Dark stencil

Adam Jagosz

This design is derived from handlettering on a T-shirt I did recently.


The objective was to keep the letters fat so that the line design could be drawn inside. I made some trials with the ornaments (a-z and some punctuation) and it looks promising, though the shape of the lines evolved in another direction. I tried to make the ornaments to mimic the shape of the letters and make up for the legibility lost because of the stencil effect.



For now though, I'd like to ask for critique of the solid version.

https://onedrive.live.com/?authkey=!AKJiwOswpLaTQcc&cid=F7DFB60722D34AE9&id=F7DFB60722D34AE9!241427&parId=F7DFB60722D34AE9!240950&o=OneUp

Ray Larabie

While non-stencil display type can be read with little or no space between letters, with stencils, the reader has to differentiate between struts and spacing. While I can see there's a difference in space/strut width, I think it's not quite enough. It'll look dorky if you space it too wide but I think the current sidebearings put it a wee bit past the frustration threshold.

But I think the tight spacing looks fine on the shirt.

Jasper de Waard

Really nice work! I'd keep the spacing as it is, actually. The t (and f) looks a bit feeble. Maybe make it just a little bit wider.

Alex Kaczun

Nice work Adam—I think you have a winner here! Because of the fluid shapes, I'd round out ever sharp corner (on everything)—especially the small serif aspects. Make the dot on the lowercase "a" smaller—like the period. The spacing is ok for display, maybe just a little more open for smaller text sizes. Lowercase "f" ok—maybe the "t" more like this. Keep up the good work & good luck!

Alex Kaczun

I'm back again—you can even introduce the "stencil-like" flow to the lowercase "t" like this.... better I think. Have fun and experiment!

Adam Jagosz

That's a good point, Ray. I guess I could try and tweak the sidebearings to make the type more useful for short paragraphs. It would also help with problematic kerning pairs — the way it is now, most of the air is sucked out and there's not much I can do with /z, /x, and /r. I admit I am rather fond of the tightness, though, and I think the target is mainly single words or short phrases.

Thanks, Jasper. For /f I made alternates with prolonged crosstrokes that appear in front of round letters and /a. Those didn't show up in the pdf because I forgot to turn on calt. While at it, I thought maybe /t deserves such treatment as well. Both /f and /t have the crosstrokes really short because the serifs on most letters (like /i  or /m) exceed the sidebearings by 2 units. This all might be a weird set of choices. Maybe I should shorten the serifs instead of changing the sidebearing of the stem?

For now I changed the default variants of /t and /f to include longer crosstrokes, while the short variants, activated as contextual alternates, appear before serifed letters. Another /f variant, with an even longer crosstroke, appears in front of round letters like /o. (See the last page of the updated pdf.) I'm not sure I like the result — if the user doesn't activate calt, they'll see the t and f bump into half of the other letters, and any longer passage looks like its t's are spiking out, piercing through the eyes. I liked the petite, overly frugal versions as the default — what do you think?

Anyway, maybe the place to implement this isn't calt? Should I just throw it together with liga? At least this one is on by default in Adobe apps. Even though what I'm doing here is de-ligating the letters, quite the opposite. But maybe that's just the stencil counterpart of what traditionally is a ligature. Ligating the whitespace.

But perhaps you do not only mean the crosstroke, but the total advance width, especially the right sidebearing? Most people are used to having more space there (after t, f, and also r). But my mind is infested with textura quadrata and that's one reason I'm trying to treat /f and /t like siblings of /i in terms of advance width.

Thanks, Alex! The big dot on the /a is one of the defining features of this face if you ask me. I love the small dot on Dala Floda, a gorgeous stencil I brought up here recently, but my design (sadly) will not even try to be it I agree though, that in a polite, correct version of the font, in another reality, the /a would feature a smaller dot (as would /i — and /r, a more modest arm). About the straight vs rounded terminals — I am uncertain about that, I decided to give the stencil effect a twist by introducing an angle in the stencil gaps. What I mean is, I tried to keep the terminals not parallel to the shape they are facing (most often curve — so parallel would be tangent here). I'm not sure if that even shows up or is conducted thoroughly. The rounded terminals were applied on strokes that are more horizontal stems than just serifs — on /t, /f, /E, /F — and the straight cut ones — whenever there is a stencil gap or serif. In the case of the crazy 10 variants of /f — the ones created as components of ligatures feature a straight cut terminal as I imagined them to connect <invisibly> to the following letter. Maybe this explanation makes this madness seem just a little bit more reasonable? On second thought, the original idea features only rounded corners. That straight edge idea was probably just out of laziness...
That last /t seems curious, I feel it would somehow resonate with my cap /T — connected but indented.

