Ernestine, a wide slab

nina's picture

Dear Typophiles,

Here is the first version of the lowercase alphabet of my first typeface*, Ernestine**. Backstory is here.

* I have made a little family of bitmap fonts before, but I'm not counting this here because Curves Are Hard.
** See below.

The intended use for this is twofold, my main focus being small text (thanks to Hrant, as well as Hans Jörg Hunziker in class, for bringing this idea to mind); the other, possibly, display: The plan is to first finish this "base" cut shown above, and then deduce two additional optical sizes from it, which will likely be of more practical use than the base itself. First, an optically corrected small size cut with more weight & looser spacing; intended point size is 3 to 8*. And second, if time permits, a display cut with less weight & more detail.

* Some quick experiments with ink traps, opened counters, and the like revealed that this can be surprisingly legible at 3pt, and rather comfortable at 4.

I'd be very grateful for crits, and any opinions really. I've been *very* immersed in this over the past couple of weeks or so, so I presume I don't really see most of the issues myself. (I must say a monoline typeface really seems to be quite a beast to try to tackle, too.) Spacing btw is still quite preliminary.

** Oh, and the name. Ernestine is the working title that I kind of got stuck on. I think it matches the feel of this rather nicely, but the word looks uh, suboptimal (very long, and somewhat boring) when set in the typeface, so I'm open for different naming ideas that would also look good. :-)

So, I'm throwing her out there. Please don't hold back, and don't be gentle for the sake of it: I'm trying to learn, and I'm wanting to actually make this usable.


nina's picture

100th post to this thread, Craig! You should win a coffee mug (or a salt shaker?) (-:

Yeah, I wonder about that "a". Type design seems to be something I can only do in brief and somewhat isolated spurts – having looked at this for a few continuous hours, I can't really tell what's up with the "a". But I did narrow it quite a lot, so that may well have been too much. Will try some variations on that.

I like the angled serifs but see your point about them "breaking off"; this is also an obvious (and visible) breach of the "monoline" premise, but I must say I'm somewhat intrigued by that. Will play around with them some more!

I wonder if the "d" is too light. But I still don't trust my new printer…

speter's picture

The bowl of the a could indeed stand to be a bit wider. The d doesn't look too light to me at this point.

BTW, Moaning Dingo is a killer name for a typeface! (Quail Blip isn't bad either.)

nina's picture

Yeah – I'll definitely keep track of these for whenever I do a wild display face! ;-)

And thanks for your feed-back Steve. Looks like the "a" is getting reconstructive surgery tomorrow!

eliason's picture

('d''s weight looks fine to me, too)

nina's picture

Here's an intermediate-width "a" (I agree this looks a lot more harmonious – thanks!).


I've also pulled the inside angle of these head serifs back up a tiny bit. Better?
(Though I realize those weights of the curves to the east of the respective joins are still a bit, uh, inconsistent…)


hrant's picture

I think you had indeed gone too far with narrowing the "a" in your first post of 3/6, but the most recent one looks great. About the b/d/p/q: it seems like the "p" is missing the "movement" present in the others; the "q" is nicely pendulous (the "d" seems close, but might need a flatter top) and the "b" is nicely perky, while the "p" seems totally straight (even though it might not be mathematically).

Revisiting the "i": I think you should raise the tittle. It depends on the "t" though (as well as the cap height - concerning which I'm getting a strong vibe that your caps should be short and unusually wide).


nina's picture

Thanks for addressing the "movement" in the bowls. I was wondering if that was showing (or working, rather). I can see the "p" does it least (mathematically, a deviation from say the "o" is definitely there); and the "d" I've been struggling with most out of these – it seems to get too droopy and sad really easily. I'll work on those some more.

"I’m getting a strong vibe that your caps should be short and unusually wide"
Ha. Interesting! I was just thinking about what my caps might look like yesterday, and caught myself wondering about making them rather short (which would probably require them to be quite wide). Wouldn't hurt for German I guess. :-)

I'll hold off on moving that tittle until I know more… (so many variables in this!)

eliason's picture

I’ve also pulled the inside angle of these head serifs back up a tiny bit. Better?

