Allows for degree of feathering to be tweaked

Allows for degree of feathering to be tweaked

[MikeSol, 2011-03-17 11:34:47]

Makes degree of feathering tweakable.
Will need a convert-ly rule and several changes in various regtests/docs, but it
works.

Cheers,
Mike

[hanwenn, 2011-03-17 12:22:57]

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc
File lily/beam.cc (right):

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc#newcode1508
lily/beam.cc:1508: Interval feather_interval = robust_scm2interval
(me->get_property ("feather-direction"), Interval (0.0, 0.0));
rename the property to something saying interval.

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc#newcode1510
lily/beam.cc:1510: int feather_dir = feather_dif == 0.0 ? 0 : int (feather_dif /
fabs (feather_dif));
sign() ?

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc#newcode1518
lily/beam.cc:1518: xl, xr, Direction (feather_dir),
this still supposes that there is just one fixed amount of featheriness?

http://codereview.appspot.com/4289052/diff/2001/scm/output-lib.scm
File scm/output-lib.scm (right):

http://codereview.appspot.com/4289052/diff/2001/scm/output-lib.scm#newcode69
scm/output-lib.scm:69: (cdr feather-dir))))
why?  you're now changing the function of the 'endpoint properties to be
different than its definition.

If you need another pair to hold an intermediate value,you can create another
property.

for easier readability, consider keeping this in C++.  I dont think it will make
much of a difference, and the c -> lisp -> c jumps are harder to follow.

[MikeSol, 2011-03-17 12:46:53]

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc
File lily/beam.cc (right):

http://codereview.appspot.com/4289052/diff/2001/lily/beam.cc#newcode1518
lily/beam.cc:1518: xl, xr, Direction (feather_dir),
On 2011/03/17 12:22:57, hanwenn wrote:
> this still supposes that there is just one fixed amount of featheriness?

You're right.
I'm honestly not sure exactly how feathering effects this function.  All of the
tests I'm running seem to pass just fine, so I'm not sure if this function needs
detailed feathering info.  Can you think of a case where this function will fail
given the new feathering?  I've been running stuff like this:


\new Staff <<
  \relative c' {
    \cadenzaOn
    \override Staff . TimeSignature #'stencil = ##f
    \override Voice . Beam #'breakable = ##t
    \once \override Voice . Beam #'feather-direction = #'(1.0 . 0.0)
    a32[ b c d e f g a ]
    \once \override Voice . Beam #'feather-direction = #'(0.75 . 0.0)
    a,32[ b c d e f8 g a ]
    \once \override Voice . Beam #'feather-direction = #'(0.5 . 0.1)
    a32[ b c d e f g a ]
} >>

[hanwenn, 2011-03-17 13:27:21]

On Thu, Mar 17, 2011 at 9:46 AM,  <mtsolo@gmail.com> wrote:
>> this still supposes that there is just one fixed amount of
>
> featheriness?
>
> You're right.
> I'm honestly not sure exactly how feathering effects this function.  All
> of the tests I'm running seem to pass just fine, so I'm not sure if this
> function needs detailed feathering info.  Can you think of a case where
> this function will fail given the new feathering?  I've been running
> stuff like this:
>

See beam.cc-

  /*
    feather dir = 1 , relx 0->1 : factor 0 -> 1
    feather dir = 0 , relx 0->1 : factor 1 -> 1
    feather dir = -1, relx 0->1 : factor 1 -> 0
   */
  Real feather_factor = 1;
  if (feather_dir > 0)
    feather_factor = relx;
  else if (feather_dir < 0)
    feather_factor = 1 - relx;

it assumes the feathering is fulll across the x range.  I suspect that
the stem lengths mismatch the beams in broken beam cases.  Try setting
beam-thickness really small and see if the stems reach up to the
beams.

>  \relative c' {
>    \cadenzaOn
>    \override Staff . TimeSignature #'stencil = ##f
>    \override Voice . Beam #'breakable = ##t
>    \once \override Voice . Beam #'feather-direction = #'(1.0 . 0.0)
>    a32[ b c d e f g a ]
>    \once \override Voice . Beam #'feather-direction = #'(0.75 . 0.0)
>    a,32[ b c d e f8 g a ]
>    \once \override Voice . Beam #'feather-direction = #'(0.5 . 0.1)
>    a32[ b c d e f g a ]
> } >>
>
> http://codereview.appspot.com/4289052/
>



-- 
Han-Wen Nienhuys - hanwen@xs4all.nl - http://www.xs4all.nl/~hanwen

[mike_apollinemike.com, 2011-03-17 15:46:23]

On Mar 17, 2011, at 9:27 AM, Han-Wen Nienhuys wrote:

> On Thu, Mar 17, 2011 at 9:46 AM,  <mtsolo@gmail.com> wrote:
>>> this still supposes that there is just one fixed amount of
>> 
>> featheriness?
>> 
>> You're right.
>> I'm honestly not sure exactly how feathering effects this function.  All
>> of the tests I'm running seem to pass just fine, so I'm not sure if this
>> function needs detailed feathering info.  Can you think of a case where
>> this function will fail given the new feathering?  I've been running
>> stuff like this:
>> 
> 
> See beam.cc-
> 
>  /*
>    feather dir = 1 , relx 0->1 : factor 0 -> 1
>    feather dir = 0 , relx 0->1 : factor 1 -> 1
>    feather dir = -1, relx 0->1 : factor 1 -> 0
>   */
>  Real feather_factor = 1;
>  if (feather_dir > 0)
>    feather_factor = relx;
>  else if (feather_dir < 0)
>    feather_factor = 1 - relx;
> 
> it assumes the feathering is fulll across the x range.  I suspect that
> the stem lengths mismatch the beams in broken beam cases.  Try setting
> beam-thickness really small and see if the stems reach up to the
> beams.

I don't see any difference...

\new Staff <<
  \relative c' {
    \cadenzaOn
    \override Staff . TimeSignature #'stencil = ##f
    \override Voice . Beam #'breakable = ##t
    \override Voice . Beam #'beam-thickness = #0.1
    \override Beam #'breakable = ##t
    \once \override Voice . Beam #'feather-direction = #'(1.0 . 0.0)
    a32[ b c d e f g a ]
    \once \override Voice . Beam #'feather-direction = #'(0.75 . 0.0)
    a,32[ b c d e f8 g a ]
    \once \override Voice . Beam #'feather-direction = #'(0.5 . 0.1)
    a32[ b c d e f g a ]
    \once \override Voice . Beam #'feather-direction = #'(0.5 . 0.1)
    a32[ b c d e \bar "" \break f g a ]
    \once \override Voice . Beam #'feather-direction = #'(1.0 . 0.5)
    a32[ b c d e \bar "" \break f g a ]
} >>