Hi, I’m wondering if someone can share a best approach or any tips for adapting a small number of glyphs from one font so that they conform to the style/appearance of another font.
Some more background: I’m working on a very simple project, adding glyphs for Japanese radical variants which aren’t represented in Unicode (either in 2F00 to 2FDF or U+2E80 to U+2EFF), as well as glyphs which are represented, but not in the desired font. For both cases I can draw either on existing glyphs as models because these share the same appearance but aren’t encoded as radicals or take an existing glyph which incorporates the radical as a component and strip away the parts I don’t need. In other words, I’m never starting from scratch. I can use all or parts of an existing glyphs as a mode. The problem however is that these glyphs are styled differently from the font into which I’d like to assimilate them.
Essentially, my question is, how to best modify an existing glyph from font A so that it looks like it belongs to font B?
Specifically, the font (style) I am trying to emulate is that used by Mplusmplus-lm-light. I have a handfull of glyphs to complete, but two examples of radicals with existing Unicode are:
Any good strategies or tips? What can (?) be done via filters or scripts, what best done by hand? That sort of thing. I realize that there may be no alternative to good old manual, visual comparison. But that’s also good to know!
My ultimate goal is to create a freeware font, in Mplus style, consisting of all Japanese radicals and all radical variants. To my knowledge, no such font currently exists (I’d love to be wrong!). Here’s my work so far (Glyphs project).
First, I would contact the creator of the font and make sure it is okay to do that.
Technically, Your best dice is probably to copy, paste & recombine existing elements from the glyphs you already have.
You could also draw the skeleton paths use the Offset Paths filter: Measure the horizontal and vertical strokes with the measurement tool (L), use half the values for the H & V offset in the filter.
M+ isn’t open source (I haven’t seen the source file anywhere), though it doesn’t restrict modification.
Anyway, the best approach would be the way mekkablue described. For making CJK, it is good to make a lot of graphical parts so that you can quickly build the letters; those parts don’t need to be tied with Unicode. Get the stroke somehow, and put the parts (like you place serif components).
Please look further down the page, and you’ll see a link to the M+ source files on sourceforge.
When you download the tarball for the current snapshot (http://sourceforge.jp/cvs/view/mplus-fonts/mplus_outline_fonts/?view=tar) or check out these files via CVS the unpacked directory contains the Illustrator and EPS files, makefile template and scripts necessary to build the TTFs with fontforge.
Ah, found it. I know some friends who can develop a typeface this way, which is fascinating. I’d never be able to work this way, though.
Anyway, since the AI file is available and it’s practically a 2-master design (five intermediate weights are only blended), it’s theoretically very straightforward to set up a multiple master Glyphs file to get a skeleton.
Ironically, now that Source Han Sans has been released (which I just discovered today) I need to start all over again since its support for radical glyphs looks to be better than M+ (on par with Meiryo or Hiragino). But it’s all good. I should just need to add the radicals to the font which don’t have unicode encodings and I’ll be done.