I would still add weight, probably by moving that line rightward a bit. As is, the head serif looks a bit like a flag on a pole, rather than sharing the slabby solidity of the other serifs. But again, this may just be me!

speter's picture

I agree with Craig on the serifs, and with the notion of shorter caps. But don't make them too wide.

nina's picture

Well, I don't even have half of the lowercase so far, so I guess there's still time to incubate those caps. To be honest, spending very serious thought on them now feels a bit like overcharging my brain with uncertainties/variables. :-]

Craig, your point about solidity is interesting – I do see what you mean.
I wonder, does the "p" work better for you than the "m" and "n"? That join is higher, so there is more black to hold the serif together with the rest of the glyph.

speter's picture

I don't think it's so much the join as the angle of the line. (For me, that is.) I would try to echo the angle of the foot serif by bringing the apex to the right.

nina's picture

Aha! Do you mean the angle on the foot serif on the "a" and "d"? If so, that angle should actually roughly correspond to the m/n/p (although the latter is a bit steeper, at least mathematically). I'll have a closer look at it.

Also, I'm thinking the x-line serifs may be too heavy overall. Adding that to the list.

nina's picture

Here's an update!

First, revised bowls:

Second, revised x-line serifs (actually, quite a bit of mini-tweakage done to serifs in general, so see PDF linked to below).

Third, new glyphs:

"e": Might be a notch too wide?
"j": I do think that descender needs more work, but I'm not sure how to do it. I don't want to make it too wide, but I'm not sure it doesn't look cramped.
"u": I wonder if anyone notices the funny thing I did to this.
So this is what I have so far, in an ugly gif:

Overall, I'm beginning to wonder if my straight stems are heavier than the rounds.

Current PDF:

hrant's picture

bdpq: Nice!

e: What's a "notch"? In my terminology I'd say it's "slightly" too wide. :-)
BTW, are you pulling a Patria with that huge eye?

j: It just looks heavy somehow.

u: I ain't seein' nothin'.

> I’m beginning to wonder if my straight
> stems are heavier than the rounds.

Yes actually.


nina's picture

Hmm… a notch is definitely more than a hair.
Actually, it's probably a hair more than a bit.  :-)

"BTW, are you pulling a Patria…"
Huh, not consciously, though I opened Patria at least five times again tonight (but mostly related to the "j"), so it's creeping deeper into my brain. Come to think of it, most of the time I can't really tell how much of Ernestine is influenced by which other fonts. Like the "a": I only consciously realized I love Patria's sans-top-terminal "a" after I decided to make mine "weak" on top; but it's still quite likely that Patria's "a" somehow registered with me previously.
Out of curiosity, for you, is it more good, or worrying, or just weird to wonder how much of Patria is in this?

"…with that huge eye?"
Hm, I was trying to make it match the bowl of the "a".
Actually, I was mostly thinking I was pulling a (slight) Ezzo on the "e", in the way it curves out and up on the right side.

"u: I ain’t seein’ nothin’."
Good.  :-)

Gonna look at the weight on those rounds. Or the straights.

nina's picture

[Edit:] Actually, I'm quite sad that most of the features I consciously love in Patria just don't seem to work in Ernestine. Like the foot terminal on the "a" mentioned above (and the "u" and "q"), and the tail of the "g" (although I do love mine too), and the right-side angle of the beak of the "r", and the weight/serif size ratio on the leg of the "k" as well as the serif treatment on the "x".
So I guess the influence really is more subliminal; or in other details. But it's definitely there in some form.

hrant's picture

> I can’t really tell how much of Ernestine
> is influenced by which other fonts.

I think that's ideal. As long as that's "fonts", plural, for two reasons: the more numerous your influences the better your results; and you're less likely to offend a single source's author. In my case however the latter does not apply at all - to me it's mostly very flattering.

FWIW, some people say "Hey, Cheltenham!" when they see Patria's caps. The first time I heard that I was taken aback - but I shouldn't have been: although I didn't reference ol' Chelt consciously, it is in fact one of the fonts I fell in love with early on (although now I've mostly moved on).

