Perfect Justification

melo's picture

What is the secret to setting such beautiful type? It's not only justified and I don't see any hyphenation. What techniques are used here to accomplish such perfect justification?

Any links or advice on such techniques would be great.

blank's picture

Let your H&J system scale text horizontally 2–3 percent and edit copy to prevent/fix really bad lines.

Nick Shinn's picture

It's hard to observe the justification at this size of reproduction.
However, there are some issues in the layout.
One should try to avoid paragraphs beginning at the top of columns.
That makes it too easy for the reader to skip text, and for instance begin at the top of the second column, rather than deal with the un-buffered jump from drop cap to text, which creates a semantic schism in the first word, exacerbated by the gutter-width gap between drop cap and body text.
Function over beauty...

dirtcastle's picture

The one time when I did a bunch of this type of layout (in the good old days of Quark), I used the image and headers as my variables -- resizing them and shifting the columns up and down until they were close to where I wanted them. Then I did what James and Nick mentioned, minus the H&J part.

I'm not even sure I understand what is meant by "system scale text horizontally 2-3 percent". But it sounds groovy.

jabez's picture

Glyph Scaling, perhaps?

dirtcastle's picture

InDesign?

nina's picture

Pointing out the obvious here, but this is also highly language dependent. Don't try no hyphenation on such narrow columns in German…

Arno Enslin's picture

@ Nick

and for instance begin at the top of the second column, rather than deal with the un-buffered jump from drop cap to text, which creates a semantic schism in the first word, exacerbated by the gutter-width gap between drop cap and body text.

I am not sure, what you mean, partly because I am amateur and partly because of my English. Although I also feel, that there is something wrong with the layout, I had not been able to say, what is wrong there. It looks a bit, like the typographer followed a scheme, that he tried to follow, but it does not work in this case. It looks a bit strained.

Would you drag the headline to the left, spare the first column and the initial? And you would not resize the image? Because of the random allocation of the fat black letter block (the initial and the headline) and the white room on the right side of the headline?

By the way, I feel not good with the block between the second and the third columns of the right page. I cannot comprehend the sense of the line at the top. But that’s not all, which I am not able to get.

blank's picture

Glyph Scaling, perhaps?

Right. The roots of contemporary H&J systems go back to the work done by Zapf and others in the 1970s and 1980s, which were designed with text scaling in mind. Most fonts can handle 2–3 percent with no noticeable distortion. Some typefaces can handle more; Vialog was designed to distort 5% before it begins to show. Unfortunately many designers never use glyph scaling because they don’t read beyond the most basic books on typography which usually insist that scaling type is always criminal.

oprion's picture

Holy jumping picas!

There I was, shamefully hiding my dirty little secret of 2-3% glyph-scaling of justified lines, certain, that if ever found, I would be dragged through the streets in a metal cage, clad in nothing but tared feathers and a foolscap.

Turns out, even the venerable typophiles approve of this practice!

Talk about coming out of the closet...
_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

.00's picture

Don't be afraid to reduce the desired word space value to something well under 100%. You'd be surprised how good 65% or 70% will make most text.

karinskij's picture

Yeah, and concerning word space, never keep that last 150%. It's so strange it's the default setting. I usually go for 80 - 90 - 100 in those boxes as a standard (but now I'll see how it look even more reduced thanks to the comment by Terminaldesign and this great thread.

blank's picture

It’s so strange it’s the default setting.

I think that the Indesign team purposely uses absurd H&J values to keep lazy designers from relying on them.

Nick Shinn's picture

It looks a bit strained.

That's perfection.
But most people prefer the Botox look.
So as a typographer, you have to know how to do slick, but then move beyond and give it a little bit of edge, of roughness, of personality.

kentlew's picture

To a large extent, that 150% (or whatever) in the Max Word Space column does not have as much control as one would hope. When the H&J routine gives up and can't solve your line justification within your parameters, then it's fallback position will be to just throw in a bunch of extra space anyway. So don't think that you're limiting the spaces with that field.

