Re: A slur across a line break with change of clef in another staff generates bad output

[Lucas Werkmeister]

Hi,

On 04.03.2018 13:54, Thomas Morley wrote:
> Hi Lucas,
>
> 2018-03-04 0:03 GMT+01:00 Lucas Werkmeister :
>> Hi everyone,
>>
>> I stumbled across a bug where Lilypond generates a bad slur if the slur
>> crosses a line break and there is a simultaneous clef change in a different
>> staff. I found a previous report for this bug in the mailing list archives
>> (2015-05-30, same subject), but with Google Code being archived and
>> SourceForge currently in Disaster Recovery mode, I’m not sure if the bug is
>> currently tracked anywhere (or even fixed in the development version), so I
>> thought I could still post my tiny(ish) example along with a workaround, and
>> perhaps it would be helpful to someone :)
> It's
> https://sourceforge.net/p/testlilyissues/issues/3287/
> not fixed yet
>
>> Example (you can replace the 7 with other numbers), with workaround
>> commented out:
>>
>> \version "2.18.2"
>> <<
>>   \new Staff {
>> s1*7 |
>> % \temporary \alterBroken control-points #'(((0.3 . 1.5) (1.3 . 2.3)
>> (2.5 . 2.3) (3.5 . 1.5)) ((3.3 . 1.5) (4.3 . 2.3) (5.5 . 2.3) (6.5 . 1.5)))
>> Slur
>> r2.. d''8( | \break
>> % \revert Slur control-points
>> c'')
>>   }
>>   \new Staff {
>> \clef "violin"
>> s1*7 s1 \clef "bass"
>>   }
>> The control points aren’t perfect,
>> since I couldn’t figure out how to get the automatically chosen control
>> points (for the second half of the broken slur, we could just reuse these,
>> since only the first half is bad, and the first half could be based on them
>> as well), so I had to try to match them by eyeballing.
> For this purpose the shape-command was developed, thanks David Nalesnik.
I wasn’t sure if it would be possible to use \shape in this case, since
I needed to not just tweak but pretty much replace the control points of
the first part, and I thought it would be more difficult to figure out
the correct offsets to an unknown starting set of control points. On the
other hand, being able to specify an empty list of adjustments for the
second part and therefore keeping the original control points is a
benefit I hadn’t considered.
> Though, you may try to set minimum-length instead, directly or via 
> \alterBroken.
> In my eyes it looks already nice then.
>
> All three possibilities demonstrated below (comment in/out to test):
>
> \version "2.18.2"
>
> <<
>   \new Staff {
> s1*7 |
> r2..
>
> %% (1) the \shape command
> \shape
>   #'(
>  ;; change cps of first part
>  ((1.8 . 0)(1.8 . 0.8)(1 . 0.8)(1 . 0))
>  ;; don't change cps of second part, thus provide an empty list
>  ()
>  )
>   Slur
>
> %% (2) override 'minimum-length
> %\once \override Slur.minimum-length = 6
>
> %% (3) \alterBroken with 'minimum-length
> %\alterBroken minimum-length #'(6 0) Slur
> d''8( | \break
> c'')
>   }
>   \new Staff {
> \clef "violin"
> s1*7 s1 \clef "bass"
>   }

Thank you very much, I really like the minimum-length tweak :) the slur
stops before the clef in the lower staff instead of going across it
(might be another effect of the same bug), but I can live with that.

>> Also, this
>> unfortunately doesn’t work if you want double slurs, since AFAIK it’s not
>> possible to set the control points in that case. But still, it’s better
>> than nothing IMHO.
> For shaping double-slurs the following may help (I used some extreme
> values for the show-effect):

In my case the minimum-length solution works for the double slur as
well, so I’ll just stick to that.

> \version "2.18.2"
>
> %% c/p from shape-function in music-functions-init.ly
> #(define (shape-curve offsets grob coords)
>   (let* ((orig (ly:grob-original grob))
>  (siblings (if (ly:spanner? grob)
>(ly:spanner-broken-into orig) '()))
>  (total-found (length siblings)))
> (define (offset-control-points offsets)
>   (if (null? offsets)
>   coords
>   (map coord-translate coords offsets)))
>
> (define (helper sibs offs)
>   (if (pair? offs)
>   (if (eq? (car sibs) grob)
>   (offset-control-points (car offs))
>   (helper (cdr sibs) (cdr offs)))
>   coords))
>
> ;; we work with lists of lists
> (if (or (null? offsets)
> (not (list? (car offsets
> (set! offsets (list offsets)))
>
> (if (>= total-found 2)
> (helper siblings offsets)
> (offset-control-points (car offsets)
>
> #(define (shape-double-slurs offsets)
>   (lambda (grob)
> (let* ((dir (ly:grob-property grob 'direction)))
>   (ly:grob-set-property! grob 'control-points '())
>   (ly:grob-set-property! grob 'after-line-breaking
> (lambda (grob)
>   (let ((cps (ly:slur::calc-control-points grob)))
> (ly:grob-set-property! grob 'control-points
>   (if (> dir 0)
>   (shape-curve (car offsets) grob cps)
>

