Re: improved lyrics handling

[Werner LEMBERG]

> > I want this:
> > 
> >   text = \lyrics {
> >     "\textit{Er:}~Ge" -- denkst ...
> >   }
> > 
> > But Lilypond doesn't like this.  Is there currently a way to get font
> > changes?
> 
> I think you want
> 
>   \property LyricsVoice . stanza = "Er:"
>   \property LyricsVoice . StanzaNumber \set #'font-shape = #'italic
> 
> or somesuch.

No.  I need the word `Er:' typeset in italics right in front of the
word `Gedenkst'.  This is a poem by Goethe, sung by a single person:


             Erinnerung
              (Goethe)

    Er:  Gedenkst du noch der Stunden,
         wo eins zum andern drang?

    Sie: Wenn ich dich nicht gefunden,
         wär mir der Tag so lang.

    Er:  Dann herrlich! ein Selbander,
         Wie es mich noch erfreut.

    Sie: Wir irrten uns einander;
         es war die schöne Zeit.


I have to indicate that `Er:' and `Sie' aren't sung, but those words
must be firmly connected to the starting syllable.

> > The alternative would be to collect all strings enclosed in '...',
> > do a TeX run with special macros and use the returned metrics.
> 
> No. Veto. We want less TeX dependence,  not more.

I basically agree, so my suggestion with '...' to make the user decide
which length should be used for the string.  But without implementing
a built-in text processor the dependency on TeX will be there, and
denying this fact is not a good thing IMHO.  A controlled dependency,
namely to typeset lyrics, doesn't harm.  In general, all major note
editing programs lack typographic quality for lyrics, why shouldn't
LilyPond be better?

Anyway, please, *please*, add a possibility to pass raw LaTeX
formatting commands in lyrics to the output.  Consider the use of
e.g. Chinese (do you still remember my small example?) -- meanwhile, I
use the character 0x7F as a starting byte for CJK letters, followed by
up to 10 Bytes (my Russian implementation uses even more): LilyPond
would be totally confused about the real length of the string.


    Werner