Re: \cadenzaOff and accidentals

[David Bobroff]

On 3/5/2012 4:41 PM, David Kastrup wrote:
> James<address@hidden>  writes:
>
> > Hello,
> >
> > On 5 March 2012 15:20, David Kastrup<address@hidden>  wrote:
> > > Neil Thornock<address@hidden>  writes:
> > >
> > > > In my mind, the real bug is that \cadenzaOff does not seem to turn the
> > > > cadenza off immediately, but rather one measure later.  I'm not sure
> > > > why it behaves as it does.
> > > It resets timing to normal, and "normal" does not mean the end of a
> > > measure.
> > Normal from what perspective? Where it left off before \cadenzaOn or
> > simply that it starts back at '0' immediately \cadenzaOff
> >
> > I'm wondering if we need something (a sentence of two) in the NR to
> > avoid what is obviously something that does confuse people.
> I still have not figured out the exact reasoning behind this.  But it
> does not seem to be much more than "stop advancing time in measure"
> "start advancing time in measure again".  If there is no material left
> to fill the bar, the measure will not get full.

I've tried a couple things:

The first example uses Neil T's idea of setting \cadenzaOff a measure 
before it ends.  It is intentionally awkward timing-wise.  Note that it 
was necessary to used \cadenzaOff one note later than \cadenzaOn and 
that I also had to add a skip to be sure the barline was in the right 
place.  It seems kludgy/ugly to me:

%%%
\version "2.14.2"

{
  \time 4/4
  c1
  \cadenzaOn
  c4 ~ \cadenzaOff c8 s32
  c16-[ c c c-] c4 c32-[ c c c c c c-] c8
  \bar "|"
  c4 c4 c
}

%% This one tests David K's comment regarding
%% time stops/time starts

\relative c'
{
  c1
  c2 \cadenzaOn \teeny f32-[ a c  f, a c f, a c-] \cadenzaOff 
\normalsize c2
  c1
}
%%%


The result I get from the second one strongly indicates that \cadenzaOn 
means time stops but that \cadenzaOff means, "Once a full measure has 
passed *from this point* the next one starts."  If it was as simple as 
time stops/starts (the way I'm intepreting it) then the first part of 
measure two of the second example would be counted and the last half 
would be counted and nothing inside the cadenza would be counted.  This 
is not what happens, though.  It would be necessary to use 'c2*2' at the 
end of that bar as David K suggested earlier.  It works and is much less 
ugly than the top example but strikes me as less than satisfactory.

I'm trying to avoid being a choosy beggar here as I certainly don't have 
the programming chops to dig in and fix this myself.  I suspect, 
however, that "remembering" the portion of a measure preceding 
\cadenzaOn and then adding it to what comes after \cadenzaOff could be 
complicated.  I can also easily imagine a situation where that approach 
would fail.  Instead, would it not be useful to have some way to signal 
that the measure is over and a new measure is to begin, complete with 
barline and accidental behavior?  I can also well imagine a scenario 
where having \cadenzaOff doing this would also fail.  I'm thinking 
something like '\endMeasure' to accomplish this.  I know that's only 
syntax and not a solution.  I'm just wondering if leaving \cadenzaOn 
\cadenzaOff as they are and having another mechanism take care of 
initiating the next measure would be the solution.

In the meantime, I'll just have to remember these workarounds.

-David