Satyr: No straight lines

Sindre's picture

This started as an attempt to draw type with curves only, and soon became an exercise in tension. I drew this several months ago, and had almost forgotten about it, until this post brought it back to mind. I don't really know if this experiment has identity enough to be considered a typeface in its own right. There's a companion italic in the works too, but I want to wait with that.

First specimen41.24 KB
Text setting31.99 KB
Revision A–Z, a–z41.14 KB
Pangrams (thanks, Craig!)39.04 KB
Testing25.33 KB
Italics32.87 KB
Sindre's picture

Thank you, Ben. I'm really uncertain about those weird serifs now, but I'm going to continue in this direction, as it initially felt slightly epiphanic.

I'm afraid I don't quite understand your suggestion, which glyph(s) are you talking about?

Yes, I've noticed the light 's'. I think that glyph (and its sibling, the 'g') needs to be too a little light in this typeface, but perhaps I've overdone it. What would your approach to fattening it slightly be? Vertical or horizontal weight gain?

No straight lines. That's a premise and a promise.

Bendy's picture

Sorry, I missed out the important bit! I was looking at the v, which seemed a little 'straight' next to the new style inky bits.

With the s, it seemed that the thins could just be a unit or two fatter, probably on the inside. What was your reasoning for making it light? I haven't come across that logic before!

Sindre's picture

New 'v', fatter 's', slightly tweaked 'i'. I'm still not sure if I understood your 'v' suggestion, Ben.

About light 's's: Never really thought about it, but I think it has to do with overall colour. Complicated, curvy letters (a, s, g) may look blacker compared to simpler glyphs from a distance than they do up close. Don't know if this is really true, or just a hangup of mine.

Bendy's picture

Yeah, I'm speaking gobbledygook. It's been a long day :)

I was imagining taking the idea of the fluid corner of the eye of the e and transposing it onto the counter of v. So the whitespace on the right serif has a larger radius/shallower notch...if that makes sense?

hrant's picture

Two units is plenty to see an effect.

Glyphs are not vases - they don't risk falling down! This is especially true in a text face. Furthermore, imbalance can just as well mean dynamism, which to me seems part of the spirit of this design. A cobra isn't a cobra without that huge head. And a cobra is awesome. So I would be very wary of taming something like the original "s".

BTW, as you suspect, the "s", binocular "g", and any glyph that's dense needs to look too light when viewed large for it to work well at text sizes.

> No straight lines.

Assuming you haven't already, you might like to study the thoughts and work of Evert Bloemsma, who used to say "A straight line is a dead line" (although IIRC he got that from one of his former teachers).

I myself believe that balancing straight and curved is what's optimal. For one thing, death is a part of life too.


eliason's picture

Looking great!
FWIW I don't see any "falling" that would be problematic.

The cap /R/ looks low-waisted, because the leg emerging so close to the stem makes the lower counter much smaller than the upper.

Right (thick) leg of /A/: I would trim back the serif on the left (interior) side.

I printed the text setting on my mediocre HP laser and saw these weight issues, for what it's worth (I don't know if the problem resides in the design or in my printer resolution (or hinting?)):
/b/f/t/u/ light
/i/l/n/p/ heavy

Sindre's picture

Hrant, your thoughts on glyph balance and weight distribution makes me happy. I'm a lot less experienced than you, but it seems we have drawn somewhat similar conclusions. Thank you for your cobra metaphor, I'll remember that. (I learned vector drawing from designing traditional Norwegian boats (descendents of Viking æra vessels, still built (in unbroken tradition) in the Western and Northern parts of Norway); their seemingly idiosyncratic hull shapes ("cod in the front, mackerel in the back") makes sense first when on water. I think a text face works much in the same way.)

Yes, I've studied Bloemsma, and even sketched a couple of faces based on his principles, more or less. Here are some glyphs from one of them:

I now realise that this is the true beginning of the typeface this thread is about.

My decision to use only curves in this typeface is one I made only because restrictions fuel my inspiration. Also, this is my last typeface made for purely self-educational reasons. Next time, it's serious. No such rigid rules then. (A critic once said that Richard Wagner's music is much better than it sounds. I feel the same way about Evert Bloemsma.)

Craig, thanks for printing. I've adressed the issues you brought up. No more drawing today, but I'll start refining the upper case tomorrow.

hrant's picture

> I learned vector drawing from designing traditional Norwegian boats

And I can think of two other type designer you might study: Starling Burgess and
Miles Newlyn. I know of the latter's interest in boats through his Flickr stream:

> this is my last typeface made for purely self-educational reasons.

