A first serif font

altsan's picture

Hi, new poster here (although I've lurked around the margins for a while)... If I may, I'd appreciate some opinions on a new serif typeface I'm designing.

This is my first attempt at a serif typeface, and my first serious attempt at a font fit for general consumption (my previous work has been for rather specialized uses).

Some remarks about my intentions... I basically set out to explore the possibilities of a humanist typeface with non-traditional styling. (The serifs are loosely intended to recall engraved inscriptions, by the way.) In the future I plan on adding at least partial Japanese support, and one of my other objectives is to see if it's possible to create a consistent style between the Latin and Japanese characters that manages to be culturally faithful to both. (Most general-purpose Japanese fonts are woefully generic-looking, especially the serif/mincho ones.)

It's primarily intended as a document font that can also function reasonably well on-screen. The serifs are deliberately understated to help the text render with minimal distraction.

Letter designs: http://www.altsan.org/creative/fonts/new_inscription/letters.png
Sample text at various sizes: http://www.altsan.org/creative/fonts/new_inscription/text_sample.png

I'm aware of a number of extant problems. The strokes need to be darker in general (and the harmonization of stroke widths needs work). I'm also thinking the contrast needs to be increased a little. For printed text, the character spacing may be a little tight. As far as the specific glyphs go, I'm not entirely convinced by a few characters, in particular some of the numbers. I'm also wondering if the serifs might not be a bit too understated.

I'd welcome opinions from more experienced hands! Thanks...

suzuran.pdf15.65 KB
suzuran_sample.pdf11.09 KB
suzuran_120618.pdf19.84 KB
suzuran_samples_120618.pdf13.09 KB
suzuran_120629.pdf23.9 KB
suzuran_120703.pdf23.98 KB
hrant's picture

Although I personally don't find this very exciting, I think it can certainly evolve into a usable font, and the Japanese angle is certainly interesting.

The first thing you should do is decide what point size range you're shooting for this, because that's what will lead you to a good balance of color versus spacing versus vertical proportions (noting for starters that the descenders are way too long compared to the ascenders).


altsan's picture

Good point about the descenders... I'd overlooked that but on closer examination I have to agree with you.

The 10-12 point range is my main target, I suppose. I plan to do a couple of other weights, though.

I'm actually a computer programmer by background... I stumbled into typeface design pretty much by accident – around ten years ago, I did a hasty mock-up of an old bitmap font to help myself create some user interface drawings, and the result somehow got adopted as the main menu & dialog font for several software programs. So I've been trying to catch up on learning all I can about proper type design ever since.

I've grown quite fascinated with the subject now... a fondness for Japanese writing and calligraphy came along separately, but I figure I should find a way to combine the interests. :)

This is a learning exercise as much as anything else.

Thanks for the feedback!

sim's picture

Interesting design. At first glance I think the /M should be a bit wider or the vertical oblique less inclined. Same comment for the /W. However the /Z seems a bit too wide. I love those glyph: ?@ß. Keep going

hrant's picture

Ah, I'm also a programmer by training (although I've become quite rusty). Wrote my first program at age 9 on a Commodore PET, but in the early 80s made bitmap fonts for my C64 (with my brother's help) and even my Computer Science degree couldn't stop my eventual move to type design - although there is very much some overlap (especially these days). One thing I think computer people are good at in type design is spacing - let me know if you feel that too.

If you're shooting for around 11 point I'd say the color (thickness) needs to be bumped up (as you say) but the x-height is good - just shorten the descenders (quite a bit though). Also, the caps are on the big side - which means they're good at reference texts (where a reader has to search for proper nouns) but not so much for immersive reading.


altsan's picture

OK, I've reworked things a bit. The descenders are shorter and I redesigned the /S and /s glyphs, as well as the /a to some extent. The /f, long-s and eszett share a slightly flatter curve now. Most significantly, I've bumped up the stroke contrast, especially on the bowls.


Hmm... the current caps height works reasonably well at the smallest point sizes. You do have a point, though, especially at the larger sizes.

I'll up the colour a bit in the coming days. I suspect I'll keep something close to the current incarnation to develop into a "light" weight, though.

I tweaked the /Z a little in this iteration – specifically, I made the top bar a bit shorter, which may help a bit. The /M and /W I'll take a look at.


riccard0's picture

I'll keep something close to the current incarnation to develop into a "light" weight

Makes good sense.
You could also use it for a "display" cut.

altsan's picture

New milestone. I've reduced the caps size very slightly, and tweaked some other characters.

