I don't really have any major issues with your o. Only thing I might do is stretch it out vertically a bit, the bottom of the e looks a lot lower than the bottom of the o doesn't it?
either that or make the e less tall
Note the *counter,* the interior space on the n and o. In your version of Garamond the o counter seems even smaller than the n. Again, counters count!
@ Ryan: The |e| and |o| were exactly the same size. I suppose they might look different due to their different proportions, but from what I can tell other fonts (like EB Garamond) also use the same overshoot for |e o| even if the letters have diverging proportions.
@ William: I've made the |o|, along with |b d p q|, a bit narrower, and I agree it looks rather good. They're probably still a bit wider than the |n|, though.
I've completed a first round of revising the lowercase letters. How does this look?
I've had some trouble optically correcting |v w y| without making the strokes look wobbly. However, I've had some rather eye-opening aha! moments with |b d p q|, which look a lot better with tapered joins and smooth curves.
PDF version for free scaling: ftp://ftp.mpia.de/pub/thalmann/eau/rev1.pdf
a long way to go still
Alright, this is clearly beyond me. I'll focus my energies on my display font projects, where such things are more acceptable as "quirkiness".
@ Ryan: If you're still interested in taking the lead on this, let's talk terms. E-mail is probably best. I'd certainly want to remain a co-author of the font and learn a thing or two from the experience.
Because the join of the n and the top of the b are at about the same angle, they should have the same thinning for the consistency of your design. My guess is that the right weight to look monotone, if you want that, is in between the two, but you decide. I think consistency is more imperative for a well-crafted face.
Check out Rosario by Hector Gattihttp://www.google.com/webfonts/specimen/Rosario
Somebody should respace that.
The Rosario fonts where in use during a lot of years at the Cosgaya Chair in the UBA for all their printed materials and they where also used in print for a few books before being released at Google Webfonts this year.
I'm now flipping through a book set in Rosario, and I don't see any spacing problem (Scan attached).
Look right there in the first word (of the GF link). Please don't tell me "rum" is balanced. In your (arguably too low-res) print sample, look at "estiró" (at the beginning of the fourth paragraph).
I wonder, how many hours did Gatti spend on spacing? Just for reference, I'm currently finishing a spacing commissions: limited character set, two weights (with very close Italics) = about 15 hours. Zero automation BTW.
I agree on "rum" but not sure about "estiró".
Also, worth saying that this kind of things usually happen to "print-world fonts" when used on the web. Web browsers do all kind of strange things to space and kern. Some sizes work better than other, may be due to rounding errors, hinting, and 'who knows what' other rendering variables that come into play.... Testing it in print may be a good idea. I know this font was highly prized/admired among the Argentinian type-scene, so I'm very happy that it's now available for everyone to use it.
> I wonder, how many hours did Gatti spend on spacing? Just for reference
I don't know. Ask Hector, or maybe Pablo Cosgaya.
Even better, since the font is now libre and the source code is available, you (or anyone) can also help to improve it.
not sure about "estiró".
You don't see how "tir" is much tighter than the rest of the word, especially the "ró"?
Web browsers do all kind of strange things
Granted, but frankly that's clearly an excuse here. Rosario is poorly spaced, plain and simple. And Google should have gotten (read: paid for) better spacing.
Please don't invent excuses just because a font saves some people money. In fact if saving money is relevant to some people (users) why would it be irrelevant to others (makers)?
this font was highly prized/admired among the Argentinian type-scene
I don't doubt that. Just like Mrs. Eaves was (and still is) so heavily used, even though the spacing is -also- strictly amateur.
you (or anyone) can also help to improve it.
The way that works is I wait for somebody to pay me to fix it. Unless there's some big sociocultural contribution involved. But Rosario is just another cutesy sans.
Don't give up, I think this has promise!
Whatever the opinion of an amateur such as myself is worth, here are some of my immediate impressions...
It looks quite 'casual' at a glance, as if it's meant to evoke neat-ish hand printing with a pencil. If that's what you're going for then it's pretty nice. On the other hand, that doesn't really mesh with 'faithful evocation of Garamond' in my mind. On closer examination, I think it's only a few elements that create that casual effect.
I realize it's a faithful interpretation of the Garamond /a's tail, but I think you're being a bit too literal. Surely it's the spirit of Garamond you want to capture more than the rigid shapes. If it helps, imagine that the tapering end of Garamond-a's tail is a serif rather than a base stroke. You dropped the serif from the top of the /a, the bottom should balance out in similar fashion. (Or so it seems to me.)
