Javis on 4/9/2007 at 16:38
I posted a thread at (
http://www.ttlg.com/forums/showthread.php?t=116588) the Shocked section, but maybe it's not attracting much notice there. I and another are working on a Korean translation for System Shock 2, and I plan to use Telliamed's thieffon application to make a Korean font. The problem is that we are limited to only 230 or so glyphs of the 2000+ preformed glyphs usually included in Korean fonts. We may be able to manage with this number, but since our custom font won't be a continuous section of any codepage, formatting our script files might be problematic. Any ideas on this front?
Zygoptera on 4/9/2007 at 22:44
You can turn most of the text in logs etc into graphics, if that would help cut things down. You'd still need to convert for things like item descriptions but anything which uses an icon can be converted to a pcx, eg (resized to 640x, the full size version lacks the jaggies)
Inline Image:
http://i226.photobucket.com/albums/dd29/Zorapter/ogport.pngIt does require altering some of the string files beyond what a standard translation would, but it isn't particularly difficult. It does require the game to be played at over a specific resolution though.
Javis on 5/9/2007 at 03:22
Thanks for the reply. If I get your drift, you've replaced the string that follows
ResRepText1 from RESEARCH.STR:
Quote:
Summary: You've learned how to best target the Hybrid for maximal damage. All damage you deal to Hybrids will be increased by 25%.
Analysis: DNA structure indicates that this organism is a hybrid of a human host and a parasitic organism. Although the parasite (resembling a yard-long worm) has deteriorated to a stage beyond useful analysis, the effects of the process are evident. Severe deterioration of higher level mental processes is caused by tumerous growths along the spinal column and brain stem. The non-human tissues seem to be formed of a composite of small wormlike creatures that have adapted into the host body and taken over a majority of motor control and decision making functions. There also appears to be direct stimulation of the autonomic nervous and glandular systems. Hence, the organism produces exceptional amounts of both adrenaline and endorphins, making it remarkably strong and aggressive. What remains can not really be called human at all. The damage done to the host by the process is irrevocable, and the organism now functions with no sense of morality or hesitation.
Recommendation: The organism is vulnerable to the same forms of stimuli as a human. However, the changes in its physiology suggest a chemistry more complex than is currently understood. Further analysis of similar organisms might provide more insight in this area.
This would certainly make for less translation work! The cost would be a large volume of graphics production though, I guess. The change that would be made to the string files would be to replace certain strings with a reference to a graphics file?
As for the issue of the number of glyphs, let me clarify. The Korean writing system consists of 24 'letters' arranged in syllabic blocks. Each glyph consists of a combination of consonants and vowels in the pattern 'cvc' or 'cvcc'. Korean fonts include 2000+ of these pre-formed syllables which are phonetically possible. However, the reason why I said that we may be able to manage with only 233 or less characters available is that far less of these possible combinations are used in actual words. If we carried out the graphical conversion that you suggest, we would possibly decrease the needed number of characters in the font, but it would only occur through statistical attrition as we eliminated words from the textual script.
What I was wondering is if, once we have made a custom Korean font that covers all of our required characters, it would be possible to convert each character of our UTF-16 encoded script to the CP850 character space as they are mapped in the font.
I have already accounted for the problem of determining what characters are used in our script. Somebody has made for us an application that performs this task. Here is a sample of its output (cropped for space)
Code:
number of files : 1
number of scanned characters : 1991
number of unique characters : 478
characters:
਍
ഢ
൐
൥
൧
൬
൮
൲
൳
൴
൹
†
•
This output can then be rendered in a web browser and from that the bitmap that shows the character set can be constructed.
Zygoptera on 5/9/2007 at 05:43
If you can get by within the 233 character limit that may be more sensible. There are a couple of potential problems with using that graphical method which I hadn't thought of- there may be a limit on how many portraits there can be and I don't know how they are handled by SS2 in memory (I doubt either would be a problem, but if either happened it would be a major problem)
Yeah, basically I made a template to replace the log portrait/ research report graphics, the pink areas are transparent masks so the background shows through- they come to 60-80 kb each so they would add up to a fairly significant size, as you would need one for each log/ email. I'm assuming that whichever graphics program you have handles Korean script, and you could therefore type the translation directly into it. That would leave only the flavour text/ item descriptions needing font conversion and hopefully reduce the number of characters you'd need. Theoretically I guess you could even take a screenshot of the text typed into Word or equivalent if your graphics program wouldn't handle the script normally.
The only change that would be needed to be made to the strings file would be to change some of the portrait names. The english text would still be displayed at the side.
Of course, I don't speak Korean, so I don't really know if that would help things practically.
