Ergam CJK

Vasil Stanev

I am Learning Mandarin as I design the glyphs in FL6, found a very polite and nice Chinese guy at my home city (and in exchange he's gonna be tought my native language by me ). What I am interested in is smart glyphs the best way to save myself work, and how do I calibrate them for a CJK font.

I know there are some issues with the stroke width, but I will clear this up in the process. I'll let the design lead itself.

I do not know if a single font can support 84 thousand glyphs, but I will try to build up to 6000 and the rest will be released in pro versions depending on sales. That's my plan.

Hin-Tak Leung

I think the typeface is interesting and unusual - not like a native person with years of (forced by schools...) encumbered use of writing with the brush and the inherited esthetic values from teachers, etc. That's a refreshing thing. Well done.

Vasil Stanev

Hin-Tak Leung said:

I think the typeface is interesting and unusual - not like a native person with years of (forced by schools...) encumbered use of writing with the brush and the inherited esthetic values from teachers, etc. That's a refreshing thing. Well done.

What is the main manual for CJK type design? What is the "Elements of style" for CJK? is it available in English?

Hin-Tak Leung

Perhaps Ken Lunde is the person to ask.

I have been reading a book about the world's writing, and it is quite interesting to see how writing style is influenced by the tools/medium. An SE Asian script have circular strokes because it was traditionally written with a stylus on palm leaves; cuneiform was written that way because it was written in wet clay with a wedge etc.

Most of Chinese scripts were in brushes; but some are wood craftings or some carvings, and others are bronze casts. And the oracle (bone/ tortoise shell) scripts, are all influenced by the medium and evolved with a style suited to that medium.

Vasil Stanev

I downloaded the trial version (64 bit PC) of FontLab VI to continue creating this font.
Constant crashes. I am beginning to wonder if the 6th version is viable to my font needs.

Paul Miller

I don't know if this would be any use to you but there is an interesting book called 'Bi-Scriptual' by various authors. This book is not about type design. Instead it is about various different languages and how they relate to the latin script from a graphic design perspective. However this book is quite useful if you wish to design a typeface which has characters in one of these scripts. It provides context.  Some of the authors do talk about what is considered good and bad design of characters for their language. Each chapter covers a different language and is written by a graphic designer who is a native speaker of that language.

The languages covered are Arabic by Lara Captan & Kristian Sarkis, Cyrillic by Eugene Yukechev, Devanagari by Vaibhav Singh, Greek by Gerry Leonaidas, Hangul (Korean) by Jeongmin Kwon, Hanzi by Keith Tam, Hebrew by Lirion Levi Turkenich & Adi Stern and Kanji/Hiragana/Katakana (Chinese and Japanese) by Mariko Takagi.

Although this last book is not about type design it does provide much useful background information when designing characters for the languages covered.

Vasil Stanev

Oh, this is an excellent find! 😊👌 I couldn't find a place for it, my home is filled with books to the brim. However, if somebody could scan this book and send me a PDF, I would be more than happy to pay for the courtesy 🍻

Paul Miller

That would be an awful lot of scanning, it's a big book.  Also it is copyright and I don't have the right to copy.

My home is also full of books but I found room for it!

Craig Eliason

If you're happy to pay, why not buy the book and scan it yourself?

Vasil Stanev

Craig Eliason said:

If you're happy to pay, why not buy the book and scan it yourself?

Besause our delivery system is more primitive and it can get lost in transfer, and it takes a lot of red tape to refund the payment. Or at least this is my information from about 4-5 years ago, but I still would rather not risk it.
(Sorry if this is getting more and more offtopic. My question was more about FontLab VI.)

Paul Miller

Vasil Stanev said:

Besause our delivery system is more primitive and it can get lost in transfer, and it takes a lot of red tape to refund the payment.

What country are you in ?

Vasil Stanev

Bulgaria (in the EU)

Rafael Saraiva

do you believe in reincarnation?

western guy learning mandarin == 1 lifetime

western guy learning chinese typeface design == +1 lifetime

and you're planning to sell your font already...

Vasil Stanev

Guys, please keep the topic free from spam. 

Simon Cozens

There isn't really a manual for CJK type design - or indeed for most craft, which is normally learnt first-hand from another person.

In case it's worth knowing, even with smart components and so on, most kanji fonts are designed by large teams over a period of years. It's not something that a single designer with no experience can easily take on; confidence and optimism doesn't really make up for that. Hangul, maybe. Kanji, not so much.

Vasil Stanev

Simon Cozens said:

There isn't really a manual for CJK type design - or indeed for most craft, which is normally learnt first-hand from another person.

In case it's worth knowing, even with smart components and so on, most kanji fonts are designed by large teams over a period of years. It's not something that a single designer with no experience can easily take on; confidence and optimism doesn't really make up for that. Hangul, maybe. Kanji, not so much.

Pardon the directness, but how can you be sure the team knows what they're doing? In my (bitter) experience, on most teams, degrees or not, 20% of the staff do 80% of the work, and the other 80% of the staff may even be creating more work than they get done. It is fully possible to have 20 years of experience in something and having spent 20 years doing it the wrong way. I've looked at some  (not all) fonts under the microscope, CJK or not, what I see is pretty far from what it should be, just looking at the Beziers. I hope I am not rubbing somebody the wrong way but those are the facts.