Otherwise, I think you've got something with definite potential here. Besides the elements I mentioned, i quite like the overall shape and feel.
@ altsan: I'm not giving up at all. To the contrary, I think teaming up with a pro to do this idea justice is the opposite of giving up.
As a matter of fact, the letters |a| and |e| that you criticize are the ones that carry the most "Garamond" spirit as I see it. The |a| is my favorite character of Garamond (at least of the prettier realizations of Garamond), and it is also my favorite character of Eau. Clipping its monkey-tail would undo a lot of the font's character and bring it closer to being "just another cutesy sans", as Hrant might put it. As for the |e|'s eye, I was following the Garamond shape more closely in an earlier version (with a soft corner on the bottom right of the eye), but didn't like the looks of it.
I agree the |g| could stand a bit of rounding. I already smoothed out the shape of the connecting stem between the two stories, which had much more tension in Garamond. Let's see how much smoothing it will stand before it loses touch with the source. I'd rather break some sans conventions than give up the premise.
As for evoking a hand-written feel, isn't that what humanist fonts are all about? ;o) Some glyphs in my draft does give me a somewhat clumsy hand-written vibe, but they're different ones that the ones you mention. What I don't like is |v x y|. They look like they're built from toothpicks next to the more alive- and grounded-looking other glyphs. Don't know what to do about that, though. The teeny weeny tip the |y|'s tail is particularly oogly, I think I'll have to take some liberties there. Maybe treat the tip as a serif, as you suggested for |a|...
A couple weeks ago, Catharsis and I entered into discussion and came to an agreement regarding Eau (temporary working name). I will now be handling the polishing and completing of the font, while continuing to incorporate Catharsis's opinions on the various proofs.
When we both agree it is finished, we will be co-releasing it, and sharing all revenue 50/50%. I imagine in the end I will have put in a lot more hours of work than Catharsis, but without the spark and initial work of Catharsis, Eau would simply never have existed. I count myself as very lucky to be part of this project, and will always view mine and Catharsis's contributions as equal.
I hope this thread will continue to get the quality of suggestions it has previously. While perhaps I am slightly more skilled than Catharis, I am certainly no master yet, and can use all the help I can get.
The first thing I did was to thicken the weight slightly, and take ever so slightly away from the monolinearity of Eau, in order to give it a more (hopefully) elegant sense of "body."
The spacing is not 100% where I would like it, but I want to get the UC, LC, and figure glyphs down before I attempt to perfect the spacing. No kerning or ligatures have been created at all yet.
As Hrant has suggested elsewhere, and as perhaps I have never fully understood until now, the lowercase g has been an ungodly pain, it has required vastly more work than any of the other glyphs, and it still has a ways to go. I have included it at a large size at the bottom of this proof for closer inspection.
I have also included at large size other glyphs that I feel need to have kind of "swashy tails." The tails of the y, the Q, and the R are intended to be the points of highest contrast at the joins in this monolinear design. They look OK to me but I would appreciate comments.
I will not be posting any pdfs or otherwise vector-based samples here. However, I hope by rendering out some large bitmaps it will provide enough to offer accurate opinion on. Also, if anyone would like to see a particular glyph or word at extra large size, I would be quite happy to create a sample and post it.
Is the wobbliness of the curves intentional?
No it's not intentional. Still ironing things out.
Could you be more specific?
That all you're going to say, Hrant?
I'm not sure how to help. If you can't see it you can't see it. Maybe it's a blessing!
But OK, here's the first one: can you see how the left half of the "Q" is "pinched"?
It does appear narrower on the left side than the right (even though technically they are both the same thickness, yes I measured, lol). But I'm attempting to give this a slight horizontal stress. If I release the "pinch" there it would look just as thick at 3 o'clock and 9'oclock as it is at 12 o'clock. OK, so then I could thicken it at 12 o'clock and 6 o'clock, but then the glyph would be too thick there compared to the other glyphs. Can you explain how you might try to fix it?
I have my own fonts to make! :-)
BTW by "pinched" I didn't mean about the thickness - I mean about the curvature.
Hmmm, do you mean it's not a perfect circle? Because that is true, its more of a horizontally wide oval on the left side, and closer to a perfect circle on the right. It's more "elongated," you could say, on the left hand side. If that's what you mean then, yes, that was intentional. It matches the deformation of the Garamond O we are basing it on.
I want to keep some of the imperfectness of the Garamond here. I'm not looking to turn this into a perfect geometrical version, to me, that wouldn't be a garamond anymore. IDK, perhaps I just misinterpreted the use of the world "wobbly?" I mean, just as a quick example, this is what I think of when I think of "wobbly."
