New Emmentaler accidentals in 3.6 alpha are too sharp-looking

• Nov 18, 2020 - 16:26

... especially the flat accidental, with too much contrast in stroke width. The bowl is so thick whereas the vertical stroke is as thin as the note stem, and the counter is so tending left. Yet in most printed scores in my experience, the vertical stroke of the flat accidental is thicker than the note stem. The same can be said of the natural accidental.

If you check how roundish the flag of the 8th note is, that's the overall style of Emmentaler. I don't quite find these new acccidentals fit because they look "sharp". Also take a look at Leland and Bravura, neither has accidentals so spiky despite that they are relatively angular fonts in general.

These are links to that change. The second includes some figures of direct comparison.

Let's see if anybody else has the same complaint, and if there's chance to get it addressed.


At first glance, the results are looking nice over here. Alternatively though, to keep "most of the people happy most of the time", the alterations could potentially be provided as an extra font choice while keeping the "original" untouched. Like, maybe "Emmentaler (Parnassus Accidentals)" could be an actual font choice?

In reply to by worldwideweary

For clarity, I'm never against an "extra font choice", but I find a modular setup is much cleaner than bundling everything into one single installation. Do you have any professional theory or experience in software design to back that my argument is utterly "interesting"? Tbh this is not the first time I receive it when I speak of modular design and I don't understand what's going on. Can you elaborate?

In reply to by Flora Canou

Hmm, yes, allowing Musical (Text) Fonts as an extension (like MDL and the HQ soundfont, via Help > Resource Manager) would be nice to have some day. Such a font would need to come with a setting as to what the fallback font should be, like here "Parnassus Accidentals" fallback to "Emmentaler", if the fallback is not "Bravura" (which is the last resort in any case, being the only 100% complete SMuFL font).
Problem with that though is the compatibility of scores between MuseScore versions, even between the same version with and without such an additional font installed.

In reply to by worldwideweary

IMHO it's be way too much overhead and duplication to keep the old "Emmentaler" and add a new "Emmentaler (Parnassus Accidentals)", for just those 2 (flat/sharp) - 5 (flat, sharp, natural, double flat, double sharp) accidentals.

Hmm, we could have an "Emmentaler (Parnassus Accidentals)" with just those changed glyphs and a fallback to Emmentaler for the rest (and that then fallback to Bravura, as we have already.
Still to much of an overhead IMHO though.


In reply to by Jojo-Schmitz

(Hm, didn’t get notified here, only in …)

It’s a bit more, about 1½ dozen accidentals are affected, but the new MScore font also contains various other corrections which you really want to have.

On the other hand (as I said in the original post), I made the new accidentals have almost the same advance width as the old ones: except for the double flat 𝄫, which sadly is a tiny bit wider, they are all the same size of the old ones or even a little bit thinner, and as such don’t affect the layout of scores (the difference is in the area of 0.02 sp). This was a thing I specifically looked out for.

The new accidentals also bother me, but it could be because I've become accustomed to the Emmentaler ones, as I don't have read a lot of engraved music.

The bowl is supposed to be this thick. I scaled the accidentals already to match the Emmentaler-derived glyphs, but if you look at decade-old prints from e.g. Breitkopf & Härtel or Bärenreiter, you see that:

b_und_h.png bärenreiter.png

Basically, the ♭ was too thin in Emmentaler already (and the ♮ looked as if a kid who barely knew notation had drawn it). The new accidentals really fit a more professional look, and I worked hard to integrate them with the overall look.

Since you prefer a ♭ with a less pronounced bowl, and taste is subjective, why don’t you have a look at the musical notation font Leland, which is currently being designed specifically for MuseScore (check Tantacrul’s Twitter for progress) and will be the default in MuseScore 4? It contains a slimmer-bowled ♭.

