Font for KanjiVG?
Torrential
Reading back through this forum, it seems that the kanjivg vectors
were derived from a Japanese schoolbook font. Does that mean that they
don't actually correspond to any current digital font? I'm not
interested in the subtleties of angle and so on, but the major
topological features and broad agreements about which way diagonals
go. I have been looking for a font that matches the kanjivg characters
in terms of such variations as whether the 'sea' kanji has a central
vertical in the right element, resembling 'every', or two dots,
resembling 'mother', and whether the rain kanji has the dots in an X
pattern or parallel, and so on. I'm yet to find a match for the
characters I'm considering, though some fonts come very close.
These characters seem useful for comparing the fonts:
(雨, 海, 才, 紙, 次, 荷)
Has anyone else found a match - a font that would draw all of the
above characters the same way as the kanjivg vectors?
Cheers,
Craig.
Dr. Ulrich Apel
> [ . . . ] it seems that the kanjivg vectors
> were derived from a Japanese schoolbook font. Does that mean that they
> don't actually correspond to any current digital font?
Japanese Schoolbook fonts are also called Kyokasho fonts. As model fonts for KanjiVG, DFKyoKaSho-W3-MP-RKSJ-H by DynaLab and A-OTF Kyoukasho ICA Pro L by Morisawa and Company Limited have been used. The Morisawa font was used for example, when the DynaLab font was lacking characters.
Japanese Schoolbook font forms are some sort of mixture of Kaisho and Mincho characters. If for example a stroke is slanted in Kaisho and vertical in Mincho, a Schoolbook font will use the vertical stroke. KanjiVG also contains normal Kaisho versions of characters which differ in Schoolbook fonts.
To my knowledge, Mincho fonts have much less variations of character forms than older historic writings. Probably their Chinese versions, the Ming fonts, may have more variations too, as in your examples.
Without doubt it would be worthwhile to do more research in character forms, their usage and a better abstract description. I am planning to do so, as soon as I have time.
By the way, our lab for Japanese studies at Tübingen now has a new second assistant for our computing needs. She is Chinese, and probably, she will be of help to adapt KanjiVG for Chinese character forms, too.
We also had a Japanese intern for four weeks and she prepared several hundred new characters for KanjiVG. Element description is still missing, but we hope to be able to include these characters in KanjiVG soon.
Best wishes
Ulrich
Torrential
Thanks for the information, and thanks indeed for the whole project.
I have found the Mincho fonts to be a reasonably close match. So far,
in my own limited kanji-learning efforts, I've only come across 才 as a
character where kanjivg and fonts like MS Mincho go separate ways -
kanjivg does not have the diagonal crossing the vertical, but MS
Mincho and other Mincho variants I've seen do have it crossing. There
are likely to be other characters that differ but I have not done a
systematic survey (if anyone has a list of non-matching kanji, that
would be great).
I'll try to chase up the fonts you mentioned but for my purposes it
would be nice to use a readily available font. If there are only a
couple of characters that do not follow the MS Mincho pattern, I may
just change those vectors to match (for my own copy of the vectors).
It might even make sense to change the shared, official version for
characters like 才, just for consistency, but that's obviously up to
you.
BTW, I looked up one of the articles you referred to previously,
ftp://reports.stanford.edu/pub/cstr/reports/cs/tr/83/974/CS-TR-83-974.pdf
, and found it quite interesting. I had thought about a similar scheme
myself, but hadn't taken it any further. It would be nice to produce a
more complex, realistic animation from the kanjivg vectors, and it
should be possible to produce an actual kanjivg font family based on
those ideas, which would solve the font issue perfectly. I could
probably do it myself but don't have a lot of spare time at the
moment, given that programing is just a hobby.
Thanks again,
Craig.