Re: accidentals on merged notes

2017-08-25 Thread Thomas Morley 
2017-08-15 18:54 GMT+02:00 Knut Petersen :
> Hi Phil!
>>>
>>> regards,
>>> Hauke Rehr (from Germany)
>>
>>
>> Looks the same as https://sourceforge.net/p/testlilyissues/issues/1134/ ?
>>
> Yes, it is the same old bug/defect/limitation. Although the problem is
> documented in the notation manual I think it actually is a serious defect.
>
> The code that automatically adds accidentals does not support different
> accidentals for the same pitch at the same time. "Merged" notes are not the
> cause of the problem (btw: lilypond never merges noteheads, it prints
> multiple noteheads at the same position).
>
> Have a look at this link  and that link
> 
> to see some related examples of music written by Chopin, Ravel and Scriabin
> ...
>
> Knut


Another extreme example:
Ornstein, Wild mans dance
http://imslp.org/wiki/Wild_Men's_Dance%2C_SO_54_(Ornstein%2C_Leo)

Perhaps my old coding here (after an idea by Mike Solomon)
https://archiv.lilypondforum.de/index.php/topic,1176.msg6932.html#msg6932
may be found helpful.

Cheers,
  Harm

___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-15 Thread Knut Petersen 

Hi Phil!

regards,
Hauke Rehr (from Germany)


Looks the same as https://sourceforge.net/p/testlilyissues/issues/1134/ ?


Yes, it is the same old bug/defect/limitation. Although the problem is 
documented in the notation manual I think it actually is a serious defect.

The code that automatically adds accidentals does not support different accidentals for 
the same pitch at the same time. "Merged" notes are not the cause of the 
problem (btw: lilypond never merges noteheads, it prints multiple noteheads at the same 
position).

Have a look at this link  and that link 

 to see some related examples of music written by Chopin, Ravel and Scriabin ...

Knut



___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-13 Thread Phil Holmes 


"Hauke Rehr"  wrote in message 
news:598f1941.7090...@yahoo.de...

Hello,

searching for this bug on sf using terms “accidentals + merged”,
I found only #546 said to have been solved for ages.
Still, the output doesn’t match at least /my/ expectations:
notes should either not be merged at all, or there should be
stacked accidentals indicating which one applies to which Voice.

find attached a minimal example

If I add a ! to the note on the third beat in the upper voice,
I get at least both a natural and a sharp sign, leaving it to
the reader which belongs to which Voice, but it still doesn’t
help with the case on the second beat.

regards,
Hauke Rehr (from Germany)


Looks the same as https://sourceforge.net/p/testlilyissues/issues/1134/ ?

--
Phil Holmes



___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-13 Thread Noeck 
Hi,

some more comments:

Knut A is fine in most cases. I would guess the first accidental is for
the upper voice. However, this can get very close to such situations:
{ cisis' cis' }
if the natural sign comes first.

Knut B looks like there are 6 quarter notes. Obviously not possible in
4/4 but confusing.

Knut A seems reasonable for an automatic solution, IMHO.

In reality, I think there are more manual options:
- enharmonic changes 
- Knut A but shifting the accidentals a little bit vertically (less than
a staff space to highlight the associated voice (unless there are other
notes in a chord)
- Knut B is clearer for beamed notes
- I've seen a split stem with two noteheads before *
- Malte C as a last resort

* About the split stem: I think I've seen this being discussed on the
user list but I couldn't find it. A single stem splits like a triangle
and is connected to two noteheads with different accidentals. This is
like Knut B but it is clear that the notes belong to one point in time.

In general, I would opt for a simple automatic resolution (better than
now, e.g. Knut A) but encourage manual improvements depending on the
context.

Cheers,
Joram

___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-13 Thread Malte Meyn 



Am 13.08.2017 um 09:53 schrieb Knut Petersen:

Am 12.08.2017 um 20:56 schrieb Malte Meyn:

Am 12.08.2017 um 17:05 schrieb Hauke Rehr:

Still, the output doesn’t match at least /my/ expectations:
notes should either not be merged at all, or there should be
stacked accidentals indicating which one applies to which Voice.


