LilyPond Notation Reference: 1.5.2 Multiple voices

[ << Musical notation ]	[Top][Contents][Index][ ? ]			[ Specialist notation >> ]
[ < Clusters ]		[ Up : Simultaneous notes ]	[ Single-staff polyphony > ]

1.5.2 Multiple voices

This section discusses simultaneous notes in multiple voices or multiple staves.

[ << Musical notation ]	[Top][Contents][Index][ ? ]			[ Specialist notation >> ]
[ < Multiple voices ]		[ Up : Multiple voices ]	[ Voice styles > ]

Single-staff polyphony

Explicitly instantiating voices

The basic structure needed to achieve multiple independent voices in a single staff is illustrated in the following example:

\new Staff <<
  \new Voice = "first"
    \relative { \voiceOne r8 r16 g'' e8. f16 g8[ c,] f e16 d }
  \new Voice= "second"
    \relative { \voiceTwo d''16 c d8~ 16 b c8~ 16 b c8~ 16 b8. }
>>

Here, voices are instantiated explicitly and are given names. The \voiceOne … \voiceFour commands set up the voices so that first and third voices get stems up, second and fourth voices get stems down, third and fourth voice note heads are horizontally shifted, and rests in the respective voices are automatically moved to avoid collisions. The \oneVoice command returns all the voice settings to the neutral default directions.

Temporary polyphonic passages

A temporary polyphonic passage can be created with the following construct:

<< { \voiceOne … }
  \new Voice { \voiceTwo … }
>> \oneVoice

Here, the first expression within a temporary polyphonic passage is placed into the Voice context which was in use immediately before the polyphonic passage, and that same Voice context continues after the temporary section. Other expressions within the angle brackets are assigned to distinct temporary voices. This allows lyrics to be assigned to one continuing voice before, during and after a polyphonic section:

\relative <<
  \new Voice = "melody" {
    a'4
    <<
      {
        \voiceOne
        g f
      }
      \new Voice {
        \voiceTwo
        d2
      }
    >>
    \oneVoice
    e4
  }
  \new Lyrics \lyricsto "melody" {
  This is my song.
  }
>>

Here, the \voiceOne and \voiceTwo commands are required to define the settings of each voice.

The double backslash construct

The << {…} \\ {…} >> construct, where the two (or more) expressions are separated by double backslashes, behaves differently to the similar construct without the double backslashes: all the expressions within this construct are assigned to new Voice contexts. These new Voice contexts are created implicitly and are given the fixed names "1", "2", etc.

The first example could be typeset as follows:

<<
  \relative { r8 r16 g'' e8. f16 g8[ c,] f e16 d }
  \\
  \relative { d''16 c d8~ 16 b c8~ 16 b c8~ 16 b8. }
>>

This syntax can be used where it does not matter that temporary voices are created and then discarded. These implicitly created voices are given the settings equivalent to the effect of the \voiceOne … \voiceFour commands, in the order in which they appear in the code.

In the following example, the intermediate voice has stems up, therefore we enter it in the third place, so it becomes voice three, which has the stems up as desired. Spacer rests are used to avoid printing doubled rests.

<<
  \relative { r8 g'' g  g g f16 ees f8 d }
  \\
  \relative { ees'8 r ees r d r d r }
  \\
  \relative { d''8 s c s bes s a s }
>>

In all but the simplest works it is advisable to create explicit Voice contexts as explained in Contexts and engravers and Explicitly instantiating voices.

Voice order

When entering multiple voices in the input file, use the following order:

Voice 1: highest
Voice 2: lowest
Voice 3: second highest
Voice 4: second lowest
Voice 5: third highest
Voice 6: third lowest
etc.

