Re: Turning autobeaming off for a stave

[Phil Holmes]

Carl,

Many thanks for your excellent explanation.  I've worked around the issue 
using a completely new voice, but was keen to understand.  I'm using 
LilyPond 2.12 so using the 2.12 Notation Reference.

On 3/18/10 4:49 AM, "Phil Holmes" <address@hidden> wrote:

> Carl,
>
> Thanks.  Doing that sort of fixes my problem, but set me off finding a
> number of other oddities in the way beaming works with implicit voices.
> Here we go....
>
> This is some excerpts from some real music, so isn't quite minimal but 
> does
> demonstrate the problems I've found.  I don't believe this has been 
> reported
> previously.
>
> BeamTest1 is what I originally started with (albeit I was using 
> \autoBeamOff
> and am not here).  As you see from the pdf, the beams that result aren't
> quite what would be expected - you'd expect the 5th/6th notes to be beamed
> with the 7th/8th, but for me it makes no odds since I normally manually
> beam.

This is a consequence of the difference between our beaming algorithms and
the Ross beaming algorithms.  Ross talks about beaming together the third
and fourth beats in a 4/4 measure as long as neither of the beats has 3 or
more notes.  LilyPond looks at the shortest duration in a beam.  In this
case, the shortest duration is 1/16, so LilyPond assumes that is equivalent
to 4 notes in a beat, and breaks the beats rather than combining them as a
unit.  I can't see a way to fix this, so you'll need to manually beam to
have it work the way you want it.


> However, the fact that the lyrics don't continue over the voiced
> notes is a problem for me, and I don't think this would be expected
> behaviour.

This is a known issue of the implicit voices construct, and is documented in
the version 2.13 Notation Reference.

> I simplified what I was using a little, to get rid of the dotted rhythm on
> the voiced notes, and this is what BeamTest2 shows.  You'll see that with
> straight quavers, the notes in the voices aren't beamed at all!  That 
> seems
> to me a bit weird.  Still get the problem of the lyrics, though.

This is an example of a known issue, which is described in the Notation
reference section 1.2.4:

"If a score ends while an automatic beam has not been ended and is still
accepting notes, this last beam will not be typeset at all. The same holds
for polyphonic voices, entered with << S \\ S >>. If a polyphonic voice ends
while an automatic beam is still accepting notes, it is not typeset."

In this case, the polyphonic voice ended while the autobeam is still
accepting notes, so the beam isn't typeset.

> If I skip a couple of test files to BeamTest5, you'll see that by 
> explicitly
> instantiating and naming the voices, the lyrics now carry through. 
> However,
> the beaming still isn't what would be expected - the upper voice is beamed
> and the lower is not.

Again, the same issue is at play -- the polyphonic voice ends while the beam
is still active.  The workaround is to manually beam the notes in the
polyphonic voice.

HTH,

Carl