Key signature and staff size

[Schneider]

Hi Octo et all,

When mixing two different staff sizes, please consider these two key signature options:

keySig.png (32.65 KiB) Viewed 13366 times

Which one's correct?

Cheers,
Pierre

[MalteM]

I think the first one is correct. At least Henle (Mendelssohn Piano Trios, 1977) and Sikorski (Shostakovich Piano Trio, 1962) do it that way.

Edit: I had a closer look: The truth lies inbetween: The spacing of the smaller key signature is a bit wider in relation to the glyph size but narrower in absolute measure.

[John Ruggero]

What an interesting observation, Schneider. One learns something new every day about music notation. I do not find this discussed in either Gould or Ross. What I see actual practice agrees with MalteM. Engravers strike a compromise so the key signatures do not differ too much in total width, as in the first example, but the accidentals themselves are not too widely spaced as in the second. Anyone know how to do this in Finale? And does Dorico do this by default?

[Schneider]

Hi Malte, hi John,
I'd be surprised if Dorico and Finale -- or even Score and Graphire -- would do differently...
Anyway, what I found out is that small key signatures padding are not steady regarding the dozen of chamber scores seen on IMSLP.
But those are pretty old ones (before 1920).
Some ex.

ex2.png (172.91 KiB) Viewed 13327 times

ex3.png (44.51 KiB) Viewed 13327 times

ex4.png (67.78 KiB) Viewed 13327 times

ex5.png (32.39 KiB) Viewed 13327 times

How about defining a standard here and now?
The 1st rule would be that key signatures (regular and smaller ones) should start on the same vertical line.
Now regarding where smaller ones should end, I did not find any arithmetical logic. Sometimes -- for optical reason probably -- the space between the regular clef and key signature is close to the space between the smaller key signature and time signature:

ex1.png (33.32 KiB) Viewed 13327 times

Another idea is to align the last accidental glyphs:

tweak2.png (21.69 KiB) Viewed 13324 times

But that's only a start...

[Florian]

Engravers strike a compromise so the key signatures do not differ too much in total width, as in the first example, but the accidentals themselves are not too widely spaced as in the second. Anyone know how to do this in Finale? And does Dorico do this by default?

I don’t think it’s possible in Finale. Then again, my tweaking days in Finale are pretty much over...

And no, Dorico does not do anything of the sort. I suppose they would be willing to add it, but as with most engraving subtleties that are not implemented yet, it will probably have to wait until the feature set is much closer to being complete. (By the way, SMuFL contains glyph variants for smaller sizes, and I imagine (hope) those variants will be used for smaller staves at some point, too.)

[benwiggy]

Dorico just scales down the staff, increasing the 'blank space' between the key sig and the time sig.

[John Ruggero]

Schneider, your rule Rule 1 regarding the vertical starting point seems correct both logically, visually and from what I have seen in practice.

Rule 2 might concern the spacing of the accidentals in the smaller key signature. The standard spacing according to Ross, Gould and the Finale default is 0 spaces. This might be expanded to about .075 spaces for the smaller key signature. Expansion beyond that limit seems a little large to me. Others may feel differently about that.

i don't think that the presence or absence of the time signature should have any influence over the spacing of the smaller key signature.

There is also the case when the key signatures don't match, as with transposed scores, but I think that the above rules would still apply as necessary.

[Schneider]

[...] i don't think that the presence or absence of the time signature should have any influence over the spacing of the smaller key signature. [...]

Ok.

[...] Rule 2 might concern the spacing of the accidentals in the smaller key signature. The standard spacing according to Ross, Gould and the Finale default is 0 spaces. This might be expanded to about .075 spaces for the smaller key signature. Expansion beyond that limit seems a little large to me. Others may feel differently about that. [...]

As a space unit, LilyPond uses the staff space height. So 7.5 % of 1 sp seems really small to me.
Here's another proposal, using the "inbetween" padding Malte mentioned and which seems to be similar to the ones observed in chamber music scores found on IMSLP:

tweak3.png (36.06 KiB) Viewed 13276 times

The other intersting thing is that this padding is arithmetically easy to calculate.
And here's the result in situ :

ex6.png (69.82 KiB) Viewed 13256 times

[John Ruggero]

Your result looks just fine to me, Schneider.

The "space" unit I referred to is the same as yours: the distance between each of the staff lines.

For me, it might be easier to know what the actual spacing would be for the smaller key signature. My reasoning is that one would want to spread the smaller key signature as much as possible to fill up the space before the time signature yet remain within acceptable norms. Thus it would be nice to know what this maximum spacing could be.

In the first of the following examples, the spacing is set at .08 space which looks OK to me. It might even go up a little. The second example is set at .2 space, which for me is too large. I too was surprised at how small the acceptable range is.

Of course, I couldn't keep the larger key signatures at the normal 0 space, since Finale doesn't seem to have that capability.

Key Signature Spacing at .08 space.jpeg (43.61 KiB) Viewed 13251 times

Key Signature spacing at .2 space.jpeg (49.04 KiB) Viewed 13251 times

[Schneider]

[...] In the first of the following examples, the spacing is set at .08 space which looks OK to me. It might even go up a little. The second example is set at .2 space, which for me is too large. I too was surprised at how small the acceptable range is.

Of course, I couldn't keep the larger key signatures at the normal 0 space, since Finale doesn't seem to have that capability.

Too bad, because it's diffucult here to evaluate the right space by eyes.
According to my calculation, the optimal padding follows a simple function: y = 2/3(1 - x) , where x is the small staff scale factor.
For example, in my in situ score, the small staff is set to 5/7 so 71%, so the padding should be 0.19...
Pretty close to .2 space!