Two New Trends in Arabic Typography

Vladimir Tamari's picture

There are two new trends in Arabic typography that need to be critically discussed by the designer and user community.

One trend is Huda Smitshuijzen AbiFares' Typographic Matchmaking effort (www.khtt.net) where Arab type designers have been paired with master designers of Latin type to produce Arabic fonts that match the Latin in style. I feel the experience of working with master designers cannot but have a positive impact on the technical quality of any resulting type. On the other hand there is the inherent danger that the resulting Arabic glyphs are constrained in style and proportion to match the x height of the Latin. What is the best way for two fonts of very different languages to appear harmoniously on the same page?

Another important trend, Tasmeem, is a plug-in to Adobe's Middle East version of InDesign released by Winsoft www.winsoft-international.com . This is the result of years of research and effort by Tom Milo and his team at Decotype www.decotype.com. Here the software automatically manipulates the placement of a small number of component glyphs to produce type for the full range of Arabic and related languages, allowing the precise control of spacing, the choice of glyph variants, and the exact placements of dots and vowels. Tasmeem was originally conceived to display traditional calligraphic styles. How will the software work with newer more geometrical styles of Arabic?

I have drawn the following cartoon to show my personal take on these efforts, which are both based in Holland. The Matchmakers are to the left, while Tasmeem operations take place in the center. A larger-resolution open-license image is also attached.

AttachmentSize
Arab-Typography-in-Holland.jpg724.89 KB
behnam's picture

Vladimir:
>>The Latin-Arabic balance in size and style is excellent. This is so in the outline and shaded display fonts more than the small text font I wonder why - probably because the added textures and stylization add an extra level of uniformity.

Yes you might be right. But I thing there is two issues here. The size and the face. Larger size may be more forgiving in magnification discrepancy. Also the bold face has a better harmony. This also may somehow be related to size issue. The outline and shadow are built from the bold face. The regular face has lesser harmony and I'm not sure how to explain it. Your remarks were quite good.
But bear in mind that the small size is size 12 all around. So in terms of magnification and line spacing it's quite acceptable. I took a picture of one paragraph of the same sample, same size, with a fundamentally Roman font, Arial. You can see the difference.

Thanks Aziz for your references.

piccic's picture

This is so in the outline and shaded display fonts more than the small text font I wonder why - probably because the added textures and stylization add an extra level of uniformity.
My opinion is harmonization largely depends on the overall features of the two alphabets (Latin and Arabic), and this is difficult because Latin may have such an infinity of stylization nowadays, but not all Latin text typefaces conform to ever-evolving models confirmed by "tradition".
Some of them may have an excess of sophistication, and I consider Palatino (used here) among them.
We tend to recognize it's a matter of taste, but I think a typeface like Palatino is objectively too "refined" and characterized to work perfectly as a book typeface.

The regular face has lesser harmony and I’m not sure how to explain it. Your remarks were quite good.
Tentatively, I'd say it's because of serifs. Latin text types evolved over so much time and the variation/structural element provided by serifs proved to be important, but most of serif faces new designs (especially from the last century), tend to have a feeling of excessive "finishing" compared to Arabic, which keeps the vigor of "handwriting" in models like Naskh.
My estraneity to Arabic does not allow me to completely confirm this, but from what I can judge visually, it's something like that…

Plus, Benham, Arial is the worst choice you can make to pair with an Arabic text face. Arial's monotony and homogeneity of widths is really far from any Arabic style. Maybe Arabic will look (also) like something like Arial in the future, but I hope not… :=)

hrant's picture

Claudio is right. Every Arab type designer and his umm seems to fixate on Arial, and the results always look like a portly gentleman in a speedo. Ubiquity is no substitute for appropriateness. What you need is a Latin font that gently leans towards Arabic in some dimensions, for example being small on the body or having [gently] reversed contrast. Take something from Excoffon or Bloemsma, and reduce the x-height.

hhp

Saad Abulhab's picture

Claudio wrote:

>>because Latin may have such an infinity of stylization nowadays,
>>but most of serif faces new designs (especially from the last century), tend to have a feeling of excessive “finishing” compared to Arabic, which keeps the vigor of “handwriting” in models like Naskh.