Fun you should mention the stroke width. The new Parnassus-based accidentals have a wider stroke width for the vertical strokes of the flat*, natural and sharp signs than the older Emmentaler-based ones, which were quite a bit thinner than the note stem (which, incidentally, is not part of the font but expressed in a spatium unit in its metadata). I had to widen the stroke on these glyphs a little, but specifically the vertical strokes are more Emmentaler than Parnassus in the new MScore accidentals.

*) the vertical stroke on the flat signs is thinning towards the bottom instead of a straight line, but that’s not a problem, it’s by design


You could be running into screen resolution issues. But if you zoom in or print it out, it’s right.

In reply to by mirabilos

If I understand correctly, you state the (average) width of the flat's vertical stroke of new Emmentaler is wider than that of old Emmentaler, and that of old Emmentaler is thinner than the note stem. So this:

old Emmentaler < note stem < new Emmentaler

In fact, your figures ( showed this:

new Emmentaler ~= note stem < old Emmentaler.

It can't be a resolution issue since it's the same for not only the downloaded alpha version but also all your figures.

In reply to by Flora Canou

It’s new MScore. Emmentaler is the musical font that comes from GNU LilyPond. MScore has long left that.

Anyway, no, not the average width, but I enhanced some widths throughout. Yes, they look sharper — mostly because they’re not boring uniform straight strokes any more. But they do look more professional.

I'd just like to chime in with some thoughts. The parnassus font was not made to mix and match with other fonts. It has a style on it's own, based on old engraved scores, like edition peeters, etc... The lilypond fonts in contrast are more styled like henle verlag or barenreiter. Both are quality designs, but they are not really made to mix.
I know parnassus is still work in progress, sadly I don't have much time to work on it...
I find that a music font should also be used together with a style for other elements, like beam-thickness, stem-thickness, note spacing, etc... The default musescore style doesn't blend well with parnassus. IMO it's a mistake to swap fonts without also changing other elements.

In reply to by kuribas

Perhaps, but the accidentals work much better, and those from Emmentaler (especially the ♮) look like an insult (and I really wanted the new ♭ as well).

I did adjust those glyphs I took from Parnassus to fit the rest of MScore better. Took me several person-days.

In reply to by kuribas

By the way, I recently read… I think Daniel Spreadbury… talking about early vs. late musical notation fonts, and how the issues I was seeing especially with the ♭ come from the low-DPI computer displays early fonts were optimised for (and printers back then also had rather low DPI, 72 or 144 not being uncommon). Hinting also was in its infancy. For these reasons, early fonts tended to use strokes of a very uniform width, not diverging too much, and certainly not filling the belly of the ♭ so much.

The thicker-belly ♭ of the current MScore indeed looks a little too “oomph” on my 92dpi laptop display, but it’s much closer to either Peters or Henle or Bärenreiter on paper, and definitely an improvement compared to professionally metal-engraved scores.

And the ♮ in Emmentaler really was an insult to a musician’s eyes.

I don't think I like the accidentals either.

Incidentally, things like lyrics and stave text look too 'ye olde' for the kind of music I transcribe (pop music). It might look okay on certain genres (classical), but it would be nice to have the option.

In reply to by chen lung

There are other options for nōn-classical use. Try MuseJazz or something.

And for text, you can use any font installed in the system (I don’t use a default font either, I have my own “house style”). It’s just musical fonts for which things are extremely tricky.

I actually like the new accidentals. However, I have not been working with emmentaler a lot recently, and that's why I may not have seen a drastic difference.
However, I would be curious to test with wider vertical lines, and would gladly point you to a musescore build with them if you provide an updated font.

There is a way that I think is most perfect. "Old Emmentaler" should be kept as a separate font and it should inherit all previous 3.5 features. "New Emmentaler" should be as it is now. BTW, I find that the "d" symbol in something like "d=80" has been severely enlarged in the new Emmentaler.

In reply to by Zongshu

It would be interesting to see a survey on this specifically. Changes are made often, we can't refuse to make improvements just because someone might subjectively think the change was a step backwards. Otherwise we'd all still doing all of our computer work on punch cards.