Re: A slur across a line break with change of clef in another staff generates bad output

[Thomas Morley]

Hi Lucas,

2018-03-04 0:03 GMT+01:00 Lucas Werkmeister :
> Hi everyone,
>
> I stumbled across a bug where Lilypond generates a bad slur if the slur
> crosses a line break and there is a simultaneous clef change in a different
> staff. I found a previous report for this bug in the mailing list archives
> (2015-05-30, same subject), but with Google Code being archived and
> SourceForge currently in Disaster Recovery mode, I’m not sure if the bug is
> currently tracked anywhere (or even fixed in the development version), so I
> thought I could still post my tiny(ish) example along with a workaround, and
> perhaps it would be helpful to someone :)

It's
https://sourceforge.net/p/testlilyissues/issues/3287/
not fixed yet

>
> Example (you can replace the 7 with other numbers), with workaround
> commented out:
>
> \version "2.18.2"
> <<
>   \new Staff {
> s1*7 |
> % \temporary \alterBroken control-points #'(((0.3 . 1.5) (1.3 . 2.3)
> (2.5 . 2.3) (3.5 . 1.5)) ((3.3 . 1.5) (4.3 . 2.3) (5.5 . 2.3) (6.5 . 1.5)))
> Slur
> r2.. d''8( | \break
> % \revert Slur control-points
> c'')
>   }
>   \new Staff {
> \clef "violin"
> s1*7 s1 \clef "bass"
>   }
>>>
>
> The control points aren’t perfect,
> since I couldn’t figure out how to get the automatically chosen control
> points (for the second half of the broken slur, we could just reuse these,
> since only the first half is bad, and the first half could be based on them
> as well), so I had to try to match them by eyeballing.

For this purpose the shape-command was developed, thanks David Nalesnik.
Though, you may try to set minimum-length instead, directly or via \alterBroken.
In my eyes it looks already nice then.

All three possibilities demonstrated below (comment in/out to test):

\version "2.18.2"

<<
  \new Staff {
s1*7 |
r2..

%% (1) the \shape command
\shape
  #'(
 ;; change cps of first part
 ((1.8 . 0)(1.8 . 0.8)(1 . 0.8)(1 . 0))
 ;; don't change cps of second part, thus provide an empty list
 ()
 )
  Slur

%% (2) override 'minimum-length
%\once \override Slur.minimum-length = 6

%% (3) \alterBroken with 'minimum-length
%\alterBroken minimum-length #'(6 0) Slur
d''8( | \break
c'')
  }
  \new Staff {
\clef "violin"
s1*7 s1 \clef "bass"
  }
>>

> Also, this
> unfortunately doesn’t work if you want double slurs, since AFAIK it’s not
> possible to set the control points in that case. But still, it’s better
> than nothing IMHO.


For shaping double-slurs the following may help (I used some extreme
values for the show-effect):

\version "2.18.2"

%% c/p from shape-function in music-functions-init.ly
#(define (shape-curve offsets grob coords)
  (let* ((orig (ly:grob-original grob))
 (siblings (if (ly:spanner? grob)
   (ly:spanner-broken-into orig) '()))
 (total-found (length siblings)))
(define (offset-control-points offsets)
  (if (null? offsets)
  coords
  (map coord-translate coords offsets)))

(define (helper sibs offs)
  (if (pair? offs)
  (if (eq? (car sibs) grob)
  (offset-control-points (car offs))
  (helper (cdr sibs) (cdr offs)))
  coords))

;; we work with lists of lists
(if (or (null? offsets)
(not (list? (car offsets
(set! offsets (list offsets)))

(if (>= total-found 2)
(helper siblings offsets)
(offset-control-points (car offsets)

#(define (shape-double-slurs offsets)
  (lambda (grob)
(let* ((dir (ly:grob-property grob 'direction)))
  (ly:grob-set-property! grob 'control-points '())
  (ly:grob-set-property! grob 'after-line-breaking
(lambda (grob)
  (let ((cps (ly:slur::calc-control-points grob)))
(ly:grob-set-property! grob 'control-points
  (if (> dir 0)
  (shape-curve (car offsets) grob cps)
  (shape-curve (cdr offsets) grob cps)

shapeDoubleSlurs =
#(define-music-function (parser location offsets)(pair?)
#{
  \once \override Slur.before-line-breaking = #(shape-double-slurs offsets)
#})


