Sabon

G T's picture

Hello peoples of Typophile,

Where in your opinions would I get the bestest digital version of Sabon please?

Thanks,

Graham

G T's picture

thanks.

Why is that one the best?

charles ellertson's picture

Adobe Garamond

You have to remember why Sabon was commissioned in the first place.

Nick Shinn's picture

One of Sabon's most distinctive features is the fact that Roman and Italic have the same metrics--a criterion of Linotype "duplexing", as I understand it.

Of course, one would imagine Tschichold stretched the italic to the roman metrics, BUT, as he was starting from scratch, I wonder if the design of the roman is informed by the italic widths in any way?

G T's picture

@Charles: you have to tell me why Sabon was commissioned in the first place (please)

charles ellertson's picture

It is a long story -- too long for a single post. But how about this: Sabon is basically a Garamond -- a *Garamond* Garamond rather than a *Jannon* "Garamond" (another long story).

Sabon was commissioned in the hot metal days, so that both the Monotype and Linotype fonts would be the same.

Lots of differences here -- , Monotype is based on the em-system and Linotype is not. With an em-based system, every character (sort) must have even units in width. Originally, an 18-unit em, I believe. No such limitation in Linotype.

BUT. Linotype used duplexed mats, usually meaning the italic and roman were on the same matrix, and so had to have the same width.

Fonts for machine-setting with Linotype/Monotype are different from foundry type, which has neither limitation. (Severe limitations in setting speed, though.) I believe originally, the commission was to have all three the same, but in the event, the machine-set fonts were the same, but different from the foundry fonts. If I have this wrong, somebody on this board will correct me.

So. Sabon was a wonderful compromise, given the objectives. None of these objectives remain in today's composition. In fact, they were absent in the earlier photocomp days (em system yes, but a 54-unit em with Linotype and a 96-unit em with the last Monotype phtocomp machine). BUT. As it happened, Sabon in Linotype photocomp just happened to work out as a very good font -- it printed well. The italic was gappy, but you lived with it. The roman printed better than most other Linotron photocomp fonts

So from my perspective, Sabon Next was completely recut to get rid of the compromises which occasioned the font in the first place. That essentially makes it another Garamond (of the "Garamond" variety). Take you pick.

For more, search the web . . .

kentlew's picture

> Of course, one would imagine Tschichold stretched the italic to the roman metrics, BUT, as he was starting from scratch, I wonder if the design of the roman is informed by the italic widths in any way?

Nick -- I refer you to this old convoluted thread (ca. Aug 2003) -- starts out about trapping, moves on to a discussion of unitizing and Mono vs. Lino, and ends up dwelling on Sabon:

http://typophile.com/node/1083?page=1

Toward the end of that page, I speculate a bit about the relationship between Sabon Roman and Italic as dictated by the Linotype duplexing constraints, including an interesting illustration of the issues.

Because of the way archives from that era got all split up when Typophile revamped, that thread continues (skipping over Hrant's 2006 postscripts) in this link (in case anyone is interested in wading through it all):

http://typophile.com/node/1022

-- K.

kentlew's picture

Charles -- the earliest Linotype phototype machines were actually limited to an 18-unit system -- something to do with the limitations of the constant-drive motor. It wasn't until the introduction of a stepping motor in the V-I-P that they were able to use a finer 54-unit system.

-- K.

charles ellertson's picture

Kent,

OK, and as I remember, the earliest Monotphoto machine was similarly limited -- but maybe not. I never set photocomp before the V-I-P, and never any Monophoto at all.

More interesting, when Linotype went to the 54-unit em, they still retained the 18-unit em for the width of a lot of type. Not wanting it to be out of date, I suppose.

G T's picture

Excellent. Thanks for the info.

Another question if people don't mind answering it (just so you know this isn't for an essay or anything, I'm just intrigued - I'm not trying to get people to do my work for me…)

Which would you say is the best digital Garamond (of the *Garamond* variety, that is)?

Thanks,

Graham

.00's picture

I suppose the best digital Garamond is the one you like best.

charles ellertson's picture

James,

your post reminds me of something that happened back in the late 1990s. True story.

We got a book manuscript in. Pretty well specified, except the typeface was spec'd only as "Garamond." I called the designer, and asked "which one?" "Why," she said, "the one I have on my computer!"

Sort of parallels the proof returned via a fax, with the instructions to only make the changes written with red ink.

Nick Shinn's picture

Thanks Kent, nice piece of detective work!

So, given that the face exists in a well-populated category, it wouldn't be Sabon without the wide roman "a", which is pretty much what Charles is saying.

As so much of Sabon's identity, as designed, is predicated on its metrics, it's hard to see how Sabon Next is really a Sabon at all, other than trading on the name.

