Asian font with non Unicode characters

[Lanimoomoo]

I have a manuscript that uses old Korean characters that are non Unicode.  These characters appear to be combinations of Unicode characters within the font ,so when I place the Word file into Indesign I end up with not 1 non Unicode character but 3 Unicode characters. If I place the Word file and strip out all the formatting while importing I get the usual pink boxes, but when I highlight and apply the Korean font the 1 non Unicode character will display correctly.  Unfortunately, I will also lose all style formatting in the file using this method. If I copy the character into FInd/Replace it becomes the 3 characters and Indesign cannot replace it with the correct non Unicode character. I also tried copy/paste the character into the Glyph find/replace and it does not show up for find, but will for replace when I type in the GID number. Does anyone know of a way I could place the Word file without the character splitting into it's 3 parts? The character in the font I've been using as a test is GID 28348, the font name in Indesign is NanumBarunGothic YetHangul.  Or does anyone have any other ideas?  Thanks in advance for any help.

EDIT [link removed by moderator]

[Joel Cherney]

I have other ideas. Are you sure that they are not Unicode? I looked up the NanumBarunGothic fonts, they seem to be up on Github with very recent activity (2-3 years). I did not download them, but I wonder if you are maybe using the wrong font format. Can you tell us some more details?

1) Are you using a Mac or a Windows machine?

2) What is the file extension of the font? Is it .ttf, .eot, or .woff2?

3) If you open the Word file in Word, are the characters composed correctly?

Also, some terminology to help you out: the three individual "letters" are called jamo. When they are composed correctly, they form up into a "block" or a "morphemic block." (Dang, I wish I knew what that was actually called in Korean.) Assuming that you are using a contemporary version of Creative Cloud, it may be worth it to try turning on the World-Ready Composer (find it in the Type menu) to see if the WRC can compose those jamo into a block.

Lastly, if you can post a few glyphs in Word format to something like Dropbox, I would be happy to download the fonts and the file and see if I can get it to work over here.

[Lanimoomoo]

Hi Joel,

I've got Windows 8 with CC2016, the font is a ttf (sorry, I put the link for the font in my original post but it got edited out), and if I open the Word file it looks fine. I put a sample Word file with the glyphs highlighted and the font in my Google Drive, Old Korean - Google Drive . The folder is shared with anyone with the link, view only rights.  Please let me know if you have any problems accessing the files.

Thanks for any help you can give me.

Wanda

[Joel Cherney]

I got it to render more-or-less correctly using the J Composer, which is a feature of InDesign that is localized in Japanese, Chinese, and of course Korean. I used the handy Japanese template from TransPacific Digital to test with the J composer.

It's still not displaying the way it does for me in Word or LibreOffice, but it's closer. I expect that, if you read Korean, you could simply download the Korean localization of InDesign CC2017, and it would (probably) work there. The UI of Korean InDesign is only in Korean (with no English UI option as there is in the Middle East edition) so I can't really use it at all.

[Joel Cherney]

Whoops, forgot the link to the J template:

TransPacificDigital: Downloads

[Lanimoomoo]

Thanks Joel, unfortunately the characters have to be the same as in the Word document. I figured out a workaround for glyphs that are in the font: I put <i> and </i> around the italics for this test in Word, then imported the document with no formatting and styles. In Indesign I used Peter Kahrel's wonderful missing glyphs script to apply the Nanum font to the Korean, and GREP to replace the italic.

Now I just have to figure out how to somehow display the correct character for the ones that are 3 jamo into 1 block that has no glyph available in the font. World Ready composer actually deconstructed a block into the individual 3 jamo when I tried to use it, when I have it off I get 2 jamo in a block and then the extra jamo like in the image you uploaded.  I don't know Korean so installing the Korean version of Indesign would not really help me.  I very much appreciate your help, and if you or anyone else has any ideas please let me know.

[Joel Cherney]

There are a bunch of ways to do it. This strikes me as the easiest:

1) Download the Japanese template that I linked to above

2) Open up that template and paste without formatting into the Japanese text

3) Apply the Nanum font to your Korean text

4) Check to see which composer is enabled. It should be the Japanese Paragraph Composer. (Check it by looking at Type -> Justification, or any one of three or four other ways)

5) Make a new paragraph style, that includes Nanum font and the Japanese composer

6) Mark some text with that style, copy it, then paste it into your document with formatting

There are a bunch of variations on this. The basic idea is that you have to get the Japanese composer into your document, and there is no UI method to get it in there without some kind of workaround. I don't remember the precise Javascript syntax, and I'm on my phone, or I'd post a script which would be much easier to use than steps 1-6 above.

