Instrumental intro on "Piano and melody with lyrics" score in LilyPond

Question

I am creating a LilyPond piano/vocal score starting with the Template A.3.2 Piano and melody with lyrics. In my piece, there will be several bars of instrumental introduction before the vocal line comes in, and I would like the staff for the vocal to be omitted for that part.

The template produces this:

What I have been doing is:

• Pad the start of the vocal line with sufficient rests to cover the duration of the introduction.
• Change the \RemoveEmptyStaves option inside the \layout block to \RemoveAllEmptyStaves, since \RemoveEmptyStaves will not affect the first system in a score.
• Include a \break on the vocal line after the last measure of the introduction to force the change from 2 staves to 3 to occur at the right point.

I now realize that I could also do this:

• Store the notes expressions for the piano intro in separate variables, (say upperIntro = { } and lowerIntro = { }
• Include two top-level \score blocks (one with 2 staves for the intro and one with 3 staves for the remainder).

Or, as indirectly suggested by Lazy's answer:

• Put the 2-staff intro and 3-staff main part into their own << ... >> blocks. (This does not seem to work inside a \score block– it throws an error and only outputs the intro part.)

As far as I can tell, the only visual difference is that the version with a single \score block and break justifies the output for each section:

while the version with two \score blocks (or two << ... >> blocks) left-aligns both sections:

Is there any other reason to prefer one method over the others, i.e. are there any side-effects to the choice that impact flexibility, modularity, appearance, etc.?

Are there other, better ways to accomplish the same thing?

Answer 1

Well, you can make the second version look like the first one. But I advocate for using \RemoveAllEmptyStaves, unless you really want to separate the introduction and the main part. Doing a break before the voice starts is not really nescessary, because even if the break is before the part you will just get a few rests in the voice. Manually specifying breaking might be nescessary if the break falls just slightly after the start of the vocal line. Else doing lots of breaks is just detrimental to spacing. Also a possibility instead of manual breaks is using \noBreak in places where a break is not good.

For example look at this:

<<
  \new Staff \with { \RemoveAllEmptyStaves }
    {
    R1*8 c''4 d'' e'' f'' c''4 d'' e'' f'' c''4 d'' e'' f'' c''4 d'' e'' f'' \repeat unfold 8 { c''4 d'' e'' f'' }}
  \new PianoStaff
  <<
    \new Staff \repeat unfold 20 { c' d' e' f' }
    \new Staff {\clef bass \repeat unfold 20 { c2 g }}
  >>
>>

<<
  \new Staff \with { \RemoveAllEmptyStaves }
    {
    R1*8 \break c''4 d'' e'' f'' c''4 d'' e'' f'' c''4 d'' e'' f'' c''4 d'' e'' f'' \repeat unfold 8 { c''4 d'' e'' f'' }}
  \new PianoStaff
  <<
    \new Staff \repeat unfold 20 { c' d' e' f' }
    \new Staff {\clef bass \repeat unfold 20 { c2 g }}
  >>
>>

here the vocal line would start two bars in, which is not bad. But specifying \break forces lilypond to make the piano very dense in the first system.