Lilypond: What's the right way to represent a song with verses, a chorus, melismata, and ties?

Question

Short question: What's the best way to represent a simple verse-chorus song with some obnoxious syllable-extension lines?

Long explanation:

I'm trying to use Lilypond to engrave an existing song with the following properties:

1. It has several verses, each of which is followed by the same chorus.
2. The chorus starts in the middle of a bar.
3. The verses end on a tied note across a bar boundary, followed by a rest.

When I try to use the melismata/extender bars (I think I'm using those terms right?) at the end of each verse, if I do it naïvely, they overshoot into the chorus for every verse but the first. I've found a couple solutions that work, but they each have their own downsides. And none of the techniques are actually producing the correct repetition behavior, as demonstrated by \unfoldRepeats.

My first approach, just using the default behavior of __, is based on §2.1.2 of the manual, in particular the "Lyrics and Repeats" subsection. I tried nesting various \new Lyrics declarations:

melody = \relative c' \repeat volta 2 {
  \partial 2 c4. e8 | g2 c~ | 4 r4
  \section \sectionLabel Chorus
  c2 | g2. e4 | c1~ | \partial 2 4. r8
}

automatic-melismata = <<
  \new Voice = "melody" \melody
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c d __
      }
      \new Lyrics {
        \set associatedVoice = "melody"
        \set stanza = "2."
        E f g h __
      }
    >>
    \repeat volta 2 {
      W x y z __
    }
  }
>>

This produces the following output, where the h __ in the second verse causes the melisma to extend through the first note of the chorus.

---

I did find two approaches that worked. One was to disable automatic melismata, and then specify the durations manually using the same technique as above; for some reason, though, the duration of the melisma in the first verse has to be one _ longer than in all the other verses.

manual-melismata = <<
  \new Voice = "melody" {
    \set melismaBusyProperties = #'()
    \melody
    \unset melismaBusyProperties
  }
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c d __ _
      }
      \new Lyrics {
        \set associatedVoice = "melody"
        \set stanza = "2."
        E f g h __
      }
    >>
    \repeat volta 2 {
      W x y z __ _
    }
  }
>>

The results are good, though:

The third approach is from Lazy's answer to Erik B's question "Lilypond: How to avoid overshooting melisma in last stanzas [incl. MWE and output]". The resulting code produces the same good output, but feels a lot more fragile to me:

split-verse-chorus = <<
  \new Voice = "melody" {
    \relative c' \repeat volta 2 {
      \clef treble
      \key c \major
      \time 4/4
      \partial 2
      c4. e8 | g2 c~ | 4 r4
      \new Voice = "chorus" {
        \section \sectionLabel Chorus
        c2 | g2. e4 | c1~ | \partial 2 4. r8
      }
    }
  }
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c \set associatedVoice = "chorus" d __
        W x y z __
      }
      \new Lyrics \lyricsto "melody" {
        \set stanza = "2."
        E f g h __
      }
    >>
  }
>>

I now have to intermingle the markup for the first verse and the chorus, which is harder to keep straight.

Moreover, all three of these approaches fail if I use \unfoldRepeats, as no actual repetition of the verses is unfolded. The first two clearly have the \repeat volta 2 in the chorus around the wrong place, and the third solution not only has no repetition, but when unfolding the repetitions the melisma line for the second verse becomes far too long.

The overall state of affairs is thus:

And the complete code for that example is as follows:

\version "2.24.4"

melody = \relative c' \repeat volta 2 {
  \partial 2 c4. e8 | g2 c~ | 4 r4
  \section \sectionLabel Chorus
  c2 | g2. e4 | c1~ | \partial 2 4. r8
}

automatic-melismata = <<
  \new Voice = "melody" \melody
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c d __
      }
      \new Lyrics {
        \set associatedVoice = "melody"
        \set stanza = "2."
        E f g h __
      }
    >>
    \repeat volta 2 {
      W x y z __
    }
  }
>>

manual-melismata = <<
  \new Voice = "melody" {
    \set melismaBusyProperties = #'()
    \melody
    \unset melismaBusyProperties
  }
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c d __ _
      }
      \new Lyrics {
        \set associatedVoice = "melody"
        \set stanza = "2."
        E f g h __
      }
    >>
    \repeat volta 2 {
      W x y z __ _
    }
  }
>>

split-verse-chorus = <<
  \new Voice = "melody" {
    \relative c' \repeat volta 2 {
      \clef treble
      \key c \major
      \time 4/4
      \partial 2
      c4. e8 | g2 c~ | 4 r4
      \new Voice = "chorus" {
        \section \sectionLabel Chorus
        c2 | g2. e4 | c1~ | \partial 2 4. r8
      }
    }
  }
  \new Lyrics \lyricsto "melody" {
    <<
      {
        \set stanza = "1."
        A b c \set associatedVoice = "chorus" d __
        W x y z __
      }
      \new Lyrics \lyricsto "melody" {
        \set stanza = "2."
        E f g h __
      }
    >>
  }
>>

