Re: Typesetting bug(?) - polyphony + lyrics + \times 2/3 + \oneVoice = breakage

[Peter Wright]

Hi Kieren,

On 10/07 14:35:27, Kieren MacMillan wrote:
> > Unfortunately I don't think it helps with my real problem though,
> > which is that I want to keep on having sections with a "new"
> > sopranoOneVoice followed by unison sections, etc... and I don't
> > think it's possible to do that and still have a single \lyricsto
> > "sopranoOneVoice" definition inside the "\score".
> 
> I think you can write all of the lyrics out in two variables, and
> then combine them in the score after the fact.

Hm... I'm not sure if I completely understand you, but I'll have a
look through the docs and see if I can work it out.

> I'm not able to give you a full demo snippet right now, but look
> through the docs and examples for such things -- if you haven't
> found the answer by the time I get back from [computer] work, I'll
> whip up an example for you.

Well, I don't want to give the impression that I'm unwilling to have
a bloody good attempt at solving it myself :-)... so I've had a look,
but the closest I've found so far is the \partcombine command
described here:

http://lilypond.org/doc/v2.14/Documentation/notation/multiple-voices#automatic-part-combining

\partcombine looks pretty cool - unfortunately it seems to be unusable
with lyrics, as mentioned in the subsection "Known issues and warnings":

  Likewise, partcombining isn’t designed to work with lyrics; when one
  of the voices is explicitly named in order to attach lyrics to it,
  the partcombining stops working.


I've now had a careful read of these two sections of the manual, which
seem to be the most informative about lyrics:

http://lilypond.org/doc/v2.14/Documentation/notation/common-notation-for-vocal-music
http://lilypond.org/doc/v2.14/Documentation/notation/techniques-specific-to-lyrics

...but I don't think there's a solution in either of those. This part
of the "common notation for vocal music" section seems pretty unambiguous:

  The Voice context containing the melody to which the lyrics are
  being aligned must not have “died”, or the lyrics after that point
  will be lost. This can happen if there are periods when that voice
  has nothing to do. For methods of keeping contexts alive, see
  Keeping contexts alive. 

The "keeping contexts alive" section, however, has another
almost-workable approach:

http://lilypond.org/doc/v2.14/Documentation/notation/keeping-contexts-alive

Artificially keeping the sopranoOneVoice context alive might perhaps
be another approach, but I suspect that'd mean I wouldn't be able to
use \oneVoice and have proper "unison" sections - I'd instead have to
have all the (fake) "unison" sections be part of voiceTwo, so stems
would always point down.

...though... hmm, maybe I could use \stemNeutral...

> > But I suppose I could have a whole bunch of distinct named voices
> > - eg. "sopranoOneVoice01", "sopranoOneVoice02", ...,
> 
> Not necessary, I believe.

I hope you're right :).

Though I've now been reminded that Lilypond variables aren't allowed
to have numbers in them, so using multiple voice names like
"sopranoOneVoice01", "sopranoOneVoice02" wouldn't be an option anyway.


Pete.
-- 
"Remember, SCSI is not black magic. There are fundamental technical
reasons why it is necessary to sacrifice a goat at midnight in order
to get a SCSI device working properly."  -- Arnoud Engelfriet