summary: Ties do not appear when shortened because of an accidental
Status: Accepted

[Google Importer]

Originally posted by: n.putt...@gmail.com

I encounter this issue all the time (also with ties avoiding flags); would it be considered bad style if the tie ignored the accidental in this case, crossing the accidental's stem? I've seen many real-world cases (mostly piano music) where ties follow this rule, in order to preserve the shape.

It would be nice to have a property such as 'avoid-accidental which one could set to allow ties to cross accidentals.

Regards,
Neil

[Google Importer]

Originally posted by: ColinPKC...@gmail.com

Issue 1219 has been merged into this issue.

[Google Importer]

Originally posted by: ColinPKC...@gmail.com

Issue 840 has been merged into this issue.

Cc: lemzw...@googlemail.com

[Google Importer]

Originally posted by: brownian.box@gmail.com

Issues 840 and 1219 have other good samples.

Labels: -Priority-Low Priority-Medium
Owner: ---

[Google Importer]

Originally posted by: tdanielsmusic

I don't regard this as an enhancement. Incorrect output is produced (the tie is effectively missing), so this is a defect. A view on the correct design is needed before work can begin: should the tie cross the accidental or rise above it, or are both required, which being user-selectable, as Neil suggested earlier?

Labels: -Type-Enhancement -Priority-Medium -Engraving-nitpick Type-Defect Needs-design

[Google Importer]

Originally posted by: marcus.m...@gmail.com

My opinion, supported without exception by the first three scores I consulted (published by Henle, UE, and Bach-Gesellschaft), is that ties and flags should be allowed to cross freely. If these "collisions" are avoided, ties become displaced too far from the noteheads they are supposed to connect. This can be seen in the example pictured above, along with other possible side effects: that ties become too short, fail to align with other ties attached to the same chords, and in extreme cases may disappear.

Based on the same principle, I suggest that ties, if colliding with accidentals, try to avoid or minimize the collision by shifting horizontally and vertically within certain limits. Again, if they shift too far away from the noteheads, the collision avoidance will be bought at too high a price.

[Google Importer]

Originally posted by: PhilEHol...@googlemail.com

Think this has actually been fixed prior to 2.12. See attached image with a tie – quite high, but clearly there.

[Google Importer]

Originally posted by: marcus.m...@gmail.com

Thanks for pointing this out. It's true, the tie between the 2nd and 3rd C, though it might be too high, is no longer absent, as it was in post #3. That's an improvement, and suggests that ties might no longer become too short by avoiding accidentals -- which is the explicit subject of this issue 298 (but not of issues 840 and 1219 which have been merged into it).

Though I haven't re-tested it, I see no reason to think that ties might not still become too short by avoiding flags. I argued (in post #8) that this is true even in examples like #2 or #9 (referring to the tie between the 1st and 2nd C), but I think the severity of the ties-too-short issue would increase from "enhancement" to "defect" if there were a line-break at the first barline in #9 -- as previously demonstrated in a similar example here:
https://code.google.com/p/lilypond/issues/detail?id=1219#c8

[Google Importer]

Originally posted by: janek.li...@gmail.com

Ties can disappear even when there are no special elements to avoid.

\layout {
  \override Score.SpacingSpanner #'common-shortest-duration =
  #(ly:make-moment 1 1)
}
{
  <a' d''>4_~ q s
}

[Simon Albrecht]