Well, if this is non-commercial then definitely
worry about ideology much more than salability.


Sindre's picture

Upper and lower case revised, pdf attached. Absolutely no kerning. Spacing still basic. Any thoughts on the letter fit?

I've tried using the turning-contrast idea on several new glyphs, most notably 'v' and its siblings. (Is this what you meant, Ben?)

Hrant, all my typefaces are non-commercial, type drawing is just a hobby. (Though the most time-consuming I've ever had.) That may change, but so far, I haven't even once thought about salability.

eliason's picture

I like that adjustment on /v/ & kin. I wonder what a similar (but more slight) treatment would do for the caps in that family.

I like the /A/ apex.

Add a cap /U/ to your text setting - I want to see that cusp in action.

You might want to make a /Th/ ligature.

Bendy's picture

Ooh, thank you for the A!

I'm too tired to give a full crit now, but I like it very much even with tired eyes ;)

Yes, that's what I was clumsily trying to explain for the you feel it's in keeping?

I'm sure others would agree you could be missing a trick by not thinking about making some of your work commercially available.

Anyway, I'll be back when I'm a bit fresher. Good night.

hrant's picture

OK, let's do this! :-)

The "g" is nearly ideal to me. I would make the ear rise though.

It would be nice to have a more forceful "j". Is "gj" that serious a problem?

The "m" could be narrower.

The "u" is wide. Probably because it's the same width as the "n"; it has to be slightly narrower to appear the same width.

The "x" is probably too dark.

The "E" is too wide I think.

The face-punched-in "G" needs work...

The leg of the "K" is too weak.

I would make the bowl of the "P" much larger.

That "Q" is way too timid; this font needs one with a kick.

The "R" is exquisite.

I would give the "S" a smaller head.

The "U" is luscious. Don't let people convince you to change it.

As in the lc the "X" is probably too dark.

BTW, looking at your body text I got this vibe that I hadn't felt in a long time. This design reminds me of Quadraat - not because of any formal attributes, but in the way it seems to say "I've been doing this for a long time - you can trust me."

Concerning the point size:
Looking at your setting, and comparing the vertical proportions and the width to some reference designs, I think it's best suited for 10pt, and can work at 11pt. (That's assuming your lc letters span nearly the entire EM.)


Sindre's picture

Thank you very much! I have no time now, but there'll be an update tonight.

Sindre's picture

Oh boy, too exhausted after a long, long day of båtpuss (preparing your boat for the season, ancient Norwegian tradition) for anything type-related tonight, the promised update must wait till tomorrow.

litera's picture

You may have a slight problem with /ae/ (and possibly /k/ as well) at small sizes. Top part of the lower closed counter (and top diagonal stem of /k/) seems a bit too thin at its thinnest point. At smaller text sizes it may turn out as if printing is bad since it may come out too thin and there may not be enough ink.

Bendy's picture

Type spectacles now on.

A appears slightly light, perhaps. One place you could add weight would be the crossbar, but this might not be the look you want. I tried this in my so far very underwhelming Dutch Antiqua's A.

The tail of C still bothers me; I'm expecting a little more definition, maybe not as far as mirroring the top serif but something. Hm, not sure.

The stem of D seems light next to the large counter.

I wonder whether an Emerson style Q would work here. (Top left image in this post)

Agree with Hrant about the lc u being a touch wide; however for me the G works beautifully.

I wonder whether z could have a spur or two on the serifs or even on its angles.

Looking forward to seeing numerals and punctuation. I'm still thinking comma and quotes are too modern looking, but I'm a stickler for commas! :)

butterick's picture

"A straight line is a dead line"

Bloemsma's ideal is nice, but the reality in digital fonts is that you're always having to cope with a grid. Lines that are long, flat, but not quite straight tend to rasterize with bumps and jogs that look like mistakes.

It's fine to avoid straight lines as part of your design concept, but if a line wants to be straight — make it straight. Don't knock it three units off plumb just on principle.

hrant's picture

I'm a big fan of accommodating the grid* (which is NOT to say I'm a fan of inventing a grid as a feelgood or a cop-out). But if the grid in question is finer than human perception (which some people believe is anything above ~1000dpi; in any case 2400dpi seems pretty safe) then one can say it does become moot.

* Proof: I love bitmap fonts.

> Don't knock it three units off plumb just on principle.

What's a better reason than principle?


