[WIP] My Sagittal Font

[Fluffeh]

Hello everyone! :D

I'm an amateur font designer currently working on developing a SMuFL-compliant music font. After looking through the basics of Sagittal notation I've decided to try my hand at designing and/or developing the Sagittal glyphs. Since this is the official Sagittal forum I figured there'd be no better place for some feedback.

Here I've started with a few designs of (tai):



The first one is based off of the standard glyph included in Bravura. The other two incorporate varying levels of a slant found in William Lynch's calligraphic depictions of Sagittal notation.

For comparison here's a few more-standard accidentals:



More to come but first it'd be great to hear any thoughts/comments/suggestions/feedback anyone has! :D

[Dave Keenan]

Hi Fluffeh. Welcome.

One problem with Sagittal glyphs is the need to make four shapes which all have to be some kind of arrow-head or directional "flag" and yet have to be very distinct from each other. Another problem is the propensity of many humans to confuse left-right reversals of the same shape. So in a sense we need to distinguish 8 different shapes of flag. The four shapes are conceptually, in order from simple to complex: straight, convex, concave, and concavoconvex (wavy). The most common alterations use the simplest shapes. These four conceptual shapes leave room for graphical interpretation. And in the existing font we attempt to distinguish the left version of a given shape from the right version by varying the width in proportion to their untempered size in cents.

In fact we attempt to make the area of the flag including the area enclosed between the flag and the shaft (and an imaginary horizontal line at the bottom of the flag), roughly proportional to the size in cents, across all shapes, not only between the left and right varieties of each shape. If not proportional, at least in the same size order. A visual sense of apparent size being more important than the literal geometric area.

There was suggestion at one stage to enhance the difference between left and right varieties by adding a serif on the end of the least common flag of each pair, but I find it aesthetically problematic to mix serif and san-serif, and the conventional accidentals are clearly san-serif.

Now to your specific examples for tai. We had similarly "round-shouldered" convex flags in an earlier version of the sagittal glyph designs, but rejected them as being too easily confused with straight flags under poor seeing conditions or when reading at high speed. I note that William's calligraphic examples show enormous variation in the angle with which a convex flag departs from the shaft (including in symbols where they are combined with other flags). But the interesting thing that makes William's convex flags very distinct from straight flags is the curving back towards the shaft, and in some cases curling upward.

However we then have to be careful that the downward version (tao) remains distinct from the conventional flat symbol under poor seeing conditions. And I note that you need to test every symbol on both a line and a space of the staff, to ensure that the features you are relying on for distinctness are not lost against a line in either position.

A good way to test for distinctness is to view the glyphs side-by-side in a very small point size. And if you could apply a blur it would be even better. A discipline that George and I used (in lieu of a blur) was to first design a bitmapped version of the symbols, on a 6-pixel staff. i.e. there were only 5 white pixels vertically between staff lines -- the staff lines themselves being 1 pixel high. We didn't know at the time, that these bitmapped symbols would actually find a use one day -- as the Sagittal "smilies" used in this forum.

Perhaps we should have forced ourselves to do it with only a 5-pixel staff, but in fact, using small bitmaps is a very artificial and particularly rectlinear kind of "blur".

So I think the lesson to take away from William's efforts is not the occasional "round shoulders" (which represent a straightening of the flag itself), but the increased convexity provided by not allowing them to straighten out and go vertical as in the existing font, but continuing to curve in the same direction right up to their end, while still remaining distinct from the conventional flat symbol.

[Dave Keenan]

Another possible source of inspiration for distinction between sagittal glyphs is the heraldic description of the symbols based on the Sagittal mythology (much of which has not yet been written). You can see these descriptions in the rightmost column of this document.
http://sagittal.org/Sagittal2_character_map.pdf
I note that the codepoints in this document are now out of date, having been replaced by the SMuFL codepoints.

[Fluffeh]

Thank you Dave for the welcome and such detailed insightful commentary!

I find it quite pleasant and very interesting to hear that the creative licence for the various accidental shapes is so liberal. Such generous freedom certainly will prove a valuable asset to all of us working on the development of Sagittal.

For clarity, by "straight flags" do you refer to the former or the latter?



For the former I did indeed notice the potential for confusion, as well as with (tao) and the conventional (flat) symbol. I did take note of the fact that William's (tai) and (tao) continued to curve rather than end in a vertical stroke. I however, omitted that for the same reason. As the standard quaver flag curves in, I thought that there was actually more potential for confusion if the convex accidental flag ended with a curve rather than a straight line.

I was slightly cautious to attempt a more generous curve, as found in William's more scroll-like (gai), (gao), (wai), (wao), (dai), and (dao), hence the somewhat conservative design, but now knowing the extent of my creative licence I will certainly experiment with some more innovative designs! :D (Though I will note that I think that I think William's scrolls aren't ideal, as I think that increasing the concavity that much could lead to a confusion with a conventional (flat), as it essentially curves back into a full revolution the same way.)

I noted that in the heraldic descriptions for the accidentals, left/right counterparts actually had different origins rather than simply being mirrors of each other e.g. (tai) originates from Archytas's arc, whilst jai originates from Demeter's' scythe. I agree that a combination of serif and sans-serif is not aesthetically pleasing, but would I be correct to therefore derive that each left/right counterpart needn't necessarily be mirrors, but in fact can have quite substantial variety as long as the general concavity (convex in this case) remains the same?

I'll follow up with a post depicting the results of a comprehensive distinctiveness test you recommend soon!

[Dave Keenan]

Thanks Fluffeh. I agree with most of what you say. To answer your questions:
While it's true that those parts of quavers are called flags (I'd call those wavy and straight from left to right), I was using the term to refer to the half-arrowheads on the Sagittal symbols. George and I adopted the term "flag" for these because it's so much shorter than "half-arrowhead". So a Sagittal symbol consists of 1 to 3 flags on 1 to 4 shafts with 0 to 3 diacritics (or accents). So we have:
Straight flag or barb
Convex flag or arc
Concave flag or scroll
Wavy flag or boathook

Unlike the straight flag on the quaver, we didn't make the straight flag on a Sagittal symbol horizontal, because we wanted it to look more arrow-like when combined with a shaft.

You will see above, that we have used the term scroll as a shorthand for concave flag, but you are quite right that some of William's convex flags could also be called scrolls although they curve toward the shaft instead of away from it. So perhaps we should stop using "scroll" to mean "concave flag".

You are right that that each left/right counterpart needn't necessarily be mirrors, but in fact can have quite substantial variety as long as the general concavity/convexity remains the same.

[Dave Keenan]

Have you seen Thomas Richter's Sagittals in http://www.ekmelic-music.org/en/extra/ekmelily.htm?
He only has a small subset and so does not include any concave or wavy flags, but I note that his "straight" flags are actually slightly concave. This makes them more distinct from the convex flags and suits the somewhat hand-written style of his font. If he were to later add some symbols with concave flags they would only need to be obviously more concave than his "straight" flags.

[Fluffeh]

Very interesting! I had read about Ekmelily but I had not taken a close enough look to realise that slight intricacy. It's a pity that Thomas's glyphs aren't designed to be viewable in a conventional typeface program; I do enjoy going through various fonts to study the composition of their glyphs.