The capitals may still look a little large next to the lowercase, but I think part of that is due to the overly-compressed letter-spacing. I need to space things out a bit more.

I'm not sure about how well the tail on the /u and /d works. Anyone have thoughts on that?

I added the proposed new uppercase Eszett character, mainly because I could. :p

Speaking of which, I've deliberately drawn the foot of the long-s without a serif, in order to differentiate it better from /f. For consistency it was thus left off the regular eszett which is derived from it. Does that seem like a good idea/bad idea/acceptable quirk?

altsan's picture

OK, I've given the face a major overhaul. The overall weight has been increased considerably, and spacing has been improved. I've also kerned about a dozen of the more problematic pairs.

Two PDFs with sample text are now attached (see above). suzuran.pdf shows all supported characters, and a block of 12 pt text. suzuran_sample.pdf has a shorter paragraph shown at a range of sizes.

And here's an image of the updated design (most but not all characters shown here):

I think this is a significant improvement. It actually works better than I expected down to quite small sizes , at least off a good-quality laser printer. :p

I think some of the curves may be a bit too heavy now. I'm also not very happy with some of the counter shapes, especially the /s and /b-/d-/p-/q... but I'll keep on tweaking.

altsan's picture

I've started adding very gentle curves and even a few subtle irregularities to the characters, which I think gives it a more old-style feel. Here are "before" and "after" images for comparison.

I think it adds some needed character to the typeface as a whole, although of course at small text sizes it's not really consciously noticeable.

PDF samples suzuran_120618.pdf and suzuran_samples_120618.pdf have been added; see above.

hrant's picture

The second one makes this worth releasing.


altsan's picture

Thanks. I do think it's getting steadily better.

There are many adjustments in the latest release, which can be seen in the attached suzuran_120629.pdf (above) or in http://www.altsan.org/creative/fonts/new_inscription/letters_5.png

Changes of note include reworking the serifs to make them sturdier-looking, adjusting the 'e' crossbar, trying to harmonize the bowl weights (still needs some work IMO), and smoothing out a lot of the counters (especially in the lowercase).

I don't know, though... there's still something not quite satisfactory about it, which I especially notice in paragraphs at 10-12pt. I think maybe the stroke contrast might need to be increased still more.

I'm also thinking I should change the numbers to text (rather than lining) figures. I'm not really a big fan of old-style figures but in this case I think they fit the style a bit more. (Since I'm using Typetool I have no way to add OpenType features, so it's pretty much one or the other.)

altsan's picture

I decided a few characters (especially /b, /d, /p, /q and to a lesser extent /a, /h, /n, and /u) were a bit too narrow relative to the x height. Besides adjusting those, I've now changed the numerals to old-style (text) figures, and I think the improvement is tangible.

And see suzuran_120703.pdf attached above.

(Just for the record, this Flash image-manager thing gives me no end of grief. I've given up on trying to use it for now.)

1996type's picture

I'm starting to like this. Keep the gentle curves going.

Top of stems like in /u/ and /n/ can be raised by at least 7-12 units.

/r u v/ are too wide.

Tail of Q should be waaayyyy longer.

Serifs /S Z s z/ could use a little more meat.

Make the bowls of P and R larger

L too wide.

W is too narrow. Looks cramped.

Get rid of the gaps in 6 and 9 or make them bigger.

Spacing is a bit loose in my opinion.

Just a little more work and you've got yourself a little diamond. Keep it up.

jcrippen's picture

It’s pointy! Be careful or you might poke your /i/ out with one of those serifs. For some reason Castlevania comes to mind, but I don’t know why.

The /z/ is lopsided. Lean it back to the left more by lengthening the base arm. I think you can get away with letting its serif poke out, but don’t let it stab anything following.

altsan's picture

Yes, I agree that /W is too narrow. For some reason I was originally reluctant to let it get too much wider than /M, but I don't quite recall why anymore.

When you say to make the bowls larger, I assume you mean vertically rather than width-wise?

The tail on /Q could be a bit more prominent, it's true. On the other hand, I'm not generally a big fan of tails that sweep under adjacent characters, which is why it's confined to its own glyph space. I may revisit that, though.

These and the other suggestions here certainly have merit, I'll look at each of them carefully. Thanks!

What's concerning me at the moment is the oblique weighting in characters like /O/o, /C/c and to a certain extent /e. I'm starting to think those characters are 'leaning' too far to the left.