That Ezzo "e" style is quite nice - and nothing like Patria's BTW: my "e" has a very hard nose, with a totally straight vertical segment at the tip - I got that (partly) from Jens Gehlhaar (who once produced an extreme incarnation of that structure).


nina's picture

Yes, definitely "fonts", plural. FWIW, these are the most important fonts I turn to repeatedly while making this (though some of them only via samples): Archer (though not so much anymore), Caecilia, a bit of Officina, and (lately) Oranda; a bit of Swift, and Rocky; Patria; Legato (mostly to remind me not to forget the white); Kabel; and Ezzo has been creeping in too. I probably forgot a few.
Plus, there's a definite influence from typewriter fonts, although I don't usually look at these actively. But I originally learned to read and write with a mechanical typewriter, so that stuff is about as deeply rooted in my system as it gets. (-:

Bendy's picture

>“u”: I wonder if anyone notices the funny thing I did to this.

Is it the left stem being off-vertical?

This is looking fantastic Nina!

I agree the stems look a fraction too dark for the bowls. I think j only needs a bit more work on the curve, perhaps a little bit of a wider base to open up the darkness there.

Does h look slightly too wide?

I'm looking forward to seeing the other letters. Especially s!

Marcelo Soler's picture


Gonna look at the weight on those rounds. Or the straights.
In a fleeting glimpse, it's quickly noticeable, say, in your "d", that the width of the stem and the top serif apparent weight are optically higher than those of the bowl, at least for my "senior" vision. Again, in your pleasant hard-worked "a", straights are heavier than rounds (always talking on their –being redundant– vertical width). Actually, the stress in the bowl is negative (in red) respect to the stem.

“j”: I do think that descender needs more work, but I’m not sure how to do it.
Probably it needs just a little more air to let the ball "float" more freely –I think of the nice way you solved the end of the loop of your "g", or the opposite in your "c".

“u”: I wonder if anyone notices the funny thing I did to this.
As Hrant says, I see nothing, except for the not exactly vertical left stem (Adobe's PostScript will hate you ;-) ). The only thing that seems odd a bit –depending on how much I put my eyes on "her"– is a slightly shortness in the right top serif, that mathematically is not.

There is something disturbing for me in your "e", but I'm not sure if I'm right –I'm not sure I'm not sure :-) –, perhaps some flatness in the tail (maybe I like too much your "c"):

Is the "h" wide a little?


nina's picture

Wow, lots of input! Thanks guys.

"u": Yeah, what I meant is the left side being off-vertical by like 2 units. Kind of an act of desperation – that side kept optically slanting out, until I physically slanted it in. Does it work? Oh, and Marcelo, why will PS hate me for that? Hinting stuff?
Btw, the right-side serif is slightly longer than the left, but I guess it may need more!

"j": I like the idea of "giving it a bit more air". I'll try that.

"e": Marcelo, that looks quite good! I had a variant with a less-flat bottom, but couldn't really get it to work nicely – it seemed to close up too much. I'll play around with this some more.

"h": I think, I think, I agree! But I'm stumped as to why it should be looking wide. The "n" isn't, is it?

Weights: I'll need some time to figure this out. It'd be a lot easier obviously to reduce weight on the straights than to add some on the rounds, but I don't want to make it too light overall.

"I’m looking forward to seeing the other letters. Especially s!"
Uh… me too. :->

Marcelo Soler's picture

When I saw your first entry, I thought Ernestine was a beautiful hit. Nevertheless, there was a lot of space for improvement I'd never imagined then. Definitely, don't touch the width of your straights.

Marcelo, why will PS hate me for that?
Just kidding. That's coz the way PostScript manages printing: when near to vertical or horizontal, but not exactly, the interpreter has to make a lot of decisions about what to do, that is, if inking or not; the digital world is discrete, while the analogous is not; aliansing is the gap where noise comes in. Nevermind. Programmers must work for human –not machine– efficiency. As the German aircraft designer Willy Messerschmidt used to say: "You can have any combination of features the Air Ministry desires, so long as you do not also require that the resulting airplane fly". Nevermind. It makes me remind the hateful discussion with that guy of the iPhone –he would surely uppercase the "i"– and his ego.