Mostly it affects the threshold above which letter-spacing will be added (if you've allowed any). And it might slightly influence the calculations of a multi-line composition algorithm.

There's a danger that having too narrow a range of parameters will actually flummox the algorithm more often and yield gappier spacing, despite your desires to control it.

Arno Enslin's picture

Thanks, Nick. I did misunderstand you.

--------

With regard to the glyph spacing I will check with my laser printer, how dependent it is from the printing resolution. I belong to those ones, that never have heard before from professionals, that glyph scaling is anything else than criminal. I often was bothered by word spacings, that were too big, but I can’t remember, when I felt bothered by hyphenation. Very interesting thread. I have to experiment more with all these parameters. There are many combinations. (I once have set up a short text, in which I changed the letter spacing from line to line. Very carefully, manually and without a big jump of the letter spacing of the line, that directly has followed, although I had heard, that this should not be done. But it worked fine.)

melo's picture

Wow, great replies thank you all so much. I now know what I should be researching in order to improve my typography. This is great direction, thanks so much.

Also, just wanted to add that this wasn't my work nor was I looking for the overall design to be critiqued (though, that was appreciated) but the work of Pentagram (http://pentagram.com/en/new/2008/10/new-work-the-atlantic.php). Was looking through their impressive portfolio and just saw this type technique a lot and finally had to ask...

Thanks again everyone, if you have anymore input/suggestions... please keep it coming! I'm eager to learn. I spent 3 years in College and none of the above techniques were mentioned.

@James Puckett - what books SHOULD I have read?

hrant's picture

That's not beautiful type - that's a caricature of type.
You don't set text so that people admire how gray it is.

hhp

Arno Enslin's picture

@ hrant

You don’t set text so that people admire how gray it is.

Would you go a bit more into details with your criticism, please? I think, the aim of having no hyphenation is exaggerated, but the evenness of the word spacing also can be influenced by the slight glyph scaling in a way, that there are no big holes between words, in which the view falls. So the legibility is partly dependent from the gray of the text blocks, isn’t it? And then we would not admire the pure beauty, but the functionality (of the beauty). And for the case, that you don’t mean the micro typography, could you propose an alternate macro typography (an alternate layout)?

James Puckett - what books SHOULD I have read?

I second that question. And I add the question for German sources with regard to the H&J system.

J Weltin's picture

And I add the question for German sources with regard to the H&J system.

Spiekermann: ÜberSchrift
Phil Baines/Andrew Haslam: Lust auf Schrift
Hans Peter Willberg/Friedrich Forssman: Lesetypographie

Gibt’s alles beim Verlag Hermann Schmidt, Mainz.

Arno Enslin's picture

Thanks, Juergen. I borrowed the first edition of Lesetypografie a while ago, but I must have skipped the part about glyph scaling.

nina's picture

So is glyph scaling officially cool now or what? For the record, I'm one of the people who never do it, on principle (and I've been cutting down on the negative tracking too). There are limits to squeezing stuff in, and essentially violating the design of the font is too much of a sacrifice in my book (scuse the bad pun).

BTW, I can't find anything about glyph scaling in Lesetypografie.

J Weltin's picture

Oh, i didn’t mean the glyph scaling is mentioned in Lesetypographie. Just the mentioning on how to set the H&J.

And in fact, i agree with Nina. If one is going to experiment with extending the scaling beyond 3%, we end up with the terrible justified setting at the beginning of DTP. And i don’t design typefaces with the intention that they should be scaled horizontally.

kentlew's picture

If you click on the image at the Pentagram blog and get a larger version of the sample Atlantic spreads, you'll see that there is plenty of hyphenation. They happen to be using InDesign's optical margin alignment to hang the punctuation, and in a small screen image the hyphens virtually disappear.

There are also plenty of gappy lines. For instance, take a look at the first line in the second paragraph in the second column [Twain's atheism was famously] or the first line at the top of the last column [(Thetans ... E-meters!). Medieval].