Though this may seem counterintuitive, it simplifies the automatic layout process. Note that the odd-numbered voices are given upstems, and the even-numbered voices are given downstems:

\new Staff <<
  \time 2/4
  { f''2 }  % 1: highest
  \\
  { c'2  }  % 2: lowest
  \\
  { d''2 }  % 3: second-highest
  \\
  { e'2  }  % 4: second-lowest
  \\
  { b'2  }  % 5: third-highest
  \\
  { g'2  }  % 6: third-lowest
>>

When a different voice entry order is desired, the command \voices may be convenient:

\new Staff \voices 1,3,5,6,4,2 <<
  \time 2/4
  { f''2 }  % 1: highest
  \\
  { d''2 }  % 3: second-highest
  \\
  { b'2  }  % 5: third-highest
  \\
  { g'2  }  % 6: third-lowest
  \\
  { e'2  }  % 4: second-lowest
  \\
  { c'2  }  % 2: lowest
>>

Note: Lyrics and spanners (such as slurs, ties, hairpins, etc.) cannot be created ‘across’ voices.

Identical rhythms

In the special case that we want to typeset parallel pieces of music that have the same rhythm, we can combine them into a single Voice context, thus forming chords. To achieve this, enclose them in a simple simultaneous music construct within an explicit voice:

\new Voice <<
  \relative { e''4 f8 d e16 f g8 d4 }
  \relative { c''4 d8 b c16 d e8 b4 }
>>

This method leads to strange beamings and warnings if the pieces of music do not have the same rhythm.

Predefined commands

\voiceOne, \voiceTwo, \voiceThree, \voiceFour, \oneVoice.

Voice styles

Voices may be given distinct colors and shapes, allowing them to be easily identified:

<<
  \relative { \voiceOneStyle d''4 c2 b4 }
  \\
  \relative { \voiceTwoStyle e'2 e }
  \\
  \relative { \voiceThreeStyle b2. c4 }
  \\
  \relative { \voiceFourStyle g'2 g }
>>

The \voiceNeutralStyle command is used to revert to the standard presentation.

Predefined commands

\voiceOneStyle, \voiceTwoStyle, \voiceThreeStyle, \voiceFourStyle, \voiceNeutralStyle.

Collision resolution

The note heads of notes in different voices with the same pitch, same note head and opposite stem direction are automatically merged, but notes with different note heads or the same stem direction are not. Rests opposite a stem in a different voice are shifted vertically. The following example shows three different circumstances, on beats 1 and 3 in bar 1 and beat 1 in bar 2, where the automatic merging fails.

<<
  \relative {
    c''8 d e d c d c4
    g'2 fis
  } \\
  \relative {
    c''2 c8. b16 c4
    e,2 r
  } \\
  \relative {
    \oneVoice
    s1
    e'8 a b c d2
  }
>>

Notes with different note heads may be merged as shown below. In this example the note heads on beat 1 of bar 1 are now merged:

<<
  \relative {
    \mergeDifferentlyHeadedOn
    c''8 d e d c d c4
    g'2 fis
  } \\
  \relative {
    c''2 c8. b16 c4
    e,2 r
  } \\
  \relative {
    \oneVoice
    s1
    e'8 a b c d2
  }
>>

Quarter and half notes are not merged in this way, since it would be difficult to tell them apart.

Note heads with different dots as shown in beat 3 of bar 1 may be also be merged:

<<
  \relative {
    \mergeDifferentlyHeadedOn
    \mergeDifferentlyDottedOn
    c''8 d e d c d c4
    g'2 fis
  } \\
  \relative {
    c''2 c8. b16 c4
    e,2 r
  } \\
  \relative {
    \oneVoice
    s1
    e'8 a b c d2
  }
>>

The half note and eighth note at the start of the second measure are incorrectly merged because the automatic merge cannot successfully complete the merge when three or more notes line up in the same note column, and in this case the merged note head is incorrect. To allow the merge to select the correct note head a \shift must be applied to the note that should not be merged. Here, \shiftOn is applied to move the top g out of the column, and \mergeDifferentlyHeadedOn then works properly.