The “n” isn’t, is it?
No, it's not. BTW, time ago I posted an uncommented log (not in the proper place) about some concerns you introduced in my mind on that matter. Still have a lot of unanswered questions.

Ben expresses my anxiety quite well 8-D .


nina's picture

Uh, sorry Marcelo, I sure as hell didn't want to remind you of that thread!
I don't really care if PostScript is swearing at Ernestine internally :-) – I was just wondering if I was setting myself (OK, the "u") up for a lot of (printing) problems by going this route with that stem.

I remember reading your "hmn" post – left me impressed, and stumped. It's interesting to look at it again now and see some of those fonts actually have sligthly differing widths between the "h" and "n"… that's good to see.

Marcelo Soler's picture

I was just wondering if I was setting myself (OK, the “u”) up for a lot of (printing) problems by going this route with that stem.
One easy way out is converting to curves (algorithms deal better with them):

But don't pay it much attention: women always deserve lots of gentleness –Ernestine is certainly a beautiful one as well ;-)–, so don't let PS bothers you.


Bendy's picture

About the h and n, your question made me look back at the pdf rather than the image above. I think the disparity may be due to the neighbouring glyphs. In the image above the h is between g and i, which are both of course narrower, whilst the n is between m and o. The hs in the pdf seem more in tune with the other letters, so we'll have to wait for the whole alphabet in block text to see if it still looks fat.

I like Marcelo's e, looks like it would go very nicely with your lovely g.

Bendy's picture

I may be unduly worrying here but to my eye the curved descenders on j and g look shorter than the straight one on p.

Perhaps the unique way the ball terminals are floating calls for extra overshoot treatment? But, I must say, the proportions of the g and p are lovely in themselves. Maybe it's because you thinned the bottom curve on g but p has quite a dark serif down there? Or perhaps it's only me :)

nina's picture

"the curved descenders on j and g look shorter than the straight one on p."
This is interesting – thanks! I'm still experimenting with the amount of undershoot the curvy descenders should get, and it's currently less than the overshoot zones at the x-line and the baseline (which I've also been figuring out more exactly recently), so it's quite possible it's not enough.
The problem is, like you said too, both the length of the straight ones and that of the curved ones looks good to me when seen in isolation. :-|
Will play around a bit and try and make them line up better.

Also, interesting point about the "h"/"n" width "illusion". I might just hold off on actually changing the "h" until the rest of the lowercase is born.

Bendy's picture

It might be a matter of enlarging the whole glyph g, fractionally, to preserve the proportions, then getting the strokes the right weight, or the other way round with the p.

I suspect the non-baseline sitting g-bowl could be playing subtle tricks on the eye in this equation too.

What do the others think?

eliason's picture

I suspect the non-baseline sitting g-bowl could be playing subtle tricks on the eye in this equation too.
That's what I was going to say.

c - does the top curl around too far before terminating?
j - is indeed cramped looking
e - I wasn't sure at first glance (too eager?) but I've decided I really like it. The large counter works great with the flavor of the font that's emerging.

Personally I don't really like the serif-lightening you're doing - foot serif of d and head serif of p in particular look wimpy to me now, at large sizes anyway. But that one inside line we were talking about before looks great to me now.

and your specimen text is very entertaining!

nina's picture

Thanks Craig.
"c - does the top curl around too far before terminating?"
I don't think I get this. :-(
Are you wondering about raising the ball / flattening the top curve?

"foot serif of d and head serif of p in particular look wimpy to me now, at large sizes anyway."
Oh, not good. I agree that cutting off that corner makes them lighter optically, which probably isn't a good thing. Though I'm not sure they don't work okay in text. I should probably put a PDF together soon that has more text at intended text sizes… but that's still going to take a moment – it's pretty hard to think up anything resembling text with only 16 letters! ;-)

nina's picture

Off-topic, sorry – I'm just curious.
"my “e” has a very hard nose, with a totally straight vertical segment at the tip - I got that (partly) from Jens Gehlhaar"
Might there be some Avance influence in it as well? I just noticed this.