\score {
  \new Staff {
\set doubleSlurs = ##t
\shapeDoubleSlurs
  #'(
 ;; upside Slur
 (
   ((0 . 0) (14 . 10)(-14 . 10)(0 . 0))
 )

 . ;; n.b it's a pair

 ;; downside Slur
 (
   ((0 . 0) (0 . -6)(-20 . -6)(0 . 0))
 )
 )
c'1( c''2 c'')
\shapeDoubleSlurs
  #'(
 ;; upside Slur
 (
   ;; first part
   ((0 . 0) (-8 . 2)(8 . 2)(-2 . 0))
   ;; second part
   ((6 . 0) (0 . 0)(0 . 0)(-6 . 0))
 )

 . ;; n.b it's a pair

 ;; downside Slur
 (
   ;; first part
   ((0 . 0) (20 . -4)(-20 . -4)(0 . 0))
   ;; second part
   ((0 . 0) (13 . -3)(-10 . -4)(0 . 0))
 

Re: A slur across a line break with change of clef in another staff generates bad output

[Richard Shann]

On Sat, 2015-05-30 at 22:23 +0200, Simon Albrecht wrote:
 Am 30.05.2015 um 18:10 schrieb Richard Shann:
  Attached is a bit of LilyPond which has a slur over a line break. At the
  same moment the clef is changing in the lower staff. The result is a
  horribly broken slur start in the top staff.
  I blindly applied the dragging-the-shape on this curve which generated
  the commented out offsets. This results in something half-way
  acceptable. The bug is sensitive to all sorts of things I think - this
  was the most I managed to pare it down - reducing the font size made it
  go away for example.
  
  Questions:
  Is there special syntax for altering the shapes of the half-slurs at
  the end and beginning of the line?
 Yes, look for \alterBroken in the docs.
  Is this a known bug?
 Yes, or at least partly so:
 https://code.google.com/p/lilypond/issues/detail?id=3287 You may
 comment in there, in case you have something to add.

Well, it looks like you have it covered there. FWIW I further whittled
my example down and found that varying the global-staff-size I could
vary the response markedly. Which is what the bug report is alluding to
when it says in tight spacing situations I guess.
Here is my short example:


\version 2.18.2
#(set-global-staff-size 36)
VoiceI = {
 fis''8 fis''8 fis''8 fis''8 fis''16( g'' a'' g'' fis''8) 
 fis''16( g''
 \break a''8) a''8 a''8 a''8 r2
}


VoiceII = {
 \clef bass r2 r4. a8(
 \clef treble d') d' a'8 d' a'8 d' a'8 r2
  }



\score {
  
   \new Staff \new Voice  \VoiceI 
   \new Staff \new Voice  \VoiceII  
 

  } 



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

Re: A slur across a line break with change of clef in another staff generates bad output

[Simon Albrecht]

Am 31.05.2015 um 10:06 schrieb Richard Shann:

On Sat, 2015-05-30 at 22:23 +0200, Simon Albrecht wrote:

Am 30.05.2015 um 18:10 schrieb Richard Shann:

Attached is a bit of LilyPond which has a slur over a line break. At the
same moment the clef is changing in the lower staff. The result is a
horribly broken slur start in the top staff.
I blindly applied the dragging-the-shape on this curve which generated
the commented out offsets. This results in something half-way
acceptable. The bug is sensitive to all sorts of things I think - this
was the most I managed to pare it down - reducing the font size made it
go away for example.

Questions:
Is there special syntax for altering the shapes of the half-slurs at
the end and beginning of the line?

Yes, look for \alterBroken in the docs.

Is this a known bug?

Yes, or at least partly so:
https://code.google.com/p/lilypond/issues/detail?id=3287 You may
comment in there, in case you have something to add.

Well, it looks like you have it covered there. FWIW I further whittled
my example down and found that varying the global-staff-size I could
vary the response markedly. Which is what the bug report is alluding to
when it says in tight spacing situations I guess.
Testing that is easiest using \paper { line-width = 50 } and compiling 
with different values.

Yours, Simon

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

Re: A slur across a line break with change of clef in another staff generates bad output

[Kieren MacMillan]

Hi Richard,

 Is this a known bug?

I don’t know for sure… but in my current score, end-of-system clefs are causing 
a lot of problems (e.g., collisions with BreathingSign, spacing oddities), so 
Lily isn’t quite Doing The Right Thing™ in there.

Perhaps after my current pressures are off, I will build a test file and submit 
a bug report highlighting what needs to be addressed w.r.t. clef changes.

Cheers,
Kieren.