I'm currently rather busy preparing to move from Japan to Canada so I don't expect to get much work done on this for the next several weeks...

altsan's picture

I can't seem to add attachments to this thread anymore... I tried to start a new one a few days ago, and again just now, but I only get "internal server error" messages.

Anyway, I've incorporated many of the suggestions made here, and I think things are really taking shape. Hopefully I'll be able to show the results soon (once the forum decides it likes me again).

altsan's picture

Here's the latest. I think it's looking quite good now, although there are still a few areas that I'm not quite sure about (for instance I'm wondering if the /e is a bit too wide).

And more extensive demonstrations:

Among the many changes since July: the tops of x-height stems have been raised very slightly, the letter-spacing has been tightened up, the /O/o/C/c/Q/e have been heavily reworked, /n and /u were narrowed a tad, and /S /s and /Z /z received significant overhauls. Various other tweaks were made to get a more calligraphic feel.

hrant's picture

I would shorten the descenders.


eliason's picture

Looking good.
Serif at the top of C and G maybe hangs too low, and its shape might need tweaking.
Extending the right baseline serif of r and f more would ease spacing gaps.
The diagonal of 5 looks too slanted to me.
The accents look rather mechanical compared to the construction of the letters.
Titles, periods, and commas look too slight.
Ear of g looks a bit sad.
Top right of U looks wimpy. Could proportions of N's serif work there?
Is the difference in width between f's hood and j's hook an issue? I'm not sure.
I agree about shortening the descenders.

altsan's picture

Will take another look at the descenders on /j /p /q, maybe /y; OTOH I've experimented with a shorter /g and didn't care for the result.

Inclined to agree about the /U. Possibly something between that and the /N serif.

I think you're right about the /g's ear - I've never been really satisfied with it but haven't managed to come up with a better one yet. Maybe it should simply be flared a little more.

The slant on the /5 is meant to loosely match the angle of /6 and /7 in the same region.

Good point about the baseline /f /r serif. As far as the hook goes, I'm not perfectly happy with either the /f's or the /j's, but I decided early on not to go with the 'inverted-f' style /j.

I reduced the serif on /c and /r considerably in the last update, but I was admittedly less aggressive in doing the same to /C /G. Will revisit.


hrant's picture

A nice dark corner of the craft:
This is a text face (where individual glyph beauty needs to be tamed "for the greater good") and the "g" happens to be exactly the glyph that needs to start looking a bit ungainly (due to being vertically cramped) for the font to be properly balanced (by not misusing the vertical space). Basically in a text font if the "g" is too beautiful more important things are going wrong.

Also: it's very rare for descenders to be at different depths (essentially for good reason) so whatever depth the "g" ends up at the others should (normally) follow.


eliason's picture

What *is* the "good reason"? If t needn't go as high as b, why must p go as deep as g?

hrant's picture

The good reason is that since the space is available to use* it should be used, to amplify boumas. On the other hand there is a case to be made in terms of bouma divergence (but much more in the ascenders than the descenders).

* Thinking of collisions.

BTW, I do make my "t"s higher than most people (and my tittles larger) but making it as high as the rest just makes it too ugly. Or maybe it's that I don't have the guts (yet).


altsan's picture

OK, descenders have been shortened (slightly). Various serifs have been improved, and the consistency of stem widths have been re-checked as well. Many other tweaks, including a narrower /e, modestly redesigned /f and /j, improvements to /3 and /5, larger bowls on /P and /R, slightly larger tittles, slight changes to accents, etc.


altsan's picture

The more I look at the "2" the more I think there's just something not right about it. I can't quite put my finger on what, exactly, is wrong, though. Does anybody else think the "2" is awkward-looking?

altsan's picture

Answering my own question... the bowl weight was poorly balanced and the counter a little off. I've adjusted both.

Also just changed the tails on '3' and '5' so they look more like oldstyle figures.

altsan's picture

Another batch of changes. 'f' and long s reworked again. Lower arm of 'E', 'L' reshaped. Stem of 'J' changed (now slightly concave on right as well as left). Widened 'AE' ligature. Increased colour of '4' and '7' a little. Also adjusted a few of the counters, especially 'b' and 'p'.

Also added several new characters, including section, 1/2, superscript 1-3, various diacritical forms and a few more ligatures. I also put a set of modified lining figures back in (no OT features, I just stuck them into the PUA); I noticed that when the PDF is generated by Scribus Acrobat only shows them as a second set of text figures (other PDF viewers seem to show them properly though), but the PDF here was generated by OpenOffice which seems OK.