How should this be a default behaviour? IMO none of those two options 
looks good and unambiguous. But here are three possible solutions to 
your problem (I like the third most):


We might decide that it is ok to provide an ambiguous engraving, but the 
engraving lilypond provides is clearly broken and should be fixed.


I agree on that and on the comments to your examples
• Hauke A (current default is buggy)
• Knut C (why Malte B is no solution)
• Malte C (don’t add staves automagically)

Some comments of mine to the other examples:
• Knut A: I agree it looks a little bit ambiguous but in chords (not in 
polyphony though) it’s the best solution for a *default* behaviour. (btw 
it has to be sharp-natural like here, not natural-sharp)
• Knut B: This works only unambiguously with much more horizontal space. 
So IMO not a good solution for default behaviour.
• Malte A: Same problem as Knut B but Knut B is better (look at the 
g-sharp/a-sharp on beat 4)
• Malte B: Not a thing for default because of cases like Knut C and the 
accidentals are too small to read.


So IMO the default should be changed from Hauke A to Knut A. This works 
only for chords but polyphony should use seperate staves here anyway.


___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-13 Thread Knut Petersen 

Am 12.08.2017 um 20:56 schrieb Malte Meyn:

Am 12.08.2017 um 17:05 schrieb Hauke Rehr:

Still, the output doesn’t match at least /my/ expectations:
notes should either not be merged at all, or there should be
stacked accidentals indicating which one applies to which Voice.


How should this be a default behaviour? IMO none of those two options looks 
good and unambiguous. But here are three possible solutions to your problem (I 
like the third most):


We might decide that it is ok to provide an ambiguous engraving, but the 
engraving lilypond provides is clearly broken and should be fixed.

Knut



\version "2.19.65"
\pointAndClickOff

\markup { "Hauke - A "}
\markup { \italic "clearly broken"}
\new Voice \relative c'' {
  <<
{ \voiceOne c4 g4 g4 ais4 }
\new Voice { \voiceTwo gis4 gis4 gis4 gis4 }
  >> \oneVoice
}



\markup { "Knut - A" }
\markup { \italic "forcing accidentals does work, but result is ambiguous"}
\relative {
  <<
{ c''4 g! g! ais } \\
{ gis4 gis! gis! gis }
  >>
}


\markup { "Knut - B" }
\markup { \italic "This is unambiguous"}
\relative {
  <<
{ c''4 \tweak NoteColumn.X-offset -3 g! \tweak NoteColumn.X-offset -3 g! ais } \\
{ gis4 \tweak Accidental.extra-offset #'(4 . 0) gis! \tweak Accidental.extra-offset #'(4 . 0) gis! gis }
  >>
}


\markup { "Malte - A" }
\markup { \italic "Unambiguous again, but I would force a " \musicglyph #"accidentals.sharp" " on beat 2"}
\relative {
  <<
{
  c''4
  \tweak NoteColumn.X-offset 2.5
  \tweak Accidental.X-offset 1.6
  g
  \tweak NoteColumn.X-offset 1.5
  \tweak Accidental.extra-offset #'(1.65 . 0)
  g!
  ais
} \\ {
  gis
  gis
  \tweak NoteHead.extra-offset #'(-1 . 0)
  \tweak Stem.extra-offset #'(-1 . 0)
  gis
  gis
}
  >>
}

\markup { "Malte - B" }
\markup { \italic "Looks good. But have a look at the next example"}
\relative {
  <<
{
  c''4
  \tweak Accidental.stencil #ly:text-interface::print
  \tweak Accidental.text
  \markup \fontsize #-5 \override #'(baseline-skip . 2) \vcenter \center-column {
\musicglyph #"accidentals.natural"
\musicglyph #"accidentals.sharp"
  }
  g
  g
  ais
} \\ {
  gis
  gis
  \tweak Accidental.stencil #ly:text-interface::print
  \tweak Accidental.text
  \markup \fontsize #-5 \override #'(baseline-skip . 2) \vcenter \center-column {
\musicglyph #"accidentals.natural"
\musicglyph #"accidentals.sharp"
  }
  gis
  gis
}
  >>
}


