Possible bug(s) with vertical alignment on Dynamics staff

[Peter Toye]

I'm trying to align dynamics with markup text in a Dynamics staff and have found several issues:

1. Just inserting the text by attaching it to a silent note using "^" or "_" works as expected. See examples 1 & 2

2. Trying "-" is the same as "_". This isn't documented, and probably shouldn't be. But it would be useful to have this syntax for horizontally aligned text. See example 3.

3. Tweaking the Y-offset by 0.0 using "_" appears to change it to having used "^" if you see what I mean. Example 4. Tweaking it to -0.6 gives the result I want. Example 5.

4. In the Internals manual, the unit for the Y-offset property is not mentioned. This wastes a lot of time in experimentation, especially given issue 3!

Regards,

Peter
mailto:***@ptoye.com
www.ptoye.com

[Peter Toye]

Please ignore previous mail - forgot to add the scores!

I'm trying to align dynamics with markup text in a Dynamics staff and have found several issues:

1. Just inserting the text by attaching it to a silent note using "^" or "_" works as expected. See examples 1 & 2

2. Trying "-" is the same as "_". This isn't documented, and probably shouldn't be. But it would be useful to have this syntax for horizontally aligned text. See example 3.

3. Tweaking the Y-offset by 0.0 using "_" appears to change it to having used "^" if you see what I mean. Example 4. Tweaking it to -0.6 gives the result I want. Example 5.

4. In the Internals manual, the unit for the Y-offset property is not mentioned. This wastes a lot of time in experimentation, especially given issue 3!

Regards,

Peter
mailto:***@ptoye.com
www.ptoye.com


\version "2.19.52"

\language "english"

\markup{\left-align "1. What I would expect"}
\score {
\new PianoStaff <<
\new Staff {
\clef "treble"

c''2 2
}
\new Dynamics {
s2\p s2_\markup {\italic "meno " \dynamic p}
}
\new Staff {
\clef "bass"
c2 2
}
}

\markup {\left-align "2. What I would also expect"}
\score {
\new PianoStaff <<
\new Staff {
\clef "treble"

c''2 2
}
\new Dynamics {
s2\p s2^\markup {\italic "meno " \dynamic p}
}
\new Staff {
\clef "bass"
c2 2
}
}

\markup { \left-align "3. I would like this to be centered vertically. It seems to be the same as using \"_\""}
\score {
\new PianoStaff <<
\new Staff {
\clef "treble"

c''2 2
}
\new Dynamics {
s2\p s2-\markup {\italic "meno " \dynamic p}
}
\new Staff {
\clef "bass"
c2 2
}
}
\markup {\left-align "4. What I would not expect - why has the text shifted upwards?"}
\score {
\new PianoStaff <<
\new Staff {
\clef "treble"
c''2 2
}
\new Dynamics {
s2\p \override TextScript.Y-offset = #0.0 s2_\markup {\italic "meno " \dynamic p}
}
\new Staff {
\clef "bass"
c2 2
}
}

\markup {\left-align "This is what I want!"}
\score {
\new PianoStaff <<
\new Staff {
\clef "treble"
c''2 2
}
\new Dynamics {
s2\p \override TextScript.Y-offset = #-0.6 s2_\markup {\italic "meno " \dynamic p}
}
\new Staff {
\clef "bass"
c2 2
}
}

[Malte Meyn]

Hi Peter,

have you read my latest answer on the user list?
(https://lists.gnu.org/archive/html/lilypond-user/2018-08/msg00412.html)

Post by Peter Toye
1. Just inserting the text by attaching it to a silent note using "^" or "_" works as expected. See examples 1 & 2

I suppose that’s not one of your “possible bugs”, is it? ;)

Post by Peter Toye
2. Trying "-" is the same as "_". This isn't documented, and probably shouldn't be. But it would be useful to have this syntax for horizontally aligned text. See example 3.

_ and ^ set the direction property of an Event (in this case a
TextScriptEvent) to #DOWN and #UP. “-” doesn’t set it at all—this means
that LilyPond decides whether to set it down or up depending on
different things (type of grob, voice direction in polyphony, …). Your
suggestion would mean that “-” sets the direction to #CENTER. This isn’t
supported at the moment, it’ll throw an error (you can try that:
\override TextScript.direction = #CENTER).

Now what should #CENTER behave like? If there is some space like in your
example, your suggestion is legitimate IMO. But in most cases one has
only two options, up and down (f. e. think of articulations). Letting
“-” set the direction to #CENTER, even if #CENTER would mean “automatic
default behaviour” would be wrong because direction information coming
from \voiceX commands would be overridden.

Post by Peter Toye
3. Tweaking the Y-offset by 0.0 using "_" appears to change it to having used "^" if you see what I mean. Example 4. Tweaking it to -0.6 gives the result I want. Example 5.

The Y-offset is calculated (see my answer to 4. below) depending on
different things including the 'direction property. So if you tweak the
Y-offset, the direction will be ignored (at least for the positioning
question).

Post by Peter Toye
4. In the Internals manual, the unit for the Y-offset property is not mentioned. This wastes a lot of time in experimentation, especially given issue 3!