<<
  \relative {
    \mergeDifferentlyHeadedOn
    \mergeDifferentlyDottedOn
    c''8 d e d c d c4
    \shiftOn
    g'2 fis
  } \\
  \relative {
    c''2 c8. b16 c4
    e,2 r
  } \\
  \relative {
    \oneVoice
    s1
    e'8 a b c d2
  }
>>

The \shiftOn command allows (but does not force) the notes in a voice to be shifted. When \shiftOn is applied to a voice, a note or chord in that voice is shifted only if its stem would otherwise collide with a stem from another voice, and only if the colliding stems point in the same direction. The \shiftOff command prevents this type of shifting from occurring.

By default, the outer voices (normally voices one and two) have \shiftOff specified, while the inner voices (three and above) have \shiftOn specified. When a shift is applied, voices with upstems (odd-numbered voices) are shifted to the right, and voices with downstems (even-numbered voices) are shifted to the left.

Here is an example to help you visualize how an abbreviated polyphonic expression would be expanded internally.

Note: Note that with three or more voices, the vertical order of voices in your input file should not be the same as the vertical order of voices on the staff!

\new Staff \relative {
  %% abbreviated entry
  <<
    { f''2  }  % 1: highest
    \\
    { g,2 }  % 2: lowest
    \\
    { d'2 }  % 3: upper middle
    \\
    { b2  }  % 4: lower middle
  >>
  %% internal expansion of the above
  <<
    \new Voice = "1" { \voiceOne   \shiftOff f'2 }
    \new Voice = "2" { \voiceTwo   \shiftOff g,2 }
    \new Voice = "3" { \voiceThree \shiftOn  d'2 } % shifts right
    \new Voice = "4" { \voiceFour  \shiftOn  b2  } % shifts left
  >>
}

Two additional commands, \shiftOnn and \shiftOnnn provide further shift levels which may be specified temporarily to resolve collisions in complex situations – see Real music example.

Notes are only merged if they have opposing stem directions (as they have, for example, in voices one and two by default or when the stems are explicitly set in opposite directions).

Predefined commands

\mergeDifferentlyDottedOn, \mergeDifferentlyDottedOff, \mergeDifferentlyHeadedOn, \mergeDifferentlyHeadedOff.

\shiftOn, \shiftOnn, \shiftOnnn, \shiftOff.

Selected Snippets

Additional voices to avoid collisions

In some instances of complex polyphonic music, additional voices are necessary to prevent collisions between notes. If more than four parallel voices are needed, additional voices can be added by defining a variable using the Scheme function context-spec-music.

voiceFive = #(context-spec-music (make-voice-props-set 4) 'Voice)

\relative c'' {
  \time 3/4
  \key d \minor
  \partial 2
  <<
    \new Voice  {
      \voiceOne
      a4. a8
      e'4 e4. e8
      f4 d4. c8
    }
    \new Voice {
      \voiceTwo
      d,2
      d4 cis2
      d4 bes2
    }
    \new Voice {
      \voiceThree
      f'2
      bes4 a2
      a4 s2
    }
    \new Voice {
      \voiceFive
      s2
      g4 g2
      f4 f2
    }
  >>
}

Forcing horizontal shift of notes

When the typesetting engine cannot cope, the following syntax can be used to override typesetting decisions. The units of measure used here are staff spaces.

\relative c' <<
  {
    <d g>2 <d g>
  }
  \\
  {
    <b f'>2
    \once \override NoteColumn.force-hshift = #1.7
    <b f'>2
  }
>>

Known issues and warnings

Using \override NoteColumn.ignore-collision = ##t will cause differently headed notes in different voices to merge incorrectly.