Sindre's picture

Phew, things take time. Been tinkering forever after the last round of critique. This is what I've done (see pdf):

A: A little weight added both on crossbar and stroke.
C: Slightly meatier tail.
D: Weight equalisitation.
E: A little narrower, some polish
G: Same shape, but slightly less extreme
J: Prettier curves under the baseline.
K: More tension in the foot.
M & N: Finishing polish
P: Significantly larger bowl
Q: Oh boy, oh boy. I've tried a lot of stuff, and wound up with this. Couldn't make a crossing tail model work. That's a shape I just don't understand, obciously.
R: Polish.
S: A lot of subtle tweaks.
V, W: Polish.
X: Less black, polish.
Y: Polish.
Z: Less black, polish.

Slightly narrower (and weight-compensated) 'm' and 'u'. 'x' is lighter, and significantly tweaked. 'k' is slightly weight-tweaked. And I've made a working "æ", the old was just a leftover from several iterations ago.

Is "gj" that serious a problem?

Oh yes. 806 words in Bokmål, 2967 in Nynorsk. Grammatically correct sentences with /gj/ in every word can easily be constructed. Very, very common in Icelandic, Faroese and Albanian too.

[...] but in the way it seems to say "I've been doing this for a long time - you can trust me."

That is some compliment. Thank you! (I wish I felt the same way.)

I wonder whether z could have a spur or two on the serifs or even on its angles.

Possibly, I have doubts about both zeds, actually. Something to look into tomorrow. Along with numerals (the top terminals on 2 and 3 must go, mustn't they?) and punctuation.

Sindre's picture


Sindre's picture

[...] my own Paphos [...]

After seeing that 'j', I'm very curious about that typeface. Any other bits of it you possibly could show?

hrant's picture

I've shown the "Q" on Typophile before, so here it is again:

The logic behind the "Q" is that making the top less like an "O" is key.

I have roughs of all the basic alphanumerics,
but it's quite possible they're all wrong!


Sindre's picture

That's ingenious. The Q is a bastard of a glyph, posing as some easy-going vowel while being the awkwardest of consonants. It's wrong that it shares 'O's shape.

Paphos needs to be finished, by the way.

Sindre's picture

Some unifying tweaks on f, r, y. (And a working ø)

hrant's picture

The "t" -a central character- is too tilty. You could keep the stem inclined
but maybe turn the crossbar a bit counter-clockwise to offset the effect.

BTW, thanks for the compliments and encouragement.
Just for that I'll show you my killer "g"... :-)


Sindre's picture

Whoa, that's a weapon for executing infidels! The negative spaces have really beautiful shapes, and somehow doesn't quite share the aggression of the strokes. I like the lemon-shaped bowl too.

Here's a (hopefully) improved 't':

hrant's picture

Yes, I think it's the second-most-violent glyph I've ever made... :-)
Negative space: it is the female in Yin/Yang after all!


The "t" works now I think.

I would balance the widths more in the "æ".

Isn't the "j" too dark?

BTW, your caps settings exhibit uncommon grace and power.


Trevor Baum's picture

I absolutely can't wait to see the italics.

We need to start coming up with names for this... :)

hrant's picture

I love Typophile.


eliason's picture

Glad you're finding my pangrams useful.

Latest /t/ adjustments are very nice.

Should the top horizontal of /Z/ be a bit longer (bring the top terminal leftward)?

That enlarged P bowl went a bit too far IMO.

I would consider bracketing the spur on eszett, at least on top.

BTW I gave my type history students pdf's of this and Telefon to have them practice their crit skills!

Sindre's picture

Thank you, guys. I need to rest now (getting type blind), but I will post some updates tomorrow.

[...] settings exhibit uncommon grace and power.

What can I say but thank you. Those words I'll remember.

Trevor, here's an early sneak preview of the italic. (I doubt the closed 'g', really a joke, will survive.)

I agree with the 'Z' and the 'P', Craig.

It would be great fun to hear what your students make of my typefaces.

I love Typophile! (Who needs a formal education in type design? I've learnt more from this website than I've ever learnt from any teacher.)

Sindre's picture

[...] the second-most-violent glyph I've ever made

I forgot to ask the obvious question. Please?

hrant's picture

OK, you twisted my arm. :-)
It's from Brutaal (part of the Daam Entity).

But that's not counting the question mark I made for
Font Aid II, since the violence there was self-inflicted...

Anyway, enough about me.


Bendy's picture

>I love Typophile.
Hear hear.

Loving the italic k, I had something similar going on with Nola.

Off topic, sorry...Hrant, that g is phenomenal. We should start an unexpected-g club or something. Is there anywhere we can see more of your stuff? I've only ever seen Patria and would love to see how your ideas translate into full alphabets.

Sindre's picture

Another slight update. Now it really is bedtime.