Sorry to make this so long. Updated pdf:

https://1drv.ms/b/s!AulK0yIHtt_3jt5hoGL7BieYIFwd-A

Craig Eliason

I like this quite a bit.
I think the very circular tittles and top terminal of /a/ stand out a bit; a more flowing shape might be more appropriate.

Adam Jagosz

I made the terminal on /a slightly lighter and along with the tittle, periods, dieresis — more squarish. The /a terminal also now sits a bit lower. What do you think now?

https://1drv.ms/b/s!AulK0yIHtt_3jt5kf1UO_B3ktXi-5A

Ray Larabie

Those changes made a huge difference; quite easy to read. I get tripped up on e and c though.

Jasper de Waard

I still think something's off about s. Everything about this typeface is nice and bold, the t feels nitty gritty.

Christian Thalmann

Nice and cuddly!

I'm not sure the centered ogoneks work, though. It should be flush with the right side of the /a, for instance. Also, the /mu is hard to parse; the tail should rather curl to the right under the body.

Great job on the /Germandbls!

Jasper de Waard

Sorry, s was supposed to be t of course.

Adam Jagosz

Thank you everyone! I appreciate your input a lot.

I made the s narrower. I tweaked the terminals to look more like teardrops in both /s and /c. I was surprised to see the more geometric shapes work better, it somehow seemed counterintuitive before. There is still room to go in this direction with /z, this one seems more tricky, though.

I adjusted the /aogonek, this time moving in the opposite direction, away from naive geometry. It looks more polite now, but I think the previous version was more playful? (Btw I'm Polish But perhaps I shouldn't use that fact to justify bad decisions).

Thanks for your /mu suggestion, Christian! That's what I've been missing.


I reverted the default forms of /f /t to use the short strokes. With calt enabled, they are substituted in this manner:

&nbsp;&nbsp;&nbsp; lookup calt_ft {<br>&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp; sub [ @f @t ]' @lc_round_left by [ @f_calt @t_calt ];<br><br>&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp; ignore sub [ @f @t ]' @lc_serif_xheight;<br>&nbsp;&nbsp;&nbsp; &nbsp;&nbsp;&nbsp; sub [ @f @t ]' by [ @f_long @t_long ];<br>&nbsp;&nbsp;&nbsp; } calt_ft;

The .calt variants are longer than .long ones and appear before @e, @o etc. This way the long-stroked variants appear also where the context is lack of context. If the user prefers the short strokes everywhere or in a particular place, they turn off calt.


https://1drv.ms/b/s!AulK0yIHtt_3juJ6Arc_Q9Mt9wYZsA

Craig Eliason

Good changes.
I think the arm of /r/ could be narrowed a bit to help close up that space underneath.

Christian Thalmann

That's a nice /five!

Yeah, the /r is rather ungainly at the moment. The drop is too heavy. Maybe it could eat into the stem a bit so as to save horizontal space.

The ogonek could still move a bit more to the right edge in /aogonek. Maybe allow it to hollow out the foot of the stem above to make itself more prominent. My /agononek is almost always a unique glyph rather than a composite since those ogoneks need a lot of manual attention to fit into their parent glyph. They are really more like limbs of the glyph than accents.

The /mu is much better, but still looks a bit like a /u with an accent below. Maybe explore a design with a continuous left stem down to descender depth.

Adam Jagosz

I know it is rudimentary to make /mu distinctive from u, but sometimes I recall this picture at these moments:

I don't know how unreliable/outdated/amateur this is, but I've been seeing it here and there. Don't know if that's how someone might actually have been writing some time ago (probably not most of the Greeks today), or if it is just a misfortunate trial of translating English cursive into the Greek script, but I have definitely seen a Greek person writing her /mu by first writing a /u and then crossing it on the South-West.

