Re: missing glissando features (bugs?)

[Marc Hohl]

Neil Puttock schrieb:
> 2009/5/15 Marc Hohl <address@hidden>:
>
>   
> > I still don't get it - when I use extra-dy = 0, I get glissando lines
> > parallel to the staff lines,
> > so a value of zero can't be right. I have attached the pdf output, the
> > upmost tab lie is the
> > standard behaviour, the middle line is extra-dy = 0, and the lowest line is
> > the one with my
> > calculated value for extra-dy.
> >     
>
> Hmm, I don't seem to be doing a very good job of explaining it. :)
>   
No, I won't go that far :-)

I wasn't just aware of the fact that your proposals with extra-dy = 0 
concerned
glissandos between different strings!
>   
> > As I can see, a value of 0 is definitely wrong here, so it seems that a
> > normal score and a tab score
> > handle extra-dy differently.
> >     
>
> Exactly: as I mentioned earlier in the thread, in engraver-init.ly
> there's a setting for 'extra-dy in the context definition for
> TabVoice, whereas there's no setting for normal glissandos.  If you
> have a look at line-spanner.cc you'll find the code which reads
> 'extra-dy: it uses robust_scm2double () to ensure a sane value is
> returned (0 in this case) if the property isn't set.
>
> To recap, the current behaviour for glissandos is as follows:
>
> 1. If a glissando is in a Voice context other than TabVoice, 'extra-dy
> = #0 unless overridden by the user.
>
> 2. In a TabVoice, 'extra-dy = #0.75 for all glissandos, even if
> they're between notes at different staff-positions.
>
> For a TabVoice glissando, we want 'extra-dy to be 0.75/-0.75 for notes
> at the same staff-position, depending on the change in pitch;
> otherwise 'extra-dy should be 0, since we don't need to add any extra
> slope to glissandos between notes at different staff-positions (or do
> we... see below ;)
>   
Great explanation - now I got it.
>   
> >     (if (and (= left-staff-position right-staff-position)
> >              (< (ly:pitch-semitones right-pitch) (ly:pitch-semitones
> > left-pitch)))
> >         -0.75
> >          0.75  )))
> >     
>
> This code can't distinguish between (equal staff-positions &
> right-pitch > left-pitch) and (different staff-positions), so will
> return 0.75 for both situations.
>
> Here's a version of the callback which will only apply 'extra-dy if
> the staff-positions are equal:
>
> #(define (glissando::calc-tab-extra-dy grob)
>    (let* ((original (ly:grob-original grob))
>           (left-bound (ly:spanner-bound original LEFT))
>           (right-bound (ly:spanner-bound original RIGHT))
>           (left-pitch (ly:event-property (event-cause left-bound) 'pitch))
>           (right-pitch (ly:event-property (event-cause right-bound) 'pitch))
>           (left-staff-position (ly:grob-property left-bound 'staff-position))
>           (right-staff-position (ly:grob-property right-bound 'staff-position))
>           (extra-dy (if (= left-staff-position right-staff-position)
>                         (if (< (ly:pitch-semitones right-pitch) 
> (ly:pitch-semitones left-pitch))
>                             -0.75
>                             0.75)
>
>                         ;; not on same staff-position -> no extra-dy
>                         0)))
>      extra-dy))
>
> Of course, you might prefer to apply 'extra-dy in all cases, since the
> above places the glissando start and end points on staff-lines (see
> attached image). :)
>   
That's the point:
First, it looks (at least for me) somewhat strange when the glissando 
lines end within the staff lines.
Second, glissandos between different strings is - if not impossible to 
play - very seldom to find, so
I would prefer the 0.75-version.

Thank you for your patience!

Marc
> Regards,
> Neil
>