\book {
  \paper {
    scoreTitleMarkup = \markup \column {
      \vspace #1
      \huge \italic \fromproperty #'header:piece
    }
    oddFooterMarkup = \markup \null
  }

  \score {
    \header {
      piece = \markup \wordwrap {
        Automatic melismata – second melisma too long
      }
    }
    \automatic-melismata
  }

  \score {
    \header {
      piece = \markup \wordwrap {
        UNFOLDED automatic melismata: verse lyrics don't repeat, chorus lyrics
        repeat only as verse 2, second melisma still too long
      }
    }
    \unfoldRepeats \automatic-melismata
  }

  \markup \vspace #2.5

  \score {
    \header {
      piece = \markup \wordwrap {
        Manual melismata – correct melismata, inconsistent code
      }
    }
    \manual-melismata
  }

  \score {
    \header {
      piece = \markup \wordwrap {
        UNFOLDED manual melismata: verse lyrics don't repeat, chorus lyrics repeat
        only as verse 2, code still inconsistent
      }
    }
    \unfoldRepeats \manual-melismata
  }

  \markup \vspace #2.5

  \score {
    \header {
      piece = \markup \wordwrap {
        Split verse/chorus – correct melismata, fragile code
      }
    }
    \split-verse-chorus
  }

  \score {
    \header {
      piece = \markup \wordwrap {
        UNFOLDED split verse/chorus: no lyrics repeated, second melisma covers
        whole chorus and continues into the next repeat, code still fragile
      }
    }
    \unfoldRepeats \split-verse-chorus
  }
}

In sum, as I said at the start: What's the best way to represent a simple verse-chorus song with some obnoxious syllable-extension lines?

Answer 1

If you do not like to have verse and chorus in mingling you can actually also do something like this:

melody = \relative c' \repeat volta 2 {
  \partial 2 c4. e8 | g2 c~ | 4 r4
  \section \sectionLabel Chorus
  \new Voice = "chorus" {
    c2 | g2. e4 | c1~ | \partial 2 4. r8
  }
}

\unfoldRepeats <<
  \new Voice = "melody" \melody
  \new Lyrics = "firstline" \lyricsto "melody" {
    \set stanza = "1."
    A b c d __
  }
  \new Lyrics \lyricsto "melody" {
    \set stanza = "2"
    E f g h __
  }
  \context Lyrics = "firstline" \lyricsto "chorus" {
    \repeat volta 2 {
      \set stanza = ""
      W x y z __
    }
  }
>>

Now to make this work with unfolding repeats we need to reuse the same "chorus" voice:

melody = \relative c' \repeat volta 2 {
  \partial 2 c4. e8 | g2 c~ | 4 r4
  \section \sectionLabel Chorus
  \context Voice = "chorus" {
    c2 | g2. e4 | c1~ | \partial 2 4. r8
  }
}

\unfoldRepeats <<
  \new Staff <<
    \new Voice = "melody" \melody
    \new Voice = "chorus" \skip\melody
  >>
  \new Lyrics = "firstline" \lyricsto "melody" {
    \set stanza = "1."
    A b c d __
  }
  \new Lyrics \lyricsto "melody" {
    \set stanza = "2"
    E f g h __
  }
  \context Lyrics = "firstline" \lyricsto "chorus" {
    \repeat volta 2 {
      \set stanza = ""
      W x y z __
    }
  }
>>

But note that this does not work out for the second stanza, because this one does not actually have a concept of repeats. So we’ll have to somehow force it to do that. The easiest way is probably using tags. Alternatively we might try to automatise this, which would result in a more beautiful structure, but complicated boilerplate.

So if we do:

melody = \relative c' \repeat volta 2 {
  \partial 2 c4. e8 | g2 c~ | 4 r4
  \section \sectionLabel Chorus
  \context Voice = "chorus" {
    c2 | g2. e4 | c1~ | \partial 2 4. r8
  }
}

lyricsVI = \lyricmode {
  A b c d __
}

lyricsVII = \lyricmode {
  E f g h __
}

lyricsC = \lyricmode {
  W x y z __
}

and

scorebase = <<
  \new Staff <<
    \new Voice = "melody" \melody
    \new Voice = "chorus" \skip\melody
  >>
  \new Lyrics = "firstline" \lyricsto "melody" {
    \set stanza = "1." \lyricsVI
    \tag #'unfolded {
      \set stanza = "2." \lyricsVII
    }
  }
  \tag #'repeat \new Lyrics \lyricsto "melody" {
    \set stanza = "2." \lyricsVII
  }
  \context Lyrics = "firstline" \lyricsto "chorus" {
    \repeat volta 2 {
      \set stanza = "C" \lyricsC
    }
  }
>>

we can do

\removeWithTag #'unfolded \scorebase

to get

or

\removeWithTag #'repeat \unfoldRepeats \scorebase

to get

In this particular case removing 'unfolded is not strictly necessary, as there are no notes for these lyrics to be assigned to.

As I said alternatively we could try to automatise this in different ways.

Note that also alternatively it would be possible (for more complex scores) to define your score by a music function that assembles the parts depending on an argument.