Nevertheless, joining the parts seems to work better anyway, and since I already have joining elements in T and Q, this is just confirming the rule. The variant with the teardrop seems more at home to me, what do you think?

I slimmed the r a bit — 15 units. In addition, the top of the stem is now thinner by 5 units, and the bottom is slightly wider than normally (it was this way before, too, to a lesser extent).

 If that's not enough, I can offer this:

The third one is not quite convincing, though. And here comes the moment I ask: do the separated elements in /B /K /R /eight work? I've been getting this weird off vibe from them. How about this?

I also changed the top arm of K. Two bottom rows show the old version — I find it quite charming, though less legible.I get the impression that the stenciled B is better, though. Maybe I could provide the connected versions as stylistic alternates? Or the other way around.

Updated pdf below, screens above are from the last page.

https://1drv.ms/b/s!AulK0yIHtt_3juN5Gd56ny8BJym9bw

Craig Eliason

Definitely the mu with the thinner connection.
Thinning of the /r/ looks good (adjusting the stem too was a fine idea). No need for the more experimental fixes IMO.
I like the more separated /B/K/R/8/. Maybe keep working on the /K/ arm. The revision is more graceful but I think might carry it over too far into /R/ territory. 
There's nothing very wrong with including stylistic alternates, but in this subtle of a case I'd probably say use your judgement to decide your favorite and just go with that.

Adam Jagosz

The ogoneks were redesigned again. Test strings in the new pdf (it is only 4 pages this time to keep it easy to look at).


I tweaked /D/V/W to occupy less horizontal space — they seemed too heavy. But then I thought maybe the construction of /W with the counters reaching almost to the baseline is a bit naive? (Old bottom, new top).


I worked more on the arm and leg of the /K. Perhaps the two variants are even compatible? (See Kraków on the second line).

Does the cedilla work?

Adam Jagosz

Hello again. I've been polishing up this project lately. I'd appreciate you taking a bit out of your time to skim through the enclosed pdf listing some OT features built into the font (and point out anything that seems odd to you). Thanks!

Craig Eliason

Whew, that's a lotta OT stuff!
I'm intrigued by the different unicase options. 
Personally I'd cut SS02, the M and N alts aren't nearly as appealing as the defaults. 
All the swashes that end in points feel a little out of place to me. 
Just love the dollar sign!

Adam Jagosz

Thank you @Craig Eliason for all the insights. I revamped the alt MN to give them more volume and curves and make them qualify better for the capital letter role despite their lowercase structure. I suppose the /N could be heavier still, but I'm wary not to overdo it?

The reason to include them in the first place was to provide alternates for a more friendly, gentle look. Think Lobster. I find the default MN a tad aggresive with the sharp triangular strokes. At the same time it is difficult to tweak them enough to get rid of this feature without dissolving their shape altogether. The EF were redesigned to eliminate this characteristic (triangular terminals were replaced with droplets) but for MN that won't work.

I understand you mean the shorter swashes that are essentially just exaggerated serifs? I admit it was kind of an afterthought. I consider the results independently in mixed case and all-caps settings. I think in the latter, the swashes help diversify the appearance of the text, introducing a kind of a secondary capital level (a supercapital letter). But indeed it looks weird especially on some letters like /E where it begins to look like a staircase. Would you suggest tweaking the current solution, making up something new instead, or just leaving it out?

Craig Eliason

Those alt N & M are working much better now. I still like the default better but they're worth including now for sure. 

I'm not a fan of the swashes. They're all so thick. Would there be any possibilities in swash forms that connect with a thin (or a stencil break)?

Adam Jagosz

Okay, so the swashes that bother you the most are the "Long swashes", the tilde-shaped ones? (The other two types I can distinguish are the enlarged serifs, and the prolonged descenders on AGJYRK).

I tried stencil/narrow connection swashes initially and when it didn't work out, I went on to the continuous forms.

The problem with stencil swashes is that it's hard to make them look like a part of the letter. They end up looking like a random thing just dropped in by accident. I could maybe revisit the narrow connection idea:

Attached is pdf with the current (old) swashes in mixed case context (in the manual it was mostly all caps to show off more glyphs in smaller space).