\mergeDifferentlyHeadedOn
<< \relative { c'16 a' b a } \\ \relative { c'2 } >>
\override NoteColumn.ignore-collision = ##t
<< \relative { c'16 a' b a } \\ \relative { c'2 } >>

[ << Musical notation ]	[Top][Contents][Index][ ? ]			[ Specialist notation >> ]
[ < Collision resolution ]		[ Up : Multiple voices ]	[ Automatic part combining > ]

Merging rests

When using multiple voices it is common to merge rests which occur in both parts. This can be accomplished using Merge_rests_engraver.

voiceA = \relative { d''4 r d2 | R1 | }
voiceB = \relative { fis'4 r g2 | R1 | }
\score {
  <<
    \new Staff \with {
      instrumentName = "unmerged"
    }
    <<
      \new Voice { \voiceOne \voiceA }
      \new Voice { \voiceTwo \voiceB }
    >>
    \new Staff \with {
      instrumentName = "merged"
      \consists "Merge_rests_engraver"
    }
    <<
      \new Voice { \voiceOne \voiceA }
      \new Voice { \voiceTwo \voiceB }
    >>
  >>
}

Setting the context property suspendRestMerging to ##t allows for turning off rest merging temporarily.

[ << Musical notation ]	[Top][Contents][Index][ ? ]			[ Specialist notation >> ]
[ < Merging rests ]		[ Up : Multiple voices ]	[ Writing music in parallel > ]

Automatic part combining

Automatic part combining is used to merge two separate parts of music onto a single staff. This can be especially helpful when typesetting orchestral scores. A single Voice is printed while the two parts of music are the same, but in places where they differ, a second Voice is printed. Stem directions are set up & down accordingly while Solo and a due parts are also identified and marked appropriately.

The syntax for automatic part combining is:

\partcombine musicexpr1 musicexpr2

The following example demonstrates the basic functionality, putting parts on a single staff as polyphony and setting stem directions accordingly. The same variables are used for the independent parts and the combined staff.

instrumentOne = \relative {
  c'4 d e f |
  R1 |
  d'4 c b a |
  b4 g2 f4 |
  e1 |
}

instrumentTwo = \relative {
  R1 |
  g'4 a b c |
  d4 c b a |
  g4 f( e) d |
  e1 |
}

<<
  \new Staff \instrumentOne
  \new Staff \instrumentTwo
  \new Staff \partcombine \instrumentOne \instrumentTwo
>>

Both parts have identical notes in the third measure, so only one instance of the notes is printed. Stem, slur, and tie directions are set automatically, depending on whether the parts are playing solo or in unison. When needed in polyphony situations, the first part (with context called one) gets “up” stems, while the second (called two) always gets “down” stems. In solo situations, the first and second parts get marked with “Solo” and “Solo II”, respectively. The unison (a due) parts are marked with the text “a2”.

By default, the partcombiner merges two notes of the same pitch as an a due note, combines notes with the same rhythm less than a ninth apart as chords and separates notes more than a ninth apart (or when the voices cross) into separate voices. This can be overridden with an optional argument of a pair of numbers after the \partcombine command: the first specifies the interval where notes start to be combined (the default is zero) and the second where the notes are split into separate voices. Setting the second argument to zero means that the partcombiner splits notes with an interval of a second or more, setting it to one splits notes of a third or more, and so on.

instrumentOne = \relative {
  a4 b c d |
  e f g a |
  b c d e |
}

instrumentTwo = \relative {
  c'4 c c c |
  c c c c |
  c c c c |
}

<<
  \new Staff \partcombine \instrumentOne \instrumentTwo
  \new Staff \partcombine #'(2 . 3) \instrumentOne \instrumentTwo
>>

Both arguments to \partcombine will be interpreted as separate Voice contexts, so if the music is being specified in relative mode then both parts must contain a \relative function, i.e.,

\partcombine
  \relative … musicexpr1
  \relative … musicexpr2

A \relative section that encloses a \partcombine has no effect on the pitches of musicexpr1 or musicexpr2.