Wider 'Z', moderated 'P' bowl, lighter 'j', equalised 'æ'. The æ shape is unorthodox, by the way.

Ouch, that 'g' looks like a combined mezzaluna, sickle, leather cutter and scythe, with attached rose thorn and sharktooth. It also looks like a blackletter from another planet.

Sindre's picture

Great k, Ben! It looks like it tests if the new ice on the river is thick enough for skating on.

Bendy's picture

> The æ shape is pretty unorthodox, by the way.

Can you say any more about that? It's another letter I don't use, so I'm interested to hear your observations.

Sindre's picture

Yes, usually the stem of the a part goes all the way down, emphasing the left part of the glyph, like this:

I've chosen to keep the left curve of the e part intact instead, breaking the a stem. Caslon does this too, so I guess unorthodox was too strong a word.

Quincunx's picture

I'm quite loving this. I must say, though, that I think the capitals are a bit... tame, compared to the lowercase. Perhaps you could incorporate some of the interestingness – if that is a word – of the lowercase into the capitals as well? The g does stand out a bit for me, the top bowl looks to be a bit light, and I think I would prefer a closed loo, although it's better in the last image you posted than in the earlier images. The æ looks fine to me.

I think your italic sketch is ok in itself, although I'm not sure how well it pairs with the roman. Maybe it's the axis of the contrast; slightly diagonal in the roman, but pretty much vertical in that italic sketch.

Sindre's picture

Thank you for commenting, Jelmar.

I've thought the same thing about the upper case myself. But isn't it always like that? I think this typeface needs some kind of serenity in its upper case, as so much is going on elsewhere. And it isn't that conventional, is it? I'd say U, J, G, R (and to a lesser extent the new Q, which no-one has commented on, by the way) are all somewhat progressive for a traditional text face. I tried injecting more of that contrast-trickery into the upper case alphabet, and found it lost character rather than gained. I think almost all upper case alphabets are less fun than their children. Unlike them, the adult letters have a unifying design and inner logic that needs to be there.

I'll have another look on the 'g'. I'm very satisfied with it waist down, but the upper part may need something.

Here's an ampersand, basic punctuation and three different question marks. Please help me decide with one deserves further development.

Another thought: Not once while drawing have I been using historical models for my letters, or even looked much at other typefaces. All my previous attempts have been quite east to classify. But which category would you put this typeface in?

hrant's picture

The ampersand is OK. Just OK.

I like the third question mark - the others are too old-fashioned.

Classification: I wouldn't worry too much, but maybe Oldstyle? Or Neo-Oldstyle? :-)


Quincunx's picture

Classification... I think the proportions are fairly classical/humanistic (i.e. height to width ratio etc). But its obviously not a straight up humanist face.

eliason's picture

I vote for question mark 1. It's better looking than 2, and more fitting than 3.

Sindre's picture

I don't worry a second about classification, I rather want this to defy classification. I've been thinking it would have to go into that hopelessly diverse transitional-drawer because of a and g, (and upper case proportions, aren't those too equalised for a renaissance face?) but then there's the ambiguous e and all those cusps ... yes, I think this is slowly coming together.

Just OK.

Then it won't do. I feel the same way; its prettiness outdone by its blandness.

I think a synthesis between question mark 1 and 3 could work. I agree that no. 3 is most interesting, but not quite right. I think I can fix that. (This is why I rarely make variants. Decisions should make themselves.)

By the way, what's your favourite ampersands, good people? I need inspiration ...

hrant's picture

I like ampersands that are topologically simple, featuring one or more of: open top, "missing" upward arm, internal part(s) "removed".

My favorite ampersand is probably the one in Triplex Italic.


hrant's picture

BTW, I'm all for avoiding frivolous variants, but sometimes it makes great functional sense. Example: your "j"! Have a strong one as the main form and a shy one for following a "g".


Trevor Baum's picture

In that last image you posted, the H and R look so much darker than the lowercase characters - maybe you should look at the stroke thickness of your capitals again?

eliason's picture

Pretty good thread on ampersands here:
And more inspiration here:

Sindre's picture

Thank you, guys.

New ampersand. I've tweaked it for countless hours, and have gone totally blind. Can this structure work? Is it wrong to adjust the arm to x-height? Is it too dark?

Revised question mark.

I've been worrying a little about the colour difference between l.c & u.c. too. I'll keep that in mind.

Hrant, The problem with the old 'j' is that no other glyph has kept its traits. Originally, f, g, r, y, j had that beak. Do you really think it could work all alone? Er ... perhaps it could be made to resemble 'g' and 's'.

Syndicate content Syndicate content