I'm not criticizing the work. I'm just pointing out that this is pretty standard fare. Nothing so "perfect" about it.

Anything relatively competent will look that gray and even when you zoom out far enough.

J Weltin's picture

Haha, well spotted, Kent!

seventy7's picture

Gutenberg made nearly 300 characters for his Bible. I've read that this was done, primarily, to achieve even word spacing. Is this true? If so, does anyone know how many varieties of certain glyphs he created? Do you view today's glyph scaling as the digital equivalent to Gutenberg's multiple glyphs?

Don McCahill's picture

> I’ve read that this was done, primarily, to achieve even word spacing. Is this true?

And I've read that he did it because he didn't want it to look mechanical, but to simulate the look of a manuscript, his competition. For that reason there are many versions of each letter.

Of course, both could be partial reasons.

Arno Enslin's picture

If one is going to experiment with extending the scaling beyond 3%, we end up with the terrible justified setting at the beginning of DTP.

But there was no question of unlimited scaling, but 2–3 percent only. If the very slight distorsion is not recognized, but the word spacing is more even and without holes, the strict principle should be broken by the more important principle legibility. And your typefaces are probably never printed in exactly the same way, as they look in your font editor. Good is, what works. If I draw an image and make a mistake – squirting ink unintentionally for example – I always check, whether the image looks better with the mistake instead of trashing it at once. My aim is not the full control about everything, that I do, but the result. The result may be not perfect without the mistakes. But naturally you should not blindly trust in your mistakes. I don’t force me to make mistakes. If then, but not then if.

oprion's picture

Roy:

Dale Guild Typefoundry, has a nice account of their attempts to recast a full set of Gutenberg's 42-line bible type.

They ended up with 245 sorts.

They wrote:
____
"The narrower letters were used to contract the length of a line and the wider ones to expand it. The space separating words remained constant." And again, "…the narrower or wider alternate characters were skillfully deployed to yield lines of identical optical width…" (Kapr *) These are faulty conclusions. Either the wider or the narrower characters are alternates. As the textual use of the wider characters out-number the narrower ones by approximately 3 to 1, the latter must be considered the alternates. What usually makes these alternates narrower is the fact that they lack serifs on the left side. They are never used or omitted to shorten or lengthen the space required by a given number of characters within a line. Gutenberg had employed them in both his earlier founts, the Sibyllenbuch and the Donatus/Kalender, but none of the extant printed specimens used justified lines. Their inception and use is purely aesthetic. They do not represent consciously narrowed letterforms. These alternates were cut and cast to create tighter fitting of certain character combinations that produced too much white-space within words. The alternates (left side sans-serifs) always follow C, E, F, X, Y, c, e, f, g, r, t, x, and y.

For the full text:
http://www.daleguild.com/B-42_Story_01.html


_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

hrant's picture

> the legibility is partly dependent from the gray of the text blocks, isn’t it?

I don't see how. We just like to see things line up because it makes us feel in control. But in fact we're losing control of the reading experience.

> could you propose an alternate macro typography (an alternate layout)?

I don't use fonts, I just make 'em. :-) But I don't mean to be flippant. I just know when something is wrong even if I'm not sure how best to fix it. But one can certainly pay attention to good typographers who are also sensitive to the micro world of fonts; and certainly not Modernists who are obsessed with Regularity.

One simple thing I can comfortably recommend
though is this: learn to love the right rag.

And glyph scaling?! Maybe on a free font.
Not one that's been crafted carefully in every detail.

BTW, tracking is probably worse than glyph scaling!

> Do you view today’s glyph scaling as the digital
> equivalent to Gutenberg’s multiple glyphs?

No way.
Scaling is brutish. Alternate glyphs can be a great idea (although it's much more work) as long as it's done for the reader, not merely the viewer.

BTW, Gutenberg didn't do that because he had high ideals. He was simply imitating the scribes* (as Don says). He had to come as close as possible to appease the critics of printing.

