Re: Ties and multiple voices

[Mario Lang]

David Kastrup <address@hidden> writes:

> Mario Lang <address@hidden> writes:
>
>> Hi.
>>
>> I recently noticed that LilyPond apparently can not deal with situations
>> like this:
>>
>>   <<{c1}\\{c'1~}>> | <<{c}\\{c'}>>
>
> Where is the problem supposed to be?  Wrapping this in { ... } looks
> just as expected to me.  There are two problems: the tie is flipped
> different from what would look nice, and the whole notes don't form a
> column.

Sorry, I definitely shouldn't write problem reports/questions from
memory... :-(

I see two problems related to ties that I can't really explain/fix given
my limited understanding of LilyPond:

1. Ties before a repeat and at the end of alternatives:

\version "2.14.2"
\score {
{
\clef "treble"
\key g \major
\time 3/8
g''4.~-4 \bar "|:" % 1
\repeat volta 2 {
  << { a''4.~}\\{g''8 fis''16 e'' d'' c''-2} >> |
  << {a''8 g''16 fis'' e''-5 d''}\\{b'4.~} >> |
  << {c''4.~-"3-5"}\\{b'8 a'16-2 g' fis'-3 e'} >> |
  << {c''8 b'16-4 a'-5 g' fis'}\\{d'4.~} >> |
  << {e'4.~-"2-5"}\\{d'8 c'16-2 b a-2 g} >> |
  << {e'8 d'16 c'-5 b a-3}\\{a4.~} >> |
  << {b4.~}\\{a16 g-1 b-5 a g fis} >> |
  << {b16 a-1 c'-2 b a-2 g}\\{g8-3 g'4~} >> |
  << {a8-"2-1" a'4~}\\{g'8-"5-4" fis'16 e' d' c'-2} >> |
  << {a'8 g'16-3 fis' e' d'-2}\\{b16 d'-2 g'-3 a' bes'8~} >> |
  << {cis'16 e'-2 a'-3 b' c''8~}\\{bes'8 a'16-3 g' fis' e'-2} >> |
  << {c''8 b'16-4 a'-5 g' fis'}\\{d'4.~} >> |
  << {e'4.~-5}\\{d'16 cis'-3 b d' cis' b} >> |
  << {e'16 d' cis' e'-5 d'-5 cis'-4}\\{cis'16 b a g-1 fis-2 e} >> |
}
\alternative {
  {<< {d'16 c' b a g  fis-1}\\{ fis8 r r} >> |  g''4.~-4  | }
  {<d' fis>4. | a''4.~-4  | }
}
}
}

I get two errors in this snippet:
* The g''4 in the first measure is not connected to the g''8 in the second
voice of the second measure (which is inside of the repeat).  At least I
guess so, its reported as unterminated tie.
* The g''4. in the first alternative is detected as unterminated tie.
I guess this is actually OK, but I am not 100% sure, mostly because all
I can work with is LilyPond sources and braille music notation, I can
not see the output of LilyPond itself (visual music notation).
 
> However, this is expected as the voices are _inverted_: the implicit
> \voiceOne is in a pitch _below_ the implicit \voiceTwo.

That is quite interesting to know.  So the order of voices is important,
and depends on the pitch of the first note?  Something similar happens
in braille music, but if I am not mistaken, there, music written for the
left hand is expected to be ordered ascending.  I will probably have to
do some reordering there.

2. Ties in chords:

\version "2.14.2"
{
  \new Staff {
    \clef "bass"
    \key g \major
    \time 3/8
    \repeat volta 2 {
      g4.-1 |
      fis4.-"5-4" |
      << {e4.}\\{r8 b-1 g} >> |
      << {r8 d-4 fis}\\{d'4 r8} >> |
      << {b,4.}\\{r8 g b} >> |
      << {c4.-"4-5"}\\{e8 a g~-"2-1"} >> |
      << {d4.}\\{g8 fis16 e fis8} >> |
      << {r8 b,-4 d-1}\\{g8 d r} >> |
      << {g,8 g4~-"1-4"}\\{R1*3/8} >> |
      << {g8 fis4~}\\{r8 d'-2 fis'} >> |
      << {fis8-4 e g}\\{b8-2 e' r} >> |
      << {cis8 a-4 cis'}\\{r8 cis' e'} >> |
      <fis-5 a~-3>4. |
      << {g8 g,4~}\\{a8 g  b} >> |
       <g, e>8 <fis,-4 fis> <a,-2 e> |
    }
    \alternative {
      {<< {d,8. d16-4 e fis}\\{d4 r8 } >> | }
      {<< {d,8. a16-4 b cis'}\\{d4 r8 } >> | }
    }
  }
}

Here the unterminated tie is inside a chord:

      <fis-5 a~-3>4. |
      << {g8 g,4~}\\{a8 g  b} >> |

And it is not recognized as being tied to the a8 in the second voice of
the following measure.

Now that I have read your explanation above, might it be because the
first note of the first voice is lower then the second voice?

How exactly is the "being lower"ness of a voice relative to another
voice calculated?  Do I need to care about voice ordering when I
autogenerated LilyPond code, or is there some magic I can enable in a
.ly file to make it irrelevant?

-- 
CYa,
  ⡍⠁⠗⠊⠕