Dangling ties

[th-we]

How should dangling ties be encoded? In any case, the current way of converting them in chords is flawed:

<chord xml:id="m-46" dur="8" dur.ges="128p" stem.dir="down" visible="false">
    <note xml:id="m-47" dur="8" dur.ges="128p" oct="4" pname="g" pnum="67"/>
    <note xml:id="m-49" dur="8" dur.ges="128p" oct="4" pname="g" pnum="67"/>
</chord>
...
<tie xml:id="m-48" endid="#m-49" startid="#m-47"/>
<tie xml:id="m-50" startid="#m-49"/>

I.e., the first tie goes from one chord note to the other, which does not make sense. The second tie is maybe O.K., but I wonder whether it would be even more reasonable to encode it as laissez vibrer. Maybe @lv should be allowed on <tie> as well, but that's a topic for the main MEI issue tracker.

The following test case is taken from a "real life" document, but it's a little strange and probably an input error in this case:

dangling-ties.zip

[rettinghaus]

There are symbols for laissez vibrer available in Sibelius. However, they have to be positioned manually and are not exported.
LaissezVibrer.zip

[lpugin]

Is @tstamp or tstamp2 a possible solution?

[th-we]

@rettinghaus Yes, I know music fonts usually have dedicated symbols - the question in this case is more about how to handle situations where sibmei finds dangling ties.

Thinking about it: Danglign ties at the end of a repeat could actually have two intended endpoints: a note at the beginning of the repeat and a note in a second ending. Would we need multiple @endids here?

[pe-ro]

@tstamp2 can be used in combination with @startid for so-called "dangling ties" --

<tie startid="#note1" tstamp2="2.5"/>

[th-we]

@pe-ro That sounds like the solution to the problem! (@lpugin Sorry, I overlooked your earlier comment about @tstamp2.)