In professional scores, voices are often kept apart from each other for long passages of music even if some of the notes are the same in both voices, and could just as easily be printed as unison. Combining notes into a chord, or showing one voice as solo is, therefore, not ideal as the \partcombine function considers each note separately. In this case the \partcombine function can be overridden with one of the following commands. All of the commands may be preceded with \once in order to have them only apply to the next note in the music expression.

\partcombineApart keeps the notes as two separate voices, even if they can be combined into a chord or unison.
\partcombineChords combines the notes into a chord.
\partcombineUnisono combines both voices as “unison”.
\partcombineSoloI prints only voice one, and marks it as a “Solo”.
\partcombineSoloII prints only voice two and marks it as a “Solo”.
\partcombineAutomatic ends the functions of the commands above, and reverts back to the standard \partcombine functionality.

instrumentOne = \relative c' {
  \partcombineApart c2^"apart" e |
  \partcombineAutomatic e2^"auto" e |
  \partcombineChords e'2^"chord" e |
  \partcombineAutomatic c2^"auto" c |
  \partcombineApart c2^"apart" \once \partcombineChords e^"chord once" |
  c2 c |
}
instrumentTwo = \relative {
  c'2 c |
  e2 e |
  a,2 c |
  c2 c' |
  c2 c |
  c2 c |
}

<<
  \new Staff { \instrumentOne }
  \new Staff { \instrumentTwo }
  \new Staff { \partcombine \instrumentOne \instrumentTwo }
>>

Using \partcombine with lyrics

The \partcombine command is not designed to work with lyrics; if one of the voices is explicitly named in order to attach lyrics to it, the partcombiner will stop working. However, this effect can be achieved using a NullVoice context. See Polyphony with shared lyrics.

Selected Snippets

Combining two parts on the same staff

The part combiner tool ( \partcombine command ) allows the combination of several different parts on the same staff. Text directions such as “solo” or “a2” are added by default; to remove them, simply set the property printPartCombineTexts to f. For vocal scores (hymns), there is no need to add “solo/a2” texts, so they should be switched off. However, it might be better not to use it if there are any solos, as they won’t be indicated. In such cases, standard polyphonic notation may be preferable.

This snippet presents the three ways two parts can be printed on a same staff: standard polyphony, \partcombine without texts, and \partcombine with texts.

musicDown = \relative c'' {
  g4 e4.( d8) c4 |
  r2 g'4( f8 e) |
  d2 \stemDown a
}

\score {
    <<
    \new Staff \with { instrumentName = #"Standard polyphony" }

      << \musicUp \\ \musicDown >>

    \new Staff \with {
      instrumentName = #"PartCombine without text"
      printPartCombineTexts = ##f
    }

    \partcombine \musicUp \musicDown

    \new Staff \with { instrumentName = #"PartCombine with text" }
      \partcombine \musicUp \musicDown
    >>
  \layout {
    indent = 6.0\cm
    \context {
      \Score
      \override SystemStartBar.collapse-height = #30
    }
  }
}

Changing partcombine texts

When using the automatic part combining feature, the printed text for the solo and unison sections may be changed:

\new Staff <<
  \set Staff.soloText = #"girl"
  \set Staff.soloIIText = #"boy"
  \set Staff.aDueText = #"together"
  \partcombine
    \relative c'' {
      g4 g r r
      a2 g
    }
    \relative c'' {
      r4 r a( b)
      a2 g
    }
>>

Known issues and warnings

All \partcombine… functions can only accept two voices.

\partcombine… functions cannot be placed inside a \tuplet or \relative block.

If printPartCombineTexts is set and the two voices play the same notes “on and off”, in the same measure, the part combiner may typeset a2 more than once in that measure.

\partcombine only knows when a note starts in a Voice; it cannot, for example, remember if a note in one Voice has already started when combining notes that have just started in the other Voice. This can lead to a number of unexpected issues including “Solo” or “Unison” marks being printed incorrectly.

