Triple thin barlines

A question of barlines. For technical, musicological reasons I need to be able to use triple thin barlines, like the currently available double one (U+E031), but with the extra third line equidistant. (I won’t bore readers with the reasons, but it’s to do with transcriptions of medieval notation.) It’s clearly not feasible to request this as a feature, since its usage would be absurdly limited, so I need to fake it somehow. My immediate thought was to use font-editing software (I use FontLab Studio) to edit one of the barline glyphs in Bravura that I am unlikely to use so often. But am I right in thinking that, according to the SMuFL specification, Dorico actually uses primitives (i.e. I presume simple graphics) to produce barlines, because of the complex drawing needs, and that the SMuFL glyphs are really only intended for use in text-based applications? Even if this is the case, I suppose I could try hiding the normal barline and replace it with a text glyph of my newly created triple barline – though this would obviously have to be done at a late stage in the editing process, as the new barline wouldn’t behave like a normal one. Does anyone have any other clever ideas? (By the way, I need to be able to retain accurate bar numberings, so hypothetical solutions involving an ultra-thin bar that juxtaposes a double and single barline would be out of the question, even if it worked, which it probably wouldn’t.) My previous favourite notation software for this particular kind of work actually incorporated the triple barline as a special request, but if I transfer this to Dorico now, I’m going to need a hack of some kind.

Professor Ron Woodley
Research Department
Royal Birmingham Conservatoire

Unfortunately Dorico does not draw barlines using font characters, but indeed using primitive lines.

If you tell me a bit more about your needs for a triple barline, perhaps we can include it in Dorico as well.

Dear Daniel,

This answers a question that occurred to me after seeing that gorgeous Petaluma font (thank you, Anthony!). And it gives me two more:
Have you – as a team – thought of implementing various different staff types or even a staff editor in Dorico that can mimick felt pen, ball pen, pencils and make staffs appear in a slightly asymmetric and/or oblique way? Kind of Harry Potter staves for anthroposophists? Never square?
It would indeed mean to implement Illustrator/Photoshop qualities into Dorico…
And, speaking of haunting saxophones, is MNX intended to include information about staves/staff appearance, too (I don’t think there was talk about it at the March meeting)?

And since I’m at it: Having read that Anthony used Photoshop, Illustrator and Glyphs (and I guess you advised him to proceed in that way) to create the Petaluma font… why Photoshop in this “vector-ious” environment?
And, if I may ask, what made you choose Glyphs?

Many, many thanks for that update! It’s holidays in Germany and I could afford to spend some time to get acquainted with the software. The update arrived just in time when I’d thought ok – got it. I hope you had a day off as a team to sip some champagne. Or hang around with some inspiring, organic Californian chicken!

Cheers and hello to the entire team!


No, honestly we’ve never given any consideration to drawing staves with non-rectilinear paths. I would never say never, but I think there are hundreds or possibly thousands of other things that will deliver more value that we can and should work on first.

Ant used Photoshop because he was drawing the characters on his graphics tablet in Photoshop and using layers to put each separate pen stroke into its own layer so that it would become its own path in Illustrator for tracing; using a single path doesn’t work properly because the tracing then “fills in” the points where the strokes cross, which isn’t how a pen works in the real world. Ant uses Glyphs because it’s the font editor he’s used since the beginning of development to produce the user interface font. I still use FontLab because I’m comfortable with it after many years of using it.

I couldn’t agree more, seeing how much work still has to be done (and you see much more) – it was just a thought I wanted to express.

As regards font creation… I don’t have any typesetting software, but am confused that it’s still necessary to use vector graphics software??
Doesn’t typesetting software include all the necessary vector drawing tools? It looks like a very cumbersome approach to me, having to use three different tools.
Sorry for keeping you busy, feel free to direct me somewhere else where you think I might retrieve useful answers to this.
In fact, this is a topic in its own right.

Hi Eddo, and thanks for the interest in Petaluma,

As Daniel mentioned, I started by drawing each of the glyphs in Petaluma using a graphics tablet to get the most natural response from Photoshop’s brush tools. I modified one of Kyle T Webster’s brushes (that now ship with Photoshop CC) to best emulate the style I was aiming for, and then used that same brush (at just two saved point sizes) to ensure consistency throughout the font. I’m fairly comfortable using Photoshop and was easily able to bring in musical examples to use as a reference.

I then had the task of vectorising each pixel-based image and Illustrator’s live trace tool gave me what I needed here, converting the hand-drawn pixels into paths that would form the glyphs of the font. I needed to draw any glyphs with intersecting lines using multiple layers in Photoshop in order to retain crisp edges when tracing to paths, however once in the flow, this was easy enough to accommodate. Moving between the three applications didn’t pose any problem. I honestly think it would have been far more cumbersome to have to create the paths from scratch in Glyphs and still manage to emulate the look of something hand-written.

I’m glad you like the font, thank you.


What, no METAFONT Anthony? (Just kidding.) But on a serious note, I remember reading many of the early posts on Daniel’s Making Notes blog and being very strongly reminded of Donald Knuth’s collection of papers, Digital Typography. I mean that as the highest form of praise.