But if you have to choose between Sabon and Sabon Next, I think the role of italics in your layout would be a deciding factor--whether you want them narrower than the roman, which is traditional and normal for a Garalde, or whether the more "modern" effect, of broad italics, which is distinctive in Sabon, is more to your fancy.

(I say "modern", because it was a feature of the "pot-hook" style of italic in the early 19th century that it had a very similar character-count to the roman.)

John Hudson's picture

As so much of Sabon’s identity, as designed, is predicated on its metrics, it’s hard to see how Sabon Next is really a Sabon at all, other than trading on the name.

It is worth bearing in mind that Tschichold designed Sabon for three different companies and their respective typesetting systems, not two: Linotype, Monotype, and Stempel 'foundry type', i.e, handset type. And despite the goal of making a metrically equivalent design for both Linotype and Monotype, the three designs were not identical. The foundry type version was not constrained by the mechanical restrictions, and if I recall correctly there is some indication that Tschichold designed the foundry version first and then modified the design to fit the restrictions of the mechanical systems. Jean François' Sabon Next is actually quite close to Tschichold's drawings for the foundry version, with additional influence from the same 16th Century specimen from which Tschichold himself worked. The specimen book and history that Linotype produced for Sabon Next is worth getting hold of if you can find a copy.

That said, it remains a fact that very few people are familiar with Sabon in its original foundry version, and most typographers know it from the mechanical and derived photomechanical and digital versions. So there is a sense that Sabon properly has weird features like the huge italic o, because that is the form with which people are familiar.

Also, I do think the regular weight of Sabon Next is too light for use below 12pt, which limits its effectiveness as a text face.

charles ellertson's picture

At our shop, there was a period where we were using both the Linotron 202 photocomp machine, and PostScript fonts, but with the output still on repro paper.

The Sabon coming out of the 202 was better than the Linotype PostScript Sabon -- both to repro -- at least to my eye.

So that's point one. I found another Sabon that didn't specifically prohibit modification, and reworked the italics. A long time ago, but as I remember, I threw out all the setwidths and reworked all the wide letters. I had sample of at least one foundry Sabon to look at, though I believe it was fairly large size. Also took the look of some characters from (of all things) Bernier, an A-M offering based on some Monotype Sabon.

Well, my drawing skills rival Jake the Plumbers, but it was OK in text sizes. Taught me a fair bit, & we still use it with jobs after informing the customer.

But the main thing it taught me was sometimes it is best not to hang onto a typeface after a couple technological changes, even though it was a superlative font for it's intended use.

Nick Shinn's picture

...most typographers know it from the mechanical and derived photomechanical and digital versions.

I must admit that's true for me.
I actually have a copy of the Sabon Next specimen, but never managed to finish reading the essay.
While I didn't use Sabon in my own design work, opting for other "Garamonds", I recognized its merit in the hands of designers such as Carmen Dunjko in shift.

You can blow up page two of this pdf and see the way the openness of the italic works to maintain the tenor of the body text.

ncaleffi's picture

I've never been fully satisfied with the way the variuos versions of digital Sabon look. I've seen fews examples of Tschichold's original type from the late '60s (in books about type and typography) and it seems to me pretty much different from its digital renditions. What strikes me most is the shape of "a", for example: to my type-amateurish eye it looks narrower than the fatter "a" of, say, digital Adobe/Monotype Sabon. And the space between letters looks bigger in the digital version. In this respect, Sabon Next seems to me like a more faithful rendition of Tschichold's original work.

Take a look here:

http://books.google.com/books?id=FiJ87ixLs0sC&printsec=frontcover&dq=Ana...

It's "Anatomy of a Typeface" By Alexander S. Lawson; go to pages 151-157 and see the differences between Linotype, Monotype, Stempel and Monophoto Sabon. They look like four distinct types; and the digital version seems drawned on the basis of the Monotype version, while Sabon Next looks closer to the Linotype cut.

And here's a (very bad) scan from a book by Jost Hochuli, showing Sabon Linotype on top and photocomp (bottom).

Nick Shinn's picture

Looks like the phototype is bigger and more tightly set. That makes sense, as during the production process the Linotype image is likely to gain, whereas the phototype image is likely to shrink.

Also, tight setting was a style associated with phototype ("because we can"), which reached its apogee (nadir, it is generally said) around 1980.

Nick Shinn's picture

BTW, 588 pixels is, IIRC, as wide as Typophile will take before introducing a scroll bar.

charles ellertson's picture

during the production process the Linotype image is likely to gain, whereas the phototype image is likely to shrink

Not been my experience. Both gain, phototype (offset from repro) gains less than letterpress. Repro from metal then printed offset can gain even more; depends on the papers used.

In the two samples above, it looks like the "photocomp" type is just slightly larger.

poms's picture

Track

Nick Shinn's picture

...depends on the papers used.