\partcombine keeps all spanners (slurs, ties, hairpins, etc.) in the same Voice so that if any such spanners start or end in a different Voice, they may not be printed properly or at all.

If the \partcombine function cannot combine both music expressions (i.e., when both voices have different durations), it will give the voices, internally, its own custom names: one and two respectively. This means if there is any “switch” to a differently named Voice context, the events in that differently named Voice will be ignored.

Refer also to Known issues and warnings when using \partcombine with tablature in Default tablatures and the Note in Automatic beams when using automatic beaming.

[ << Musical notation ]	[Top][Contents][Index][ ? ]			[ Specialist notation >> ]
[ < Automatic part combining ]		[ Up : Multiple voices ]	[ Staff notation > ]

Writing music in parallel

Music for multiple parts can be interleaved in input code. The function \parallelMusic accepts a list with the names of a number of variables to be created, and a musical expression. The content of alternate measures from the expression become the value of the respective variables, so you can use them afterwards to print the music.

Note: Bar checks | must be used, and the measures must be of the same length.

\parallelMusic #'(voiceA voiceB voiceC) {
  % Bar 1
  r8 g'16 c'' e'' g' c'' e'' r8 g'16 c'' e'' g' c'' e'' |
  r16 e'8.~   4              r16 e'8.~   4              |
  c'2                        c'2                        |

  % Bar 2
  r8 a'16 d'' f'' a' d'' f'' r8 a'16 d'' f'' a' d'' f'' |
  r16 d'8.~   4              r16 d'8.~   4              |
  c'2                        c'2                        |

}
\new StaffGroup <<
  \new Staff << \voiceA \\ \voiceB >>
  \new Staff { \clef bass \voiceC }
>>

Relative mode may be used. Note that the \relative command is not used inside \parallelMusic itself. The notes are relative to the preceding note in the voice, not to the previous note in the input – in other words, relative notes for voiceA ignore the notes in voiceB.

\parallelMusic #'(voiceA voiceB voiceC) {
  % Bar 1
  r8 g16 c e g, c e r8 g,16 c e g, c e  |
  r16 e8.~ 4        r16 e8.~  4         |
  c2                c                   |

  % Bar 2
  r8 a,16 d f a, d f r8 a,16 d f a, d f |
  r16 d8.~  4        r16 d8.~  4        |
  c2                 c                  |

 }
\new StaffGroup <<
  \new Staff << \relative c'' \voiceA \\ \relative c' \voiceB >>
  \new Staff \relative c' { \clef bass \voiceC }
>>

This works quite well for piano music. This example maps four consecutive measures to four variables:

global = {
  \key g \major
  \time 2/4
}

\parallelMusic #'(voiceA voiceB voiceC voiceD) {
  % Bar 1
  a8    b     c   d     |
  d4          e         |
  c16 d e fis d e fis g |
  a4          a         |

  % Bar 2
  e8      fis  g     a   |
  fis4         g         |
  e16 fis g  a fis g a b |
  a4           a         |

  % Bar 3 ...
}

\score {
  \new PianoStaff <<
     \new Staff {
       \global
       <<
         \relative c'' \voiceA
         \\
         \relative c'  \voiceB
       >>
     }
     \new Staff {
       \global \clef bass
       <<
         \relative c \voiceC
         \\
         \relative c \voiceD
       >>
     }
  >>
}

1.5.2 Multiple voices

Single-staff polyphony

Explicitly instantiating voices

Temporary polyphonic passages

The double backslash construct

Voice order

Identical rhythms

Predefined commands

See also

Voice styles

Predefined commands

See also

Collision resolution

Predefined commands

Selected Snippets

See also

Known issues and warnings

Merging rests

Automatic part combining

Using \partcombine with lyrics

Selected Snippets

See also

Known issues and warnings

Writing music in parallel

See also

LilyPond — Notation Reference v2.21.0 (开发分支).