It may sound funny, but Ergam CJK is something I'm doing in my free time because I like being productive. It may never be released, but it's a great tool for myself to get to know the script. That's why I would like FontLab 6 to not crash and was asking in effect if only I had that experience with the demo. Considering the polite help above from Mr. Miller, I remembered I had seen much of the material on Youtube - Granshan 2015 - so I'm actually familiar with most of it. In general, I believe it's best to leave to local professionals to do the version of a given font in their native script, and this CJK is the one time I break this rule.

Khaled Hosny

just looking at the Beziers

That is a (modern?) type designers obsession, what matters is how the text looks when rendered at the intended size(s) and on the intended medium(s).

Vasil Stanev

Khaled Hosny said:

just looking at the Beziers

That is a (modern?) type designers obsession, what matters is how the text looks when rendered at the intended size(s) and on the intended medium(s).

I agree. The mistakes I saw in some faces were more serious and caused certain glyphs to crash alltogther. For example, in one font, the zero was drawn with the inner ovals having only two cardinal points each. Some faces had no east or west cardinal points on the outer ovals, causing issues with the decimal point when spaced. I remember seing an otherwise pro serif font, in which each serif ended in a bowl with this issue. In some the lowercase only looked to be the same as the uppercase, but certain elements were off by 10-20 points. And there were many fonts, I would say even the majority, that did not understand where the ogonek connects and how it should be shaped, with bad coding for Scommaaccent or Tcedilla, and the list goes on. It is like cinema to me, on a different time scale - what production value was passable 30 years ago would be laughable now.

Paul Miller

Vasil Stanev said:

Pardon the directness, but how can you be sure the team knows what they're doing? In my (bitter) experience, on most teams, degrees or not, 20% of the staff do 80% of the work, and the other 80% of the staff may even be creating more work than they get done. It is fully possible to have 20 years of experience in something and having spent 20 years doing it the wrong way. I've looked at some  (not all) fonts under the microscope, CJK or not, what I see is pretty far from what it should be, just looking at the Beziers. I hope I am not rubbing somebody the wrong way but those are the facts.

In my experience when designing glyphs for a language you are not familiar with and not a speaker of it is easy to make what you think is an innocuous design decision which may look extremely bad to speakers who are familiar with the language.

CJK would be an extremely big project to take on, even doing all the Latin blocks plus Cyrillic, Greek, Armenian and Georgian would be an enormous project, scale that up ten times and that is what you are looking at with CJK.

Good Luck !!!

Ray Larabie

I don't doubt that a person with little knowledge of kanji could draw a CJK typeface provided they had a sufficient team of native Chinese and Japanese readers proofing the output. While a large number of character are radicals in blocks that can be arranged on a grid, there are a lot that are not. Radicals get smashed together to the point where they're barely recognizable. The way they merge together has a threshold that's very hard for non-Japanese/Chinese readers to ascertain. Sure, you can see a lot of outlandish Chinese and Japanese typefaces that push the boundaries of legibility but only a native reader will know which characters might be misconstrued. I've seen this effect in low resolution Japanese pixel fonts. I've seen pixel glyphs or characters where the lengths of the horizontal strokes in high resolution can be distinguished but appear to be even length in some low-resolution radicals and exaggerated in others by adding/removing pixels to avoid ambiguity with another character.

The abstraction of radicals differs in Chinese and Japanese. The way radicals squoosh together varies between Chinese and Japanese. My wife will point out the mistaken use of a Chinese font for a Japanese word on a sign. Sometimes I can see it...like all those X shapes they use more often in Chinese but sometimes I can't see the difference. Even though they're the correct radicals, the way they're squished is wrong.

Your proofing team would need to be diverse. While some readers may be able to identify a large number of characters, they may not be as familiar with some. The importance of specific characters can vary by where the reader resides. I watched a Japanese comedy show on television a few months ago that showed how the names of many subway stops in Japanese cities can't be read by people from other cities. In my city, there are characters that are very familiar to locals that would likely be unknown to most readers in other cities. Even Google Translate can't consistently translate the name of my local subway stop. Main point is: you would need a sufficient team of proofers. If you could receive enough, quality feedback and you have enough time, you could probably do this. But I can't see it taking much less than US$1 mil to hire a team of qualified Chinese and Japanese proofers to work for 4 years or so.

Thomas Phinney

Vasil Stanev said:

I do not know if a single font can support 84 thousand glyphs, but I will try to build up to 6000 and the rest will be released in pro versions depending on sales. That's my plan.

No, current font formats are limited to 64K glyphs. 

But trying to hit that many as an individual would be more than a little challenging. Chinese and Japanese fonts are usually developed by teams, because of the huge amount of work involved, and the complexity of the task.

Although feedback is great for almost any typeface design, I think that the larger the font project, the more it benefits from an extra pair (or several pairs) of eyes on it, during development.

Vasil Stanev

I have cancelled this project permanently. But this thread will remain, I hope, as a list of valuable advice to future designers who are thinking of doing a CJK font alone.