* So why did they do it? Not sure.
Maybe because hyphenation was
seen as weak, hence unholy?

hhp

hrant's picture

> If the very slight distorsion is not recognized

By the consciousness? Maybe.
But we don't read with our consciousness.

And if scaling is OK, that means well-crafted [text] fonts are pointless.

hhp

Arno Enslin's picture

@ hrant

But we don’t read with our consciousness.

That’s true, but I try to consciously make use of my subconsciousness. If the view falls into holes, you are pulled out of the unconsciously reading process, aren’t you? I try to make comparative checks and I ask myself, with which alternative I feel better. Noone here said, that distorsion should not be avoided.

And if scaling is OK, that means well-crafted [text] fonts are pointless.

No. The conclusion is not allowed, because ‘if …, then …’ does not necessarily have the same truth value as ‘then …, if …’.

hrant's picture

Not absolutes, but:
The more it makes sense to scale type the less
it makes sense to spend money on text type.

hhp

oprion's picture

Eh, guess I should go back into the closet.
_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

hrant's picture

Why not just fix up the closet?
No closet is perfect.

hhp

Arno Enslin's picture

@ hrant

No, it is not not absolutely true, it is wrong. A well crafted body text typeface might tolerate more mistreatment than a poorly crafted one. Might! And I assume, that the Indesign/Quark algorithms are not scaling all glyphs in all lines and all words; if they would, the scaling would be almost useless with regard to the evenness of the word spacing or the hyphenation. Nobody here said, that there should be established a general rule, that scaling is a good thing. James Puckett already wrote in the beginning, that most fonts can handle 2–3 percent with no noticeable distortion. I don’t know, how many are most, but for sure most is not the same as all. And we are talking about body text, but not about headlines, logos or any other text, that is set in big point sizes. We all have these deterrent examples in mind. But I know extreme examples set in big point sizes only. I never have tested scaling body text with 2–3 percent, neither globally nor with the Indesign algorithm.

But perhaps it is easier, if I illustrate the logic of your sentence in another way: Imagine two men. The one was born with two arms and legs, the usual condition. The other one was born with one arm and two legs. Would you say now, that the one-arm-man better can abandon his arm? In fact, you cannot conclude anything on this base, because it may be, that the one-arm-man is palsied.

Hehe, probably the illustration is harder to comprehend. But I don’t delete it.

hrant's picture

> A well crafted body text typeface might tolerate
> more mistreatment than a poorly crafted one.

But that's not the only way a text font can be well-crafted.
In fact that's pretty rare, probably because it's too much of
a design compromise. It's sort of planning for failure.

> with no noticeable distortion.

To me that means I shouldn't be moving the vertices in
my text fonts by small amounts. But I don't believe that,
because I don't design [only] for the consciousness.

hhp

Arno Enslin's picture

@ hrant

But that’s not the only way a text font can be well-crafted.
In fact that’s pretty rare, probably because it’s too much of
a design compromise.

But didn’t you mean the care, with which a font is made? You were regarding to the technical quality with the term well-crafted, but not to the (raw) design, weren’t you? If you mean the design, I agree with you, that there are well-crafted fonts, that are more robust than others. But if a font already is horizontally scaled with 2–3 percent by a dilettante typemaker, the user would move it into the recognizable (by the subconsciousness) extreme, if he additionally would scale it horizontally. (In the same way as you loose information with every analog copy.) So the typedesigner should not say: ‘I don’t craft my fonts well, because the users don’t handle them well.’ This would be a defiance reaction.

To me that means I shouldn’t be moving the vertices in
my text fonts by small amounts.

Why should the slight softening of the non-scaling-rule of the user change the rules of the typeface designer?

hrant's picture

> So the typedesigner should not say: ‘I don’t craft my fonts well, because
> the users don’t handle them well.’ This would be a defiance reaction.

True.