I guess that hrant is talking of 'wobbliness' as something more like "geometric continuity".
There are a few tools that can help you improve your outlines, like the RMX Harmoniser (1) or Yanone Speedpunk (2). Using these tools you will also learn to see the 'wobbliness', by comparing the before and after results.
Where do you see the wobble, Pablo?
In those places:
This is all very educational, keep going. :)
So by wobbliness you mean a discontinuity in curvature? But then, is it really possible to achieve a non-wobbly transition between a straight segment and a line with Bézier curves?
I also didn't know you couldn't make a circle with Bézier curves. They sure look circular to me.
I wish Speed Punk was available for PC.
If I got rid of all those whatever-they-ares, would the design actually be better? Maybe it would only be more bland?
Only one way to know: Fix a few letters and compare both version.
For example: You can fix /h/a/m/b/u/r/g/e/f/o/n/s/t/i/v, generate a new font, and quickly compare both versions using the drag-and-drop testing page at http://www.impallari.com/testing/
Crappy curves are not what you want to give your typeface interest.
Hmmm, perhaps it's time to take this over to TTF. Quadratic curves to the rescue!
Craig, I wish that were true. :-/ We've been seeing so many wannabe-fauve fonts become so popular lately.* But they're intentionally so, and that's key; Ryan, you need to know what's going on. If you can't even see how/where it's dorky, I can't see this working out. I mean, some parts might be non-dorky hence out-of-character, and you wouldn't even realize it!
Let me add to Craig and Hrant. This is the blind leading the blind. You guys (Christian and Ryan) need to study what Briem says, and follow it, for a start. Those strokes are hideous.
Lol, hideous, he says. Hey don't hold back on our account, William, tell us what you really think.
Ryan has stepped back from the job for now (and from the forum in general, as I understand).
Does anyone of you pros feel like picking up the torch? The baseline offer is 50% of the net earnings after MyFonts' cut, but we can negotiate.
Hi, I did not comment before since there have been several interesting attempts in producing an almost monolinear "sans" version of Garamond models.
I think it would be good to check out how things have been approached by:
- Alan Meeks - Claude Sans (1988): http://www.myfonts.com/fonts/itc/claude-sans/
(it’s rounded, but it doesn’t matter: it is well drawn).
- Yasuhiro Yamaoka - Practica (1991-2002): http://yoworks.blogspot.it/2009/02/yofonts-practica-at-beginning-in-1991...
(forms are based on the proportions of Apple Garamond).
- Frantisek Storm: Jannon Sans (2011): http://www.stormtype.com/family-jannon-sans.html
(not Garamond, and pretty modulated, but may give some useful ideas).
@ Piccic: I'm aware of Claude Sans; it's been mentioned before in this thread. Personally, I think it's butt-ugly, so I can't feel threatened by it. :P Practica doesn't strike me as particularly Garamondesque; probably due to the condensed proportions of the Apple Garamond it's inspired by (I think "based upon" would be saying too much). Likewise, Jannon Sans suffers from being based on a particularly uncouth version of Garamond — I honestly don't understand why people took Jannon seriously in the olden days, much less why they still do these days. It's just plain malformed. Just look at that sad, beer-bellied |a|! ;o)
Anyway, the offer stands.
Claude Sans is well drawn, not so much usable, but well drawn (not a surprise since it’s from Alan Meeks).
Practica is really well done if you consider Yasuhiro did not have any formal type design education, and went on drawing and redrawing, which is the only way to get the forms right. Considering Apple Garamond as the starting point, I think he managed to keep the essential characteristics of the forms while designing it.
I have always liked Jannons, so it doesn’t make much sense for me to comment about those: clearly they aren’t "proper Garamonds", but saying "malformed" doesn’t make sense.
Considering Apple Garamond as the starting point
I guess that was the problem. :-)
I wouldn't call Jannon "malformed", but I don't like it either. That "a" does need to be flogged in a Parisian square.
I did not like Apple Garamond so much myself, but Practica is good.
I mentioned it because you can always look at the solutions he adopted and apply them to an original Garamond.
@ William: That's an interesting proposition. Just looking at those other sanses inspired by Garamond, I can see there's a lot of different approaches to this theme, which are all notably different from what I have in mind. It makes sense, then, for me to establish a solid definition of what I'm trying to achieve before I pass it to someone else. I'll try to put together a polished hamburgefonstiv as a first step (though I'm sure I'll need a lot of help on the way*). I'll worry about completion and polishing if and when I get there.