This is an excellent way to explain it. Enforcing cursive (i.e. handwriting) today limits both design creativity and utility progress of the Arabic script and typography. The all out letters' connectivity adapted by Arabic, among several other scripts of the Arabian Peninsula and surrounding areas, in the 2nd or 3rd century was a revolutionary step, but it had outgrown its main goals. Cursiveness should be preserved as an option not as a rule for design.

-Saad

AzizMostafa's picture

@ Enforcing cursive (i.e. handwriting) today limits both design creativity
and utility progress of the Arabic script and typography.

MirEmad of MaryamSoft proves Otherwise:
http://typophile.com/node/48495
Other tools are useless compared to it.
Without Exaggeration + with due Modesty + Flowers

behnam's picture

Thanks all! I'm listening.
But vertical space remains a physical truth. The choice of Arial by the way, was not so much optional. Arial, Times NR and overwhelmingly Tahoma are the fonts that present Arabic to most computer users. Where, the combination of Arabic and Roman is present the most.

hrant's picture

> vertical space remains a physical truth.

Vertical space is only one factor. Truth? What about the truth of readability, and that of cultural authenticity? When you impose vertical proportions across two scripts you either kill one of the scripts, or both. True functionality sacrificed at the altar of Modernism.

And guess how Latin and Arabic would have to [not] relate!

> Arial, Times NR and overwhelmingly Tahoma are the
> fonts that present Arabic to most computer users.

Not if you make a substitute. There are virtually no Latin fonts that are good companions (as opposed to masters) to Arabic. Please make one.

hhp

piccic's picture

Please, don't get me wrong, I think Benham is doing a great work by extending these families and working on overall design proportions (BTW Benham, were those open source types or types produced by private designers and offered publicly?).
Hrant is right about the problem of vertical spacing, but it's not always so crucial; what I think it's important is experimentation, to have your intutions tested in real usage. Manuscipts and earlier inkunabula mixing different alphabets turned out excellent because of this, because a theoretical approach should not be seen as real life.

I also think Saad and Aziz's approaches are both important at the same level. My tendency to abstraction and conceptuality would bring me close to Saad's daring simplifications, but at the same time I feel the need for a typeface to incorporate all the tradition gathered over centuries of history, and so I cherish the approach gathering from manuscripts.
Should I approach an Arabic typeface, I would get as classical as I can, while having a sort of set of "small capitals" experimenting along Saad lines…

My phrase, quoted by Saad, was in fact to say Latin text typography has become too often "excessively slick". This is fine in many cases, but not for books. For example, staying with Zapf, it's fine for Optima but not for Palatino.
It's clear that Aziz has a "vision", so do I, but I think we should have deep, complementary visions, which may enrich each other, while following the lines of an excessively refined approach, like that of Palatino or Arno Pro, ends up in self-closing.

I think this is the reason there is often a misunderstanding about "calligraphy" or "chirography". I do not even remotely think type should be "separated" from writing, because type is an aspect of writing.
I mean that, no matter how you approach letters, letters have a form, and this form is alive. By refining it too much, it becomes a representation of life, not life. This tension becomes even stronger when letters bring us important things.

Saad Abulhab's picture

@ Enforcing cursive (i.e. handwriting) today limits both design creativity
and utility progress of the Arabic script and typography.

>>MirEmad of MaryamSoft proves Otherwise
>>http://typophile.com/node/48495
>>Other tools are useless compared to it.

Maryamsoft looks truly a great work to reproduce calligraphy stretching OpenType very creatively, as a typographic design tool. I wish I can use a demo. However, reproducing already "created" historic forms is a *limited* venture by nature as far as applications are involved. In my statement I meant to say cursiveness *limits* designer ability to concentrate freshly on the individual letter making it difficult to manipulate letters individually or preserve letters integrity.

