Re: [Lilypond-auto] Issue 34 in lilypond: Grace synchronization

[lilypond]

Comment #29 on issue 34 by address@hidden: Grace synchronization
https://code.google.com/p/lilypond/issues/detail?id=34

Ok, let me try another solution sketch here.  Inside of sequential music,  
we have grace fixups that shorten any music of non-zero duration followed  
by an arbitrary number of music with zero duration followed by a music with  
a non-zero grace duration by the specified grace duration, so that all the  
music with zero duration (like overrides etc) after it are executed at a  
point of time _before_ their nominal point of time.

Basically, we need a grace fixup for simultaneous music as well that, in  
the presence of a graceful start time in the next event, sucks forward all  
zero-length music in parallel music to occur at the same graceful start  
time.

Now here is the rub: in sequential music, the order in the music _supports_  
the contention that overrides etc occuring before a grace phrase should be  
executed _before_ the grace phrase, namely in sequence.

In simultaneous music, we have no such indication.  As an extreme example,  
if we have
<<
  \new Staff { \grace { e'8 } g'1 }
  \new Staff { \once\override NoteHead.color #red c'1 }

we don't want the \once\override to be over before it has had a chance to  
affect c'1.  On the other hand, for
<< \new Staff { \grace { e'8 } g'1 }
   \new Staff { \once\set Staff.instrumentName = "Fagott" c'1 }

we don't want to have the instrumentName setting delayed until it can no  
longer affect the staff.

So \once has to become more complex, with the _onset_ obeying simultaneous  
grace fixups, but the revert coming on time.

The reversal of \once is an obvious candidate where we don't want a  
simultaneous grace fixup to occur.  It can probably be dealt with with a  
variation on the finalization hook procedure being told the expected time  
of delivery, just requeueing when it is too early yet.

I consider it unlikely that this is the only problem though.  Something  
like c1*0 is a zero-length event.  Nobody complained so far about the  
output of, say,
{ c'1 g'8*0 \acciaccatura { e'8 f' } c''1 }
but I'm not sure everything else will be similarly good-natured.

For something like
<<
  \new Staff { c'1 \time 3/4 \acciaccatura { e'8 f' } c''2. }
  \new Staff { c'1 \time 3/4 c''2. }

we probably want grace adjustment.

And even a separate timing track like
<<
  \new Staff { c'1 \acciaccatura { e'8 f' } c''2. }
  \new Staff { c'1 c''2. }
  \new Devnull { s1 \time 3/4 s2. }

would arguably benefit from grace adjustment.  But will it be for every  
case?

--
You received this message because this project is configured to send all  
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings