Re: Following voices in chords?

[David Kastrup]

Carl Sorensen <address@hidden> writes:

> On 5/7/10 7:21 AM, "David Kastrup" <address@hidden> wrote:
>> 
>> Hi, how would I do something like the following _properly_:
>> 
>> { \clef bass  <<
>>                  { <e gis b>\glissando s4 <dis a b>4 }
>>                  \new Voice { \hideNotes <gis b e>\glissando s4 <a b dis>4  }
>>                  \new Voice { \hideNotes <b e gis>\glissando s4 <b dis a>4 }
>>>> 
>> }
>> 
>> Note that this has several deficiencies: We get clashing notecolumn
>> warnings, and the s4 that is required for proper length glissando lines
>> takes musical time.  The obvious solution, writing s4*0 instead, does
>> not change the spacing at all!
>
> Have you looked at 
> <http://thread.gmane.org/gmane.comp.gnu.lilypond.general/56071/focus=56151>

Just right now, after a pointer from Kieren.

> I recognize that it takes a different tack than you want, because it only
> goes note for note instead of chord for chord.  But it shows the way to get
> the spacing you want and to avoid the clashing note columns.
>
> \once \override Glissando #'minimum-length = #5
> \once \override Glissando #'springs-and-rods =
>    #ly:spanner::set-spacing-rods
>
> is the way to get the spacing.

Ok, that helps.  Not sure I understand this, though.

> To avoid the clashing note columns, you could do
>
> \override NoteColumn #'ignore-collision = ##t
>
> before your function, and
>
> \revert NoteColumn #'ignore-collision
>
> after the function.

This does not change the composition of the chord?

> It would be pretty simple for you to adjust the inner workings of
> chord-glissando.ly to make it work by rotating the chord, rather than
> by carving out individual notes.

Well, looks like a fair piece of work.  And if one invests all this
work... I guess it would be nicer if one could write <c\glissando
e\glissando g\glissando> <d e f> and notes got matched one by one.  And
possibly let <c e g>\glissando be the same as that spelled-out first
chord.

Putting aside the obvious "patches will be thoughtfully considered" to a
later point of time, anybody with a hunch why this would be a bad idea
and/or terribly complicated to implement and/or leading to a lot of
unpredictable behavior?

Thanks,

-- 
David Kastrup