As for utility (user side text processing experience) limitation, I do not want to ask how much letter dancing is involved to produce these beautiful calligraphic texts, or how it would work to use it in a software to teach Farsi.

-Saad

behnam's picture

>>Please, don’t get me wrong, I think Benham is doing a great work by extending these families and working on overall design proportions (BTW Benham, were those open source types or types produced by private designers and offered publicly?).

I am an accidental font maker. My initial concern (and still my main concern) was to get my language working on my computer. I was very poorly prepared for the issues that you brought-up and when I learned more over time, I already had put too much work on my fonts to let them go! Having said that, I'm prepared to retract any of my fonts from circulation if there is a conflict.
The Arabic characters were collected from freely distributed fonts, mostly from Iranian sites. They already had some Roman for the most part but I often changed them to something else to create a better proportionality. I don't even remember were I got the Romans because as far as I'm concerned, I wasn't making a Roman font at all.
Case in point, this latest Kayhan which I do remember. I knew that Kayhan newspaper based on which the font is made, used Palatino for Roman. To avoid any discrepancy from the free version I had, I did use the original.
But I don't have a Roman eye. I couldn't possibly instantly recognize Palatino as you did, if it wasn't myself who put it there. Roman characters in my fonts -and this is not intended to offend anyone- are thought as necessary accessories. Their role is to fill out a couple of words here and there within an Arabic text. And that is why it is so unacceptable to me that a Roman font dictates how an entire Persian text is written.
Anyhow I can track how many fonts I have distributed and it is 45 for Kayhan so far. Non of those who downloaded this, couldn't care less what was the design of Roman. The mere fact that it is proportionate is all they care. The rest, is the font ability to work on the Mac or its language coverage... plus the concern that they may have that Arabic text maybe typesets well!
So I really didn't focus on copyright issues as I should have done. I don't even know how I could do it beside giving credit to the fonts based on which my fonts are built. As I said I'm prepared to retract my fonts but this will only deprive few dozen Afghanis or Kurds or Persians from fonts for their language that they will otherwise won't have.

>>Hrant is right about the problem of vertical spacing, but it’s not always so crucial; what I think it’s important is experimentation, to have your intutions tested in real usage. Manuscipts and earlier inkunabula mixing different alphabets turned out excellent because of this, because a theoretical approach should not be seen as real life.

My first portion of reply covers part of this portion. I may not be a suitable person for your intended experimentation. If I didn't have to put Roman to cover few words here and there, I wouldn't have put it there at all. This is clearly not what you have in mind in overall harmonization of Arabic and Roman.

Now if it is not objectionable, I'm quite interested experimenting XB Zar with few Roman fonts mentioned by hrant!

Vladimir Tamari's picture

Claudio is right. Every Arab type designer and his umm seems to fixate on Arial
In my time it was Gill Sans that I (initially and partly) fixated my AlQuds font on. This included the thick and thin strokes typical of Latin fonts. Recently I redesigned the font and made it monoline, and created a Latin that can go with it (see example above). Vertical spacing is indeed important in matching the two fonts, as Hrant found for Armenian. Making the x-height the same as typical Arabic shapes made the Latin look too small, so I made it larger. In my inexperience I drew the Latin outlines rather hastily, but I was satisfied as long as the Latin did not intrude on the Arabic, which was the main thing. However first Karsten Lueke (k.l.) and then Claudio Piccinini (piccic) could not accept such a crude Latin and kindly volunteered to refine the outlines and spacing. The very subtle changes in thickness keep the monoline and other features that matched my Arabic, yet the Latin now looks elegant, assured...and Latin. I am grateful for the skill and integrity of these two amazing designers, with flowers. I will post examples when I detail all the changes that this font went through since around 1962!