Ernestine update coming soonish too. :-)

hrant's picture

I've never looked at the outlines of Avance, and that segment is too small to see by eye. As you can check, in my "e" the segment is so tall even the inside counter has part of it.


eliason's picture

I don’t think I get this. :-(
Are you wondering about raising the ball / flattening the top curve?

I'm looking at the vertical relation of the top and bottom ends.

The ball at top stretches over the bottom - it shares this with the 'e' (it's what I meant by "eager" in that context) but IMO it doesn't work as well in the 'c'. At first I was going to suggest rolling back the top end, but now I'm thinking to retain the roundness of the whole, it might work better to extend the lower end out (and probably up). I think it can be more different from the bottom of the 'e' (which, with its crossbar, is a different context) than it is, while still maintaining a genetic resemblance.

nina's picture

Ahh, thanks for the explanation, Craig! Does it look top-heavy/overbite-y? Somehow when I moved the bottom out more, it looked strange (like it was having an underbite). I'll try again though, and maybe post some variants.

Cheers Hrant – I was just wondering. :-)

nina's picture

OK. I got immersed in the "s", and it has opened up a little can of worms. :-)
(The changes discussed earlier are still coming up; just got quite caught up in this.)

To recap, the "s" I originally intended has one ball terminal on top and no ball/serif on the bottom (like #2 below). I like the concept, but it kept looking pretty lopsided/contorted/unnatural. So just to open up the field again, I did some thinking about my options; this is what I came up with:

Options I'm definitely unhappy with in grey.
Also, I know the curves still need some work; but I'm hoping not to do the serious tweaking six times! So I hope these are worked out enough to tell if the respective concepts could work.

Some notes/comments –
1: Works, but probably looks too naked in the context of a slab serif font. Also makes me queasy when I think about the caps.
2: Still my favorite, but still a bit awkward; I'm hoping it's only a matter of execution. Would nicely match the "c" while not clashing with the "a".
3: Seems like the obvious solution, but looks too cute and too cluttered/busy to me. Also makes the weak-topped "a" look like a mistake.
4: This struck me after looking at my old typewriter again – typewriter fonts tend to have ball terminals on things like the "a", "c", "j", &c., but outwards-pointing serifs on the "s" (as well as on many caps). I'm not at all happy with it, and don't think it works for Ernestine (see the "q"), but happy I tried it.
5: Tried to combine #2 and #4 in an effort to make #2 less top-heavy. Looks like a zombie up close, and is probably too distracting in text, but was interesting to try.
6: Variation on #4, moving the serifs inside the curves. I like this better than #4, and think it works okay, but doesn't click with too much in the font (except the "z", which isn't there yet).

I'm not quite sure. I think, I think, I'd probably most like to continue on the #2 route; but I'm wondering if you guys think it's actually possible to make it look truly good and balanced (I'm finding it tough). Also, when I think ahead to the terminals on the caps, I'd probably feel most at ease with #2 or maybe #3; though I'm wondering about #6 as well, and any thoughts on the "permissibleness" of #1.

So here's a PDF with all 6 variants tested in a short paragraph at 9pt:

eliason's picture

My first impulse was to say 2 and 3 are the only legitimate options, but you may want to give redrawing #4 a chance. Move the top serif down and let it close the aperture a bit - so that the inside curve of the top counter almost hits vertical (becoming the edge of the vertical slab serif) and do the same for the bottom serif. (Don't know if that description made sense.)

I never realized how many ball-terminal fonts don't use ball terminals on 's' until you made me go look!

hrant's picture

Considering your "c" is probably final, I think you have to use a ball form, not a serif form. So either #2, #3, or... ball on the bottom and nothing on top! Assuming it doesn't look too peculiar, it would: pull away from the "c" while still being related; nicely echo the "a" on top; and nicely echo the "g" on the bottom.


nina's picture

"I never realized how many ball-terminal fonts don’t use ball terminals on ’s’"
Same here, Craig! I was pretty surprised.

Re #4 above, here's the acs treatment in
- my old typewriter (a 1958 Hermes 3000),
- Fine Line, a typewriter font from IBM,
- Alphabet Centre Georges Pompidou (1976) by Frutiger, based on Fine Line.