Image: http://www.altsan.org/creative/fonts/suzuran/suzuran_20121017.png
PDF: http://www.altsan.org/creative/fonts/suzuran/suzuran_20121017.pdf

altsan's picture

I've concluded the "pointy" serifs were a bit too much so... at display sizes they gave the impression of ink running off the page, and didn't have enough weight at text sizes.

I also decided the "beak" on lowercase /c and /f wasn't working.

Old characters in blue, modified characters black. I've strengthened both upper- and lowercase serifs, and significantly reworked both /c and /f.

I've also addressed a longstanding flaw: the apostrophe, double-quote and asterisk were all very weak. They've been completely replaced with new versions.

hrant's picture

That's a nice shuriken of an asterisk! But I still prefer five-pointers (and personally I make them upside-down).


eliason's picture

I like those changes on the letters, especially the "beak" change.
Punctuation I'd want to see in context to judge.

altsan's picture

@hrant: I settled on six points because I thought it was a bit more reflective of the fifteenth-century designs which I'm trying to pay tribute to. (I just made a quick try with five instead of six arms but I didn't really think it worked.)

@eliason: Thanks. The new quotes in a short sentence:

sim's picture

I like the six-pointers asterisk. For the five-pointers, is the space between the points should be the same width.

eliason's picture

That punctuation really starts to look Victorian to me.

altsan's picture

@eliason: You mean 19th-century in general, or a more specific style? (I suspect I'm not quite done playing with them yet.)

@sim: Yes, the first one was created somewhat experimentally - I'm still trying to figure out how best to measure rotation angles properly. I think the new asterisk does a better job (I've adjusted it some since the above was taken) but I'm still basically relying on my eyes.

altsan's picture

@hrant: After thinking more about your comments re the /g, I've adjusted it some more. You're right: I think making it less elegant-looking makes it actually more harmonious.

(Again, old version in blue, new in black.)

altsan's picture

Major overhaul of the capitals:

I'm starting in on the lowercase next. Taking a several-week break was a big help, I think.

eliason's picture

Spine of /S/ looks thinner in the middle than on the curves.

altsan's picture

You're right. Thanks. I was staring at the counters for so long that I somehow didn't see that.

eliason's picture

Check consistency of thins, too: crossbar of /A/, crossbar of /H/, and thin stroke of /X/ all look quite different widths to me.

altsan's picture

The /A crossbar was intentionally heavier than/H, because I thought it helped to counteract the more emphasized open space of the /A. However, I have moved the /A crossbar up significantly, and I'm thinking perhaps I should lighten it some more.

/X, /K /Y thin diagonals do worry me a bit. I was wondering if they could stand to be darker. Your comments suggest that line of thought is on the right track. :) Thanks!

altsan's picture

Lowercase and numerals now revamped, plus a few more adjustments to the uppercase.

I think I've finally found a /g ear style that I'm happy with.

sim's picture

Nice update. Is the |U| should be a bit wider?

altsan's picture

@sim: I dunno. You have a point. OTOH in classic proportions I thought the /U was supposed to be about the same optical width as /V. Maybe the /V is too narrow as well, though.

There's still something not quite right about the overall harmony but it's hard to say what.

I'm not really happy with the /a just yet. I quite liked the earlier, wider form (seen in various posts above), but it didn't really fit in anymore with how I'd redesigned everything else over time. I think I'll try extending the top end a little more to the left, and maybe try reducing the bowl height some more.

I also can't quite shake the feeling that /b and /d are faintly off-balance, somehow...

sim's picture

I think the issue with the |U| is the bottom. Try to down the curve, this will open the bottom. It also be enlarge to be between the H and the V. You will see after if you have to enlarge the |VWXY|. About the harmony I would say it's because some letters are too condensed. Compare the |s, g and u| to my point of view the |s| should be wider. For the |a| you could go up the left part of the top where the serif is, this will open the letter between the serif and the bowl. Keep going.

altsan's picture

I think you're right, this is a distinct improvement.

/a and /s both look more convincing. I also modifed the tails of several lowercase characters.

altsan's picture

Here are some more complete samples of the current design (as it's been a while):


Happy New Year, all.

altsan's picture

This may be slowly getting somewhere. As before, taking a fresh look after a hiatus seems to be pretty useful.

This round, I've redesigned quite a few characters, especially /a and /e.

Sample PDF (1.5 MB) including my first rough (hand-drawn) sketches of what some of the Japanese characters could look like.

Syndicate content Syndicate content