What do you mean by unit? Yes, the IR
(http://lilypond.org/doc/v2.19/Documentation/internals/user-backend-properties)
doesn’t say whether Y-offset is measured in staff-spaces, half
staff-spaces or some different unit. But I suppose you mean “value”
here? That value is calculated by a procedure that uses information
about the context (type of grob, neighbouring grobs, font sizes, line
breaks, voice direction, …). That’s why the manual cannot give the
default value as a number—there is no such thing as a default value.

I hope I made clear what the problems are here. If not, just ask ;) (And
have a look at the \offset command I mentioned on the user list :))

Cheers,
Malte

[Peter Toye]

Hi Malte,

Thanks very much for this.I'd not read your comment - I've not had a Lilypond Digest yet today so didn't know you'd commented.

Your answer below is far fuller. My comments are inserted into it.

Sorry for my ignorance - I don't use LP much and this is the first time I've had to tweak anything seriously.

Best regards,

Peter
mailto:***@ptoye.com
www.ptoye.com

-------------------------

Post by Malte Meyn
Hi Peter,
have you read my latest answer on the user list?
(https://lists.gnu.org/archive/html/lilypond-user/2018-08/msg00412.html)

Post by Peter Toye
1. Just inserting the text by attaching it to a silent note using "^" or "_" works as expected. See examples 1 & 2

I suppose that’s not one of your “possible bugs”, is it? ;)

Err.... No :)> I included them to use as comparisons to the other three.

Post by Malte Meyn

Post by Peter Toye
2. Trying "-" is the same as "_". This isn't documented, and probably shouldn't be. But it would be useful to have this syntax for horizontally aligned text. See example 3.

_ and ^ set the direction property of an Event (in this case a
TextScriptEvent) to #DOWN and #UP. “-” doesn’t
set it at all—this means
that LilyPond decides whether to set it down or up depending on
different things (type of grob, voice direction in polyphony, …). Your
suggestion would mean that “-” sets the
direction to #CENTER. This isn’t
\override TextScript.direction = #CENTER).
Now what should #CENTER behave like? If there
is some space like in your
example, your suggestion is legitimate IMO. But in most cases one has
only two options, up and down (f. e. think of
articulations). Letting
“-” set the direction to #CENTER, even if
#CENTER would mean “automatic
default behaviour” would be wrong because
direction information coming
from \voiceX commands would be overridden.

In a dynamic staff there aren't engraved notes to attach the text to. My issue (which I wasn't 100% clear about) is that dynamics and text are treated differently - dynamics ( s2\p ) appear to be centred on the staff (at least, they're half-way between the two piano staves) and text is either raised or lowered. In this context, #CENTER would have a legitimate meaning. Or are the dynamics shifted and the text centered? Not easy to tell.

Post by Malte Meyn

Post by Peter Toye
3. Tweaking the Y-offset by 0.0 using "_" appears to change it to having used "^" if you see what I mean. Example 4. Tweaking it to -0.6 gives the result I want. Example 5.

The Y-offset is calculated (see my answer to 4. below) depending on
different things including the 'direction
property. So if you tweak the
Y-offset, the direction will be ignored (at
least for the positioning
question).

I didn't know that. Thanks.

Post by Malte Meyn

Post by Peter Toye
4. In the Internals manual, the unit for the Y-offset property is not mentioned. This wastes a lot of time in experimentation, especially given issue 3!

What do you mean by unit? Yes, the IR
(http://lilypond.org/doc/v2.19/Documentation/internals/user-backend-properties)
doesn’t say whether Y-offset is measured in staff-spaces, half
staff-spaces or some different unit. But I
suppose you mean “value”
here? That value is calculated by a procedure that uses information
about the context (type of grob, neighbouring
grobs, font sizes, line
breaks, voice direction, …). That’s why the manual cannot give the
default value as a number—there is no such
thing as a default value.

I meant the unit (staff-spaces, etc.). Without knowing that, it's not exactly easy to work out what value to insert to shift the text by the amount needed. Hence my comment on experimentation. Maybe it's a documentation bug.

Post by Malte Meyn
I hope I made clear what the problems are here.
If not, just ask ;) (And
have a look at the \offset command I mentioned on the user list :))

You've pointed out the problems, but not of course the answer to my new point on number 2.

I shall look at \offset. It's not mentioned in the Learning manual at all. Thanks again.

Post by Malte Meyn
Cheers,
Malte
_______________________________________________
bug-lilypond mailing list
https://lists.gnu.org/mailman/listinfo/bug-lilypond

[Thomas Morley]

Post by Peter Toye
In a dynamic staff there aren't engraved notes to attach the text to. My issue (which I wasn't 100% clear about) is that dynamics and text are treated differently - dynamics ( s2\p ) appear to be centred on the staff (at least, they're half-way between the two piano staves) and text is either raised or lowered. In this context, #CENTER would have a legitimate meaning.

I don't see any good in special-casing "-" for Dynamics-context.
The main reason why DynamicText and TextScript behave different is the
different procedure used to calculate the Y-offset.

DynamicText uses `(scale-by-font-size -0.6)` to calculate the
Y-offset. You'll find it in define-grobs.scm, the definition of
`scale-by-font-size` in output-lib.scm.
Using this for both works here:

\new Dynamics
\with { \override TextScript.Y-offset = #(scale-by-font-size -0.6) }
{
s2^\p
s2_\markup \dynamic "p"
s2^\markup { ggTT \dynamic "p" }
}



Cheers,
Harm