With the need to accommodate a serif on the outside, the curves definitely get a different character. I tried to bend them in as gently as I could, and I suspect you're right that I'd have to do it more strongly to make the shape work; but closing up counters and making curves less flat/open is precisely one thing I'm reluctant to do in Ernestine (which is also why I gave up on putting a spike on top of the "q"). It seems out of character somehow; and this treatment in particular makes a lot less practical sense to me than it seems to add noise. Though maybe I'm not seeing something.

"ball on the bottom and nothing on top"
Hrant, good point. I did try that previously, but it seemed even more lopsided than the top-balled one, so it didn't make it into this round. Looking at it again, I probably ruled it out too quickly. (Please forgive bumpiness of curves.)

I won't put balls on both ends unless there's good reason to believe I should.
But this might really be a good alternative to the top-balled one. I especially enjoy how it relates to the "c".

eliason's picture

As long as you're casting a wide net for possibilities, these asymmetrical terminals make me think you might want to play with asymmetrical bowls (if that's the right word). For example what would this last option look like with a narrower and shorter top loop?

hrant's picture

No, I think you were right to dump the form I suggested... :-/

Personally I'm starting to favor the double-ball one - not mosty because the single-top-ball one looks strange, but because I think the double one makes the "s" belong in the font more.

BTW, I wouldn't make an alternate just for the "s"; if you do that then you should have a... "ballsy" :-) option to fire up a bunch of extra ball terminals, for example on the "a".


nina's picture

You mean something like this, Craig?

Looks interesting, but maybe slightly too funky? I may have to play around with the proportions. I suspect that if the difference/asymmetry is too slight, it may be off-putting; and if I make the top even narrower than this, the whole thing seems to get too curly and round. Interesting idea though!

nina's picture

Oh, cross post.
Hrant, do you think the top-balled version works better than the bottom-balled one?

"I think the double one makes the “s” belong in the font more"
Really? Even next to the "a"?
It looks so cute. :-/

I'm not really thinking too much about alternates right now. Sure there are some things that might be nice to have as alternates, but I'm kinda first wanting to find "The One Shape I Really Want To Be In The Font And Makes The Most Sense In It" (ha).

Bendy's picture

I find 2 and 6 work very well. Version 2 is a less obvious and more unique shape, though I think the spine could move up a few units? I personally like the shape of 6 and it reads equally well in the sample text.

hrant's picture

> do you think the top-balled version works better than the bottom-balled one?

Do you mean top-only vs bottom-only?
Yes, the bottom-only is pretty lousy.

> Really? Even next to the “a”?

I dunno. Maybe it's too early to decide.


nina's picture

Yeah, I'm keeping a few versions of the "s" (tentatively, at least all of the balled ones, and the inside-serifed one) and am now working on the rest.

Ben, I think the spine on #2 has the problem that it needs to be a tiny little "too" low, otherwise the top looks cramped and the bottom too empty. But I'll try and see if it can handle another couple units.

Marcelo Soler's picture

So, what's wrong with this?

(Sorry for the bad image)
I think it's cute as a whole, indeed.
Probably I'd make some adjustments in angles on the "s", but I think it's amazingly synergic in all ways.


nina's picture

"what’s wrong with this?"
Hmm, the doubts I have are that it:
1) makes the lack of a ball terminal on the "a" (a decision I would like to consider as final) look strange, or like a mistake;
2) strangely looks too flat overall, like it doesn't have enough overshoot? (And I'm not sure I want to use larger overshoot values just for this one glyph – but if I bend the curves in more to make it less "flat", the terminals close up the counters a lot and make things even busier.)
3) looks primarily pretty and maybe a bit conceited, as opposed to serious and working; maybe also evoking too much of the usual Archer association. But maybe I'm judging that too harshly… in the end, other glyphs are a bit "special" too.
I just don't want this to go too far down the "decorative" route.

Bendy's picture

It looks twiddly to me.

nina's picture

"Twiddly"? Sorry. Does that mean something like fiddly?

Bendy's picture

Yes, exactly.
Kind of needless and ticklish.

Syndicate content Syndicate content