@ I agree that Practica is a nice font, it just doesn't strike me as very Garamond-y. As for the intentions behind Eau, I'm certainly not aiming at a book font. I see its niche more along the lines of, say, Gill Sans Light — an airy display font capable of carrying short texts.
*I think I understand the concept of a "wobbly" curve now... but as for Briem's illusions, I think I'll simply have to follow the recommendations blindly. I don't see those things yet. Actually, I see the heavier strokes of the Claude Sans capitals as almost comically out of place among the lowercase letters — and that font has been quoted as being particularly well-drawn... :P
I'm wondering whether returning to a lighter work weight might make that process easier or harder... If I only ever complete one weight, I'd want it to be a light one.
I'm not that impressed with the other attempts to get sans from Garamond, which Claudio posted. That tells me that it's quite a challenge, which is what I suspected.
I commented here because I thought there was something in the letter shapes that was good. But I also felt that the proportions needed to be harmonized, and the strokes and curves needed some serious work. Christian, what merits this has so far came from you. And the weaknesses. If you have someone 'take it from here', it is sufficiently incomplete that my feeling is that it will will inevitably change and probably lose what good it has. If you had a polished a-z, or even hamburgerfonts, someone else could develop it, but at this point I think it's your baby. Unlike a real baby, you also put it aside and come back to it with fresh eyes later. —I think that's pretty common amongst designers.
to be honest, to me it doesn’t even make sense the basic idea of having “a sans” out of a typeface.
Serifs are not something attached, and if that was born that way, it means serifs were part of the designs.
I mentioned those because are well done, but clearly any attempt is a departure from the original (especially Claude).
In the attempt of Catharsis I just see an attempt to get at the "skeleton" of Garamond, which is not, strictly speaking, designing a linear version.
it doesn’t even make sense the basic idea of having “a sans” out of a typeface.
Serifs are not something attached
With that I totally agree.
Claudio, I agree that sans are a different animal. You can be 'inspired by' a serif font, and I don't see anything wrong with that. But I agree it is inevitably going to be very different if it is any good. The designer is going to have to have design ideas different from the serif. It's going to be a creative process on its own, which stands or falls on its own merits. I tried to argue something like that early in this thread.
Hi William, I did not read your previous comment, as I was mostly looking at how Catharsis approached the work. I see you pretty much already focused on the problems.
To be clear, I am not saying you can’t try to capture the essential qualities of an original text face, but I think it needs more to start than just identifying its "skeletal" forms, so to speak, and I think it’s hard to produce a book-aimed typeface. But again, maybe it wasn’t Catharsis intention.
That’s why I mentioned those three: Claude Sans is clearly a titling face, decorative and not meant for long text, but it evokes Garamond-like features.
@Catharsis: Now that I see you are aiming at display/titling, why don’t you concentrate just on the lighter weight? Marian by Paul Barnes individuates the skeletons of the forms (of many lead types, or "models"), while you could try to get more into Garamond’s essence by adding some modulation or stroke treatment.
This could prove interesting, and less abstract. Something like what Storm did with Jannon, but simpler and with attention to the lead Garamonds.
Re: Claude Sans. I was speaking strictly of drawings, not of how the overall typeface works.
But I do think it works (I have used it and of course not for text): I think Meeks followed something in the lines of the Imprimerie Nationale Garamond or the like, where there is a pretty different treatment of uppercase and lowercase, and where the Italics have a markedly different angle between the two.
Alright, I've gone over hamburgefonstiv of my rather light "Regular" weight and implemented Briem's recommendations for countering optical illusions. I've used the same weight ratio for horizontal vs vertical and straight vs round stems as does Gill Sans Light, since our fonts are pretty similar. That didn't change much, since I had already been close to those weight distributions after my first try. I regularized the thickness of tapered-down connections, though, and gave round strokes an ever so slight stress along the forward diagonal.
I've also tried to make the resulting curves smooth and pretty. It sure helps to know what it is I'm trying to avoid! Not sure how well it worked, though.
BTW, I've just installed the demo version of Glyphs, and quite like the handling. The above was done in Glyphs. I suppose I'm going to buy it soon, though first I'd like to figure out whether I can import my other fonts into Glyphs (Glamour & Glory failed last time I tried).
Looks better, but I think you should cut most of the terminals (at least those of f and t) in a way which feels less “automatic”: it’s those subtleties which improve the overall design, besides the harmony of curves.