Kieren MacMillan, composer
‣ website: www.kierenmacmillan.info
‣ email: i...@kierenmacmillan.info


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

Re: A slur across a line break with change of clef in another staff generates bad output

[Richard Shann]

On Sat, 2015-05-30 at 09:46 -0700, Mark Stephen Mrotek wrote:
 Richard,
 
 Tried putting the \clef treble before the a8.

yes, it is fairly sensitive to minor changes... 
I think I have found the answer to my other question - the manual says
that \shape can take a list of two sets of displacements.

 \shape #'((0.0 . 0.0) (1 . 1.2) (3 . 1.2) (4 . 1) ()) Slur

Richard

  Looks good to me, yet I do not know what other constraints you have.
 
 Mark
 
 -Original Message-
 From: lilypond-user-bounces+carsonmark=ca.rr@gnu.org 
 [mailto:lilypond-user-bounces+carsonmark=ca.rr@gnu.org] On Behalf Of 
 Richard Shann
 Sent: Saturday, May 30, 2015 9:11 AM
 To: lilypond-user@gnu.org
 Subject: A slur across a line break with change of clef in another staff 
 generates bad output
 
 Attached is a bit of LilyPond which has a slur over a line break. At the same 
 moment the clef is changing in the lower staff. The result is a horribly 
 broken slur start in the top staff.
 I blindly applied the dragging-the-shape on this curve which generated the 
 commented out offsets. This results in something half-way acceptable. The bug 
 is sensitive to all sorts of things I think - this was the most I managed to 
 pare it down - reducing the font size made it go away for example.
 
 Questions:
   Is there special syntax for altering the shapes of the half-slurs at 
 the end and beginning of the line?
   Is this a known bug?
 
 Richard
 
 



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

RE: A slur across a line break with change of clef in another staff generates bad output

[Mark Stephen Mrotek]

Richard,

Tried putting the \clef treble before the a8. Looks good to me, yet I do not 
know what other constraints you have.

Mark

-Original Message-
From: lilypond-user-bounces+carsonmark=ca.rr@gnu.org 
[mailto:lilypond-user-bounces+carsonmark=ca.rr@gnu.org] On Behalf Of 
Richard Shann
Sent: Saturday, May 30, 2015 9:11 AM
To: lilypond-user@gnu.org
Subject: A slur across a line break with change of clef in another staff 
generates bad output

Attached is a bit of LilyPond which has a slur over a line break. At the same 
moment the clef is changing in the lower staff. The result is a horribly broken 
slur start in the top staff.
I blindly applied the dragging-the-shape on this curve which generated the 
commented out offsets. This results in something half-way acceptable. The bug 
is sensitive to all sorts of things I think - this was the most I managed to 
pare it down - reducing the font size made it go away for example.

Questions:
Is there special syntax for altering the shapes of the half-slurs at 
the end and beginning of the line?
Is this a known bug?

Richard



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

Re: A slur across a line break with change of clef in another staff generates bad output

[Simon Albrecht]

Am 30.05.2015 um 18:10 schrieb Richard Shann:

Attached is a bit of LilyPond which has a slur over a line break. At the
same moment the clef is changing in the lower staff. The result is a
horribly broken slur start in the top staff.
I blindly applied the dragging-the-shape on this curve which generated
the commented out offsets. This results in something half-way
acceptable. The bug is sensitive to all sorts of things I think - this
was the most I managed to pare it down - reducing the font size made it
go away for example.

Questions:
Is there special syntax for altering the shapes of the half-slurs at
the end and beginning of the line?

Yes, look for \alterBroken in the docs.

Is this a known bug?
Yes, or at least partly so: 
https://code.google.com/p/lilypond/issues/detail?id=3287 You may 
comment in there, in case you have something to add.


Yours, Simon


Richard



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


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

Re: A slur across a line break with change of clef in another staff generates bad output

[Richard Shann]

On Sat, 2015-05-30 at 12:26 -0400, Kieren MacMillan wrote:
 Hi Richard,
 
  Is this a known bug?
 
 I don’t know for sure… but in my current score, end-of-system clefs are 
 causing a lot of problems 

right - in fact I guess you could say there are two bad things that the
end of system clef has caused here, the (badly truncated) slur in the
top part and the (rather short) slur in the other part.

Richard



 (e.g., collisions with BreathingSign, spacing oddities), so Lily isn’t quite 
 Doing The Right Thing™ in there.
 
 Perhaps after my current pressures are off, I will build a test file and 
 submit a bug report highlighting what needs to be addressed w.r.t. clef 
 changes.
 
 Cheers,
 Kieren.
 
 
 Kieren MacMillan, composer
 ‣ website: www.kierenmacmillan.info
 ‣ email: i...@kierenmacmillan.info
 



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