In reply to by Jojo-Schmitz

I think our problems all started with the wheel. If dragging heavy objects on the ground was good enough for my great-geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-grandparents, it should have been good enough for my geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-great-geat-great-great-grandparents too.

In reply to by Marc Sabatella

Punch cards -> computers was a technological improvement.
You can't say that the new accidentals are somehow technogically more advanced than the old ones, can you?
it's not like reverting the accidentals would create 1000 bugs in the software, or preventing people from writing scores properly, right?
If you like the new accidentals so much, duh just create a new font. Old fonts are not meant to be changed.

In reply to by Zongshu

Improvements are improvements, whether technological or aesthetic. But they are also subjective, and of course there will always be some minority who subjectively prefer the old even if in general most prefer the new. The fact that some minority might prefer the old has never stopped people from making aesthetic improvements to designs. Including to fonts. Yes, they are improved for aesthetic reasons all the time. There is nothing even remotely unusual about making changes to the appearance of certain glyphs from one version of a font to the next.

So again, it's unfortunate that you are in the set of users that doesn't like this particular change, but please do understand that these sort of changes are perfectly normal. It is impossible for each and every aesthetic change to please each and every one of our millions of users.

In reply to by Marc Sabatella

> of course there will always be some minority who subjectively prefer the old even if in general most prefer the new
> It would be interesting to see a survey on this specifically.

You have never carried out a survey, how do you even assert this is an improvement at all?

In reply to by Flora Canou

We seldom carry out surveys for each and every change we make to fonts; instead it's based on input from interested and motivated experts who volunteer to get involved - this is open source software, after all. That's how virtually every single improvement made to MuseScore has happened over the years.

It is certainly conceivable that for the first time in history of MuseScore, a survey could be constructed specifically to get input on a specific feature. But again, being an open source project, this effort too I'd expect to come from interested and motivated experts who volunteer to get involved.

In reply to by Flora Canou

Just the same thing I can say about every single change that has ever been introduced into MuseScore: the people who volunteer their time producing the software had their reasons for considering it an improvement, or they wouldn't have taken the time out of their day/week/month to make it happen. And the change was accepted because the people advocating for the change did a sufficiently good job of convincing the people in charge of such things that it was an improvement to merge the change. That doesn't mean I personally remember every single reason for every single change that has ever happened. In this particular case, I recall discussions among multiple users, developers, and full-time professional designers and engravers regarding flaws in the previous version of Emmentaler. It didn't seem unusually controversial at the time.

In any case, it is of course impossible to "prove" any aesthetic change is an improvement, because it's completely subjective. All one can do is point out that motivated experts thought it was, enough so to take the time to make the change. like any of thousands upon thousands upon thousands of other changes that were seen subjectively as improvements over the years but maybe two or three people out of the millions of users didn't like.

Although without interest for most of this conversation, I give my own personal opinion on the side:

I think it would be neat to be able to have -- in addition to choosing separate accidentals by means of a replacement mechanism for loading external glyphs -- a version of "Emmentaler" that made use of the original but smaller scaled version (as mentioned here: there are something like 7 versions), but resized to be 'default' sizing.. They'd be bolder, and I think I'd be interested in testing them out that way. I saw somewhere on lilypond's website a demonstration of this and thought it looked nice. Even the note heads were a little different in scaling, so it'd be like a whole different font. I'd imagine most people wouldn't be interested in it though.

In reply to by worldwideweary

I’d love to obtain that flexibility!

However, as long as loading musical notation truetype/opentype fonts still involves code changes, I’m afraid every new option is really expensive, plus it will differ across versions.

Perhaps mu͒4 will solve this… in theory, for a SMuFL font, you only need one additional JSON file and no code changes (if everything is implemented “correctly” on both sides, renderer and font)… I wonder how the current MScore will fare in Dorico or other notation software…

Do you still have an unanswered question? Please log in first to post your question.