One thing that makes traditional Arabic look 'Arabic' is the way the slant of the nib create thick and thin lines. The nibs for both Latin and Arabic scripts both leaned to top right as the waw and the e in the diagram below show. Actually in Arabic the stroke is less inclined to the vertical, resulting in a thicker horizontal and a thinner vertical line, but this is only for the sake of demonstration. As scripts and then fonts evolved people accepted these variations in thickness as part of the written language. In slavish imitation of Latin typography, however, many recent Arabic fonts seem as if they were a mirror-image of the Latin shapes, drawn with a reversed nib, for example a Latinized waw (in the center). This helps give the incongruous look to these fonts that Hrant so hilariously described.

piccic's picture

So I really didn’t focus on copyright issues as I should have done.

It's not so much your responsibility (I mean, here it's quite outside our reach), if Zapf hasn't been able to vindicate his rights on shameless copying of Palatino so far. If you read the wikipedia entry you'll see what I mean.

And… yes, of course, most of Latin faces accompanying Arabic (or Japanese) typefaces have nothing to do with the style of the main typeface (Arabic or Japanese).
While it's absurd an Arabic should conform to an ugly Latin alphabet for some reasons of "default choice", the opposite is also true. Each one of your Arabic alphabets look quite nice, professionally drawn, I mean. So it would be appropriate to have a specific Latin component designed as a subset.

Where needed, of course. :=)

Vladimir's [e] (very nice… :=) ) confirms what Hrant and I were saying: you need a "spontaneous" Latin approach to marry Arabic.
An idea could be Oldrich Menhart's work, something like Manuscript (PDF is a study by Veronika Burian).
See also here (a set on Flickr with a book design by him).

behnam's picture

Vladimir:
>>Claudio is right. Every Arab type designer and his umm seems to fixate on Arial

It also may have something to do with what is visually easier to read, for a non Roman eye, rather than aesthetics.

piccic:
>>An idea could be Oldrich Menhart’s work, something like Manuscript (PDF is a study by Veronika Burian).
See also here (a set on Flickr with a book design by him).

Thanks! It is very interesting. But unfortunately I'm not a designer and if I were, I don't see how a Roman design -any design- could maintain its optimum line spacing AND proportionality with Arabic, AND keeping a level of 'standard look'. Forcing a design to be optimized for a line spacing (rather than the other way around) is the same thing as forcing Arabic for Roman optimized line spacing. An excess of line spacing for Roman is of a lesser sacrifice seems to me, particularly for a font that doesn't intend to be of Roman denomination.

But I'll collect as many Roman font as I can and I'll compare their look beside the Arabic part of Zar to see. 'Mellel' word processor for the Mac has a wonderful feature that you can select one font for one script and another font for another script, and adjust their size separately, within a single mixed text. This gives you a visual indication that for a similar magnification of Roman part, how much excess of line spacing is produced. Until now, all Roman fonts that I had, produced excessive line spacing, or they had a design completely incompatible with the Arabic side.

piccic's picture

@benham: uuh… If Arial looks "visually easier" to read maybe you have never had read Latin texts of length, like books, with the due treatment for their content, too (see Aziz rightly complaining the Qur'an has to be set according to its importance). Arial has inherent features (excessive homogeneity and lack of modulation for one) to be almost useless in books.

This is a problem which belongs to all sans-serif designs produced in the 20th century following the proportions of "modern" forms, with all their problems (excessive homogeneity of witdths above all), but choosing a sans-serif with good features for a more prolonged read (like, for example Metro, Quadraat Sans, Scala Sans or Sebastian could produce a decent readability while pairing them with a more "monolinear" Naskh modelled face, if I get it right…

Of course, it's hard to find good open source faces meeting those requirements, but you could look among the Linux Liberation fonts, SIL Gentium or similar efforts from open source teams…

EDIT: A great face to suggest Arabic designers to pair with "medium modulation" Arabic types could be Candara, which comes as default with Windows Vista. Cambria is also excellent, although a little too wide spaced between letters, but often it does not harm for onscreen apps.
Plus, they both have non-lining numerals, which makes for excellent typography paired with the usual "warmth" of Naskh calligraphic character.

Syndicate content Syndicate content