But what I was saying is: somebody who thinks it's OK to scale fonts also thinks it's OK not to worry about small details at the letterform level.

And going back in the discussion, to something more practical: don't scale fonts, not merely because some type designers mind small details, but because you can assume that the type designer of the font at hand minded the small details* and a simple horizontal scaling will do more harm than good (especially if the goal is a Modernist succubus).

* Otherwise you should be using a font where that's the case.

If you need to scale type, learn how to do it well at the type design level, and maybe use OT for width alternates (assuming the EULA of the font you're using allows "internal" modification*). Don't just hit 103% in InDesign.

* And the ones that don't are being unreasonable, in my book.

hhp

seventy7's picture


No way.
Scaling is brutish. Alternate glyphs can be a great idea (although it’s much more work) as long as it’s done for the reader, not merely the viewer.

I agree.

BTW, Gutenberg didn’t do that because he had high ideals. He was simply imitating the scribes* (as Don says). He had to come as close as possible to appease the critics of printing.

But the scribes weren't producing texts with columns as tidily justified as Gutenberg's. I agree that Gutenberg's multiple glyphs enabled his texts to look more natural, but I do believe that the number of alternate glyphs and ligatures were also created so he could achieve a strict justified column with even word spacing. As pointed out here, the strict justified columns are one of the ways scholars distinguish Johann's work from manuscripts. So, it appears that the even columns and spacing in his work were not part part of his goal of imitating the scribes.

From the above link:
One way to distinguish between Gutenberg's printed Bible and a manuscript Bible is that in the printed book, all lines of text are brought to equal length in the column, forming justified lines. The strictly justified lines of Gutenberg's Bible were considered by some to be superior to the slightly uneven lines of a manuscript Bible.

Personally, I'm against scaling glyphs and spaces. I was taught by a thoughtful typographer that doing so was to butcher the balance and detail put into the font by the type designer. (Not to mention the readability issues). But practically speaking, I'd wager most typographers, book designers publication designers and graphic designers in general don't have the skill (Or the time. As you pointed out, it's much more work) to scale type at the type design level.

jordy's picture

I am against scaling glyphs. Period. Ragged right is the way. Justification, sure, go for it, but don't use glyph scaling to make up for lousy practices. And don't leave last lines in a paragraph with one word. I think that is a basic practice and a no-no. Glyph scaling is a cheap way to make text look better!? How?

Arno Enslin's picture

Glyph scaling is a cheap way to make text look better!? How?

I wonder, whether really all of you, that insist in the strict rule, have tested the slight glyph scaling in body text. I mean the algorithms of Quark XPress and Indesign. I did not test it. But maybe I am coming to the same result.

I think, there is one thing, that you forget, when you talk about well-crafted fonts. The very most body text fonts don’t have optical styles. And almost nobody of you typedesigners tell the potential customers on your websites, under which printing circumstances and in which point sizes your fonts should be used.

Typetogether writes about Athelas: "Athelas also takes advantage of the great advances and technical developments made in offset printing." Nevertheless the letterpress print of Athelas seems to work. And it really looks dark in the letterpress print (and probably lower contrasted).

I own a book, that is printed in Stempel Garamond. A dirty print in my opinion, because on the one page the text is relatively thin and on the other really fat. In other words, on one of the pages the original outline does not match with the printed outline. Nevertheless I enjoy reading it. It is very legible.

What would you think, if the user enbolds one of your well-crafted fonts? Is the digital version of Stempel Garamond well-crafted? It is not the same as scaling, I know, but probably it is also not that, what you have wished, when you have designed your typefaces.

oprion's picture

Done a little test.


_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

hrant's picture

But Arno, weren't you the one that said poor usage of a font
shouldn't make the type designer regret doing a good job? :-)

Ivan, why are the first paragraphs of the "glyph scaling"
columns shorter? Is it OK that there's more narrowing
than widening?

hhp

dirtcastle's picture

oprion... thanks for the example!

The example defitely shows the technical differences, but with only 5-6 words per line, it's not a realistic application, imo.