\markup { "Knut - C " }
\markup { \italic "This proves that stacking accidentals is not a reasonable solution"}
\relative {
  <<
{ c''4   ais } \\
{ gis4   gis }
  >>
}


\markup { "Malte - C" }
\markup { \italic "Ok, but nothing we should do automatically"}
\new StaffGroup \new Staff = "orig" \relative {
  <<
{
  c''4 c c c
  \new Staff \with {
alignAboveContext = "orig"
  } \relative {
\omit Staff.TimeSignature
\omit Staff.Clef
c'' g! g a
  }
  c c c c
} \\ {
  gis gis gis gis
  gis gis gis gis
  gis gis gis gis
}
  >>
}



issue.pdf
Description: Adobe PDF document
___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

Re: accidentals on merged notes

2017-08-12 Thread Malte Meyn 



Am 12.08.2017 um 17:05 schrieb Hauke Rehr:

Still, the output doesn’t match at least /my/ expectations:
notes should either not be merged at all, or there should be
stacked accidentals indicating which one applies to which Voice.


How should this be a default behaviour? IMO none of those two options 
looks good and unambiguous. But here are three possible solutions to 
your problem (I like the third most):


\version "2.19.63"

\relative {
  <<
{
  c''4
  \tweak NoteColumn.X-offset 2.5
  \tweak Accidental.X-offset 1.6
  g
  \tweak NoteColumn.X-offset 1.5
  \tweak Accidental.extra-offset #'(1.65 . 0)
  g!
  ais
} \\ {
  gis
  gis
  \tweak NoteHead.extra-offset #'(-1 . 0)
  \tweak Stem.extra-offset #'(-1 . 0)
  gis
  gis
}
  >>
}

\relative {
  <<
{
  c''4
  \tweak Accidental.stencil #ly:text-interface::print
  \tweak Accidental.text
  \markup \fontsize #-5 \override #'(baseline-skip . 2) \vcenter 
\center-column {

\musicglyph #"accidentals.natural"
\musicglyph #"accidentals.sharp"
  }
  g
  g
  ais
} \\ {
  gis
  gis
  \tweak Accidental.stencil #ly:text-interface::print
  \tweak Accidental.text
  \markup \fontsize #-5 \override #'(baseline-skip . 2) \vcenter 
\center-column {

\musicglyph #"accidentals.natural"
\musicglyph #"accidentals.sharp"
  }
  gis
  gis
}
  >>
}

\new StaffGroup \new Staff = "orig" \relative {
  <<
{
  c''4 c c c
  \new Staff \with {
alignAboveContext = "orig"
  } \relative {
\omit Staff.TimeSignature
\omit Staff.Clef
c'' g! g a
  }
  c c c c
} \\ {
  gis gis gis gis
  gis gis gis gis
  gis gis gis gis
}
  >>
}

___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond

accidentals on merged notes

2017-08-12 Thread Hauke Rehr 
Hello,

searching for this bug on sf using terms “accidentals + merged”,
I found only #546 said to have been solved for ages.
Still, the output doesn’t match at least /my/ expectations:
notes should either not be merged at all, or there should be
stacked accidentals indicating which one applies to which Voice.

find attached a minimal example

If I add a ! to the note on the third beat in the upper voice,
I get at least both a natural and a sharp sign, leaving it to
the reader which belongs to which Voice, but it still doesn’t
help with the case on the second beat.

regards,
Hauke Rehr (from Germany)
%{
  prints natural sign and merged g note on 2nd quarter,
  sharp and merged gis on 3rd quarter
%}
\version "2.19.58"
\new Voice \relative c'' {
  <<
{ \voiceOne c4 g4 g4 ais4 }
\new Voice { \voiceTwo gis4 gis4 gis4 gis4 }
  >> \oneVoice
}
___
bug-lilypond mailing list
bug-lilypond@gnu.org
https://lists.gnu.org/mailman/listinfo/bug-lilypond