Oh, so Petaluma is yours? Congratulations, Anthony?

You’re rocking Glyphs? Pro or Mini? Did any of you port any of the scripts written by Ben for use in FontForge?

Thank you!

I use Glyphs Pro, and Ben was kind enough to port the scripts for me. He’s good like that.

Thank you, Anthony!

May I kindly ask who Ben is (is he working for Steinberg, too, or a free lance that works for the world…?) and if the scripts you are talking about refer to the SMuFL metadata?

Daniel: I posted quite a lengthy reply to my OP, giving some more details, but it doesn’t seem to have reached the forum – has it got lost in the ether somehow? If so, I’ll try again…

Ron Woodley

It doesn’t appear to have made it, Ron! Sorry to cause you to duplicate your work.

Eddo, Ben is Ben Timms, who is the head of the Dorico team, and yes, the scripts are for doing things like generating SMuFL metadata, and transforming Petaluma into Petaluma Text by performing the appropriate scaling operations, etc.

Thank you, Daniel!

I’m very much interested in having my own recognizable font, hence my penetrating questions.
I wonder if it might be possible to turn to you once I have a first result, to get it transformed into a text font.
But I’ll ask again in due time.

Any chance you guys would be kind enough to share those in the Steinberg github as well?

Daniel: OK, thanks. Since you asked for a bit more detail on my OP, here goes:

One of my main research areas is late medieval and early Renaissance notation and its underlying theory, esp. relating to the writings of the 15th-c. theorist Johannes Tinctoris. You may remember that we had a brief email exchange on this a year or so ago, when I was talking about my own font development for late medieval mensural notation.

When I am editing transcriptions of some of this repertory from its original 15th-c. notation into modern CPN, it is often very useful to be able to articulate structural units larger than the single bar. In advanced mensural notation (unlike later purely binary divisions of note value), the breve units, say, can be grouped into either twos or threes, depending on whether its next larger note value, the longa, is imperfect or perfect. In turn, the longas can be grouped into twos or threes, depending in whether its own next larger value, the maxima, is imperfect or perfect. So if (as commonly happens) the modern transcription uses a reduction ratio whereby the original breve value is represented by a bar, the particular structural, metrical permutations of these higher levels of mensuration (called minor and major modus in the theory) need to be articulated in the edition by groups of bars, and then groups of groups of bars. These could in principle be represented in a number of ways, such as some kind of indication above the staff at the point of the relevant normal barline. But in the past, when I have been doing this kind of work, I have found it useful to adopt the double thin barline as a simple representation of the first supra-bar level, and the triple thin barline to show the highest level of mensural structure (groups of groups of bars).

I have basically transferred nearly all my editing work to Dorico now, and I really think it is a terrific application, and destined to become even more so. But I have been forced to hold back for my medieval stuff, partly because of the missing triple barline. I have previously used NoteAbility Pro for this medieval editing, since its developer Keith Hamel at the Univ. of British Columbia very kindly added the feature for me. I know there are various ways I could fake it in Dorico, for example by adding the extra third stroke to relevant double barlines at the end of the editing process by exporting to e.g. Illustrator – and indeed I already have to do this kind of thing for other drawn features such as ligature indications, which need to be stretchable and tiltable horizontal square brackets. But it would certainly be a great boon to have the triple barline implemented natively in Dorico. (The other drawing features, many of us are hoping, may eventually make an appearance before too long, anyway – cross fingers!) I have no idea whether other editors, composers, etc. might have some other use for such a barline – I suppose it’s quite possible – but the above is what prompted my own query, anyhow.

Sorry for the long post, but you did ask for more detail!

All best,

Ron Woodley
Professor of Music
Royal Birmingham Conservatoire

Ligature brackets can be faked inside Dorico with a faux tuplet: select the notes from the ligature, press ; (semicolon), type “2:2” (for a two-note ligature), press return, and adjust the properties of the resulting tuplet so that the number is hidden and the bracket is always shown above. Creating a tuplet with a 2:2 ratio (or any n:n ratio) will not affect the rhythm at all, but nicely behave as a tuplet anyway. The bracket will be stretched/slanted accordingly.
BTW, I think the triple barline could be a nice innovation.
[edit: ; is a semicolon, not colon…]

Thanks for the detail, Ron. I will do what I can! We might not add them to the UI but perhaps we can add them as a little secret thing you can enter into the Shift+B popover…

Thanks very much, Daniel – it would be excellent even to just have this as a popover option.

And thanks, too, Peter B. for the tip about using tuplet brackets. I’ve fiddled with these a bit now and they will work pretty well, I think, especially with the property options to use Scale to increase the weight of the lines a little, combined with the adjustment of the hook length. It doesn’t quite have the flexibility, I think, for me to indicate coloration brackets (i.e. just small corner brackets without the connecting line), but that’s another story…

Ron Woodley

Even if the triple-barline is only a “secret option,” it would be nice if the popover code were listed (as a kind of Easter egg) in the popover list, if this option becomes available.

FYI I’ve had the time to add the triple barline to the Shift+B popover today, so this will be in the next update when it comes.