Of course.
The point I was making: letterpress has ink gain to begin with, whereas phototype was likely to lose fine details, e.g. hairlines, as the image was repeatedly copied during the production process. This was why faces like ITC Garamond and Galliard were dodgy for text, pre-digital.

billtroop's picture

>Nick: it wouldn’t be Sabon without the wide roman “a”,

>Jean François’ Sabon Next is actually quite close to Tschichold’s drawings for the foundry version, with additional influence from the same 16th Century specimen from which Tschichold himself worked.

DUH. Please remember that metal Sabon is not ONE design but many SIZE-OPTIMIZED designs.

The wide roman a without which Nick says Sabon is not Sabon - - is not present in the larger sizes. (Nor is it present in the Scangraphic version which was digitized from a less familiar metal source.) (Didn't Stephen Harper go into all of this in the early 90s?)

You can't discuss historical type design while you're still thinking in terms of a typeface that has a unitary, one-size-fits-all design.

And that's the problem with type design since 1960!

>The point I was making: letterpress has ink gain to begin with, whereas phototype was likely to lose fine details, e.g. hairlines, as the image was repeatedly copied during the production process. This was why faces like ITC Garamond and Galliard were dodgy for text, pre-digital.

Some would argue that the limitless resolution of a good photo master provides infinitely more subtlety and information than any Postscript version with its crude 1000x1000 unit grid. Especially for a face like Galliard or ITC Garamond. (The spectacular crudeness of Adobe Galliard is a special case.)

Nick Shinn's picture

The wide roman a without which Nick says Sabon is not Sabon - - is not present in the larger sizes.

John Hudson has already called me on associating "Sabon" with the single master digital cut.

...the limitless resolution of a good photo master...

Limitless, until put it into production. Every reproduction in the process degraded the image, and there were many between galley and printed image. In my worst experience, by the time the photoset Galliard and ITC Garamond I had spec'd for text made it to press, the lower case "o" was in two halves.

billtroop's picture

>Every reproduction in the process degraded the image

Of course you're right about that, Nick, and this is one respect where digital workflow is much cleaner, but surely it isn't (wasn't) strictly necessary to have several intermediate generations?

Nick Shinn's picture

The original artwork of the typeface letters was reproduced photographically four times:

  1. Original artwork of letters, from which a negative image was shot onto, for instance a glass grid (Diatronic). The original artwork had pre-emptive "ticks" to counteract image loss, however.
  2. Galley made--this was a positive image exposed from the font's negative image, through a lens, onto photo paper.
  3. Galley pasted onto artboard, negative film shot from artboard, through a lens.
  4. Film assembly contact printed as positive image onto printing plate.

Compare with direct-to-plate digital imaging.

In theory, it was possible to make the galley directly onto negative film, and skip step 3, but I didn't know anyone who did that. One reason being that mechanical assembly artists were paid less than filmstrippers. Also more difficult to make proofs and do "author's alterations" to the copy (text).

charles ellertson's picture

Don't forget the effects of processing silver materials -- Lith film, and "lith" repro paper did not have a 90-degree slope.

With most photocomp fonts, we "overexposed" and "overdeveloped" to get a little more gain with the fine strokes. Of course, the nature of a silver halide crystal being what it is, the heavy strokes picked up even more density (gain). The price you paid to get the fine lines from fonts not properly adjusted for the technology.

And in making the negative, the printer tended to "underexpose" to reduce the lines needing opaquing. More gain.

On Nick's skipping repro & going straight to negative material -- I believe there was one time when we ran film, as a negative, on the Linotron 202. Frightfully expensive.

On the whole matter of different masters in metal, of course. In the earlier days of photocomp, you could purchase different master sizes for a few fonts -- an 18-point, 12-point, and 8-point were the most common master sizes. Sometimes you simply used the 8-point for text, as it was the only one where the fine lines stood a chance of printing.

And I remember Richard Eckersley using a font -- Vendome(?) for the text of a book. It was only available in an 18-point master, which meant we had to play a lot of tricks to get it to work.

By in large, photocomp was pretty screwed up, but in some ways, such as different masters, it was ahead of where we are with most digital type. I see the 4- and 5-"masters" available from Adobe as a step toward recovery. Countering this is graphic designers (not type designers) falling out of love with a typeface about as fast as casual listeners fall out of love with the latest hit song, so there isn't time for us comps (end users) to fix things, even when a font has a lot of potential.

This fixing isn't just a matter for type designers. It takes the graphic designers and the comps too; nobody, no matter how clever, can create a typeface that satisfies all needs.

ncaleffi's picture

"The wide roman a without which Nick says Sabon is not Sabon - - is not present in the larger sizes. (Nor is it present in the Scangraphic version which was digitized from a less familiar metal source."

Bill, you're a great source of historical information as always. I've never seen any version of Sabon by Scangraphic - and it would be interesting to see a sample.

Syndicate content Syndicate content