A 3-col layout with 6-9 words per line would give us a better feel for its readability.

Arno Enslin's picture

@ oprion

On screen we probably have to take the hinting into account. I doubt, that the differences can be estimated on screen. Did you print it out? And I think, that also a column should belong to the test, in which all glyphs are scaled, because I assume, that the probability, that the differences can be noticed, is higher then.

But to which conclusion did you come with regard to the distorsion of glyphs?

@ hrant

But Arno, weren’t you the one that said poor usage of a font
shouldn’t make the type designer regret doing a good job?

Yes, but where is the contradiction? Everyone in the chain should give his best. If I take a picture on a 1000 ASA 35mm film with cheap lenses and a converter, I hardly can work out details in the darkroom.

oprion's picture

@ Arno
I printed out that page, and personally, can't see any noticeable distortions, while the benefit in tight spots (like extremely narrow columns and unusually long words) seems to be pretty obvious. I certainly wouldn't use this when setting in a comfortable measure, but can see myself resorting to this trick when all else fails.

In the days of metal type, glyphs-scaling (or rather squeezing) was rightfully frowned upon as it damaged the type (and could potentially crack up the chase,) but in our illustrious age...seems a tad less criminal.
_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

hrant's picture

> can see myself resorting to this trick when all else fails.

Sure.
This is not a religion - everything has its usefulness.

BTW, who told you that metal type can be squeezed? The most you can do is file the sidebearings (and not just flatly - you would even take chunks off to accomodate kerning). But that wasn't really frowned upon. In fact my letterpress teacher (our Gerald Lange) once told me that you could tell how good a printer was by looking at how much filing was evident on his fonts!

hhp

oprion's picture

@ hrant
There is a reason, why people in composting rooms carried mallets. Hastily and badly crafted forms can be hammered into the chase under pressure, as improperly justified lines get squeezed by a few millimeters when the quoins are tightened. These forms, were always in danger of bursting in a little shower of lead, so the form makers sometimes poured beer on the underside to help sorts stick together. "Just a thimble full for the comp" - as it was called.

There's a great old song:

'Twas a busy day in the print shop, with orders coming in fast.
The stoneman, in a frightful stew, was harried and harassed
When up to the stone came a pressman with an injured look on his pan
He started to sing his song of woes as only a pressman can.
"That form's the bunk. The lockup's poor. The corners don't even meet.
The leads and slugs are bastard, and the type is off its feet.
The cuts are warped and rock like hell; the furniture is sprung.
The quoins are all ass-backward and the type's filled up with dung.
I've done my best, used all my kinks, but it won't take the hint;
For two whole days I've sweated blood and I can't make it print.
The stoneman heaved a mighty sigh, his eyes a glassy stare.
He threw away his mallet as he wildly tore his hair.
"You've hammered and you've battered like a blacksmith on a drunk.
You've filled it up with cardboard and other kinds of junk.
You've pounded down the work-ups and you've nailed them with a tack.
You've turned the damn thing upside-down and daubed it with shellac.
You've screwed it up in every way that's known from A to Z.
And now that it's a dreadful mess you bring your grief to me.
That form left here in tip-top shape, the lockup was the class.
Now you can take the gosh-darned thing and shove it up your"
_____________________________________________
Personal Art and Design Portal of Ivan Gulkov
www.ivangdesign.com

hrant's picture

You can (and should) tighten the lines slightly,
but you physically can't squeeze the letterforms!

hhp

lotter's picture

@hrant

Early implementations of microtypographic extensions for TeX explicitly allowed one to specify, in cases where a multiple master width-axis for a font was available, different designs in place of glyph scaling. The general feeling though is that it isn't worth the effort; unless the printing is being done to very high standards it's difficult to discern glyph scaling of less than 2% [try having your masterpiece printed on low quality paper and see how that effect overwhelms everything else — I speak from bitter (recent) experience.]

Syndicate content Syndicate content