[Lanimoomoo]

I need to have the character highlighted in yellow (this is from Word):

When I use World Ready composer I get this:

When I turn off World Ready composer or use the method described above for J composer I get this:

It appears that if the glyph combined from 3 jamo is within the font I can get the character with the workaround I used in a previous post, but if the glyph is not in the font the most Indesign will let me do is combine 2 jamo (the first character is a glyph in the font). Unless I am doing something wrong in J composer.

If there is anything else I can try please let me know.

[Joel Cherney]

I don't think that you are doing anything wrong, to be honest.  Are you maybe working on archaic text, here?

For my own part, I am out of suggestions, but I will ask around on your behalf - I have a few Korean linguistics grad students in my contacts list, perhaps I can get some more information and find a solution for you. Cross your fingers.

[Lanimoomoo]

Yes, it is a manuscript about "old" Korean.

The best I could do was input the 3 unicodes of the individual characters and kern them into position (character in green):

It's just so frustrating since I can insert the 3 unicodes in Word and the correct character will appear.

I will post in the Type and Typography forum, maybe someone there has some ideas.  Thanks for all your help.

[David W. Goodrich]

I deal with a fair number of non-Unicode CJK chars., almost always archaic Chinese forms, and the simplest solution usually is to make custom chars.  Fonts are very complex, though you can try Indyfont within InDesign instead of a font editor.  Another route I've used is to insert an in-line vector graphic; drawing the individual components freehand is not easy, but there are open-source and copyright-free fonts from which one can "borrow" parts.

The basic issue here is how NanumBarunGothic creates the combinations of elements you need, and I fear they may use custom software.  I don't read Korean, but when I tried using Win7's Hangul keyboard I could not make your combination, suggesting something else may be going on.  I'm not about to find out by running their installation *.exe (NanumFontSetup_TTF_BARUNGOTHIC_hangeulcamp.exe) lest it affect other things on my computer.  I also note that Bugzilla@Mozilla suggests, if you read to the end, that the Korean version of Source Han Sans may prove more useful for Mozilla's purposes precisely because it does not use custom encoding. 

Source Han Sans is fully Unicode, and appears not to include the "am" glyph in your first example.  As I say I don't read Korean, but it looks as though it would fall in the area around U+C556.  Glyphwiki can give you an SVG, which you could modify with Adobe Illustrator or Inkscape to adjust the "box" at the lower left.

Good luck!

David

[Lanimoomoo]

I've made fonts to create non unicode (and missing unicode) characters before, but the publisher I work with does not want me to do this any more. I'll check out Indyfont.

The publisher I work with also does not want inline graphics.

Microsoft has an IME called "Old Hangul" which I installed on my PC but it seems to only work with Word, not the CC suite.

Have modified Source Han Sans to create characters I needed in the past, did not know there was a Korean version of it so I'll check that out also.

Thanks for your help David.

[Joel Cherney]

I  suspect that the problem here is that we English-speakers do not know if there is any composer in InDesign that has support for the Korean-specific OpenType glyph substitution feature "trailing jamo forms" (tjmo), or maybe one of the other Korean-specific GSUB features. Does Korean InDesign have its own composer?  I suspect it does not, and that getting this to work without attention from developers (and without any of the already-discussed workarounds) may not be possible. However, I am not done digging.

[David W. Goodrich]

I hope Joel can dig something up.  I thought that folks with Creative Cloud subscriptions could run East Asian versions with an English interface after careful (re-)installation -- perhaps this doesn't work for Korean.  I would hate to be stuck using a font that only works in MS Word: I've twice set books with Word, a decade apart, and both times swore never again (the less ancient was for a job requiring furigana back when FrameMaker offered that feature but only with a fully Japanese OS).

Over the years I've run many odd-ball characters via custom fonts, but for electronic documents custom encoding confuses searching and indexing: pictures don't have codes.  One of these days, ID is supposed to get SVG back, and SVGs can carry XMP metadata to identify the characters they represent (of course, SVG support is already built into HTML5 and Epub3).  If PDF 2.0 extends object-level metadata to vector graphics, well, gee, that might work there as well.  But we're not there yet.

Good luck,

David

[David W. Goodrich]

I had another thought overnight.  Might this be a case of an Opentype feature that depends on the "language" setting?  ID became more strict about that, back around IDCS5.5.  The "Korean" language attribute doesn't come with western-language versions of ID, but they will pick it up from word-processor files.  Applying "Korean" this way didn't help my IDCS6 sample compose non-Unicode chars., but perhaps old Korean forms need a particular language attribute (along the lines of "Portugese: Orthographic Agreement", or "German: 1996 Reform").

That is pretty much a shot in the dark, and a long one at that: the Japanese in Source Sans Han left out one or more Opentype features for historical text that are available in the Kozuka fonts bundled with ID.

David