Japanese text encoding standard for MSX?

Page 3/3
1 | 2 |

By Takamichi

Supporter (12)

Takamichi's picture

13-03-2018, 15:54

MSX Kana Filter is an application that converts MSX 1 byte hiraganas to 2 byte ones.
http://tatsu.life.coocan.jp/MySoft/index.html#MSX
I can translate the enclosed manual but probably it's unnecessary because most people subscribing here seem to know more Japanese than me :)

And yes, MSX Kanji BASIC cannot handle 1 byte hiraganas so CALL AKCNV, the command that converts 1 byte characters to two byte ones, doesn't work with 1 byte hiraganas.

About the character table in Izumic Ballade, I can guess most of these characters are. Is there someone who wants me to write them down? And as is obvious, there are nonstandard characters like a finger (column 12 row 1), "Bs" and "Cl" written as single characters (column 0 and 1, row 3).

By sd_snatcher

Prophet (2880)

sd_snatcher's picture

14-03-2018, 19:09

Thank you for your tips, Takamichi! And welcome back to the forums. It has been some years since we last saw you around here.

Do you know if the MSX 1-byte hiraganas followed some formally defined standard, or was it custom tailored for this architecture?

By eimaster

Master (218)

eimaster's picture

15-03-2018, 01:30

The following site provides some documents and tools used for translating games roms. There are lots of books and documents about Japanese character set which might be knowledgable and of use to you.
http://www.romhacking.net/?page=documents&startpage=1

By wyrdwad

Paladin (678)

wyrdwad's picture

15-03-2018, 07:13

Yes, thank you for that very detailed explanation! Regarding Izumic Ballade, it turns out the "encoding" being used is just straight ASCII, with a custom character set loaded into memory via an image file. I've been able to account for this in the MSX-BASIC source code via a substitution table, so that particular hurdle has been passed and the game's translation is very much underway:

https://www.youtube.com/watch?v=iSnNQ3raGq8

(Note that this video uses a completely different font from a completely different image file, but like the small version, it's still more or less just straight ASCII, and I was able to work with it via a substitution table and a little bit of graphic editing support from MP83.)

It's interesting to read such a detailed account of Japanese encoding standards -- I knew they were a bit of a mess, but I didn't realize just how much of one! I work in video game localization at XSEED Games, and I'd say about 75% of the modern Japanese games we work on have their text encoded via Shift-JIS, which I greatly appreciate since Shift-JIS is very easy to parse using simple programming -- any routines you write that expect standard CSV formatting with quotation marks around text fields and commas as field delineators will work perfectly on Shift-JIS text more or less 100% of the time.

When we have games that use UTF-8 encoding, however (which basically accounts for the remaining 25% of instances), simple text parsing flies out the window, since you can't really trust bytes read as quotation marks or line breaks anymore.

Fortunately, the need to parse text manually is somewhat rare, as the Japanese developers typically do that for us -- but on those few occasions when they don't, it's proven to be a headache and a half for sure.

That's why personally, I'm a big fan of Shift-JIS. It's nowhere near as dynamic or flexible as UTF-8, but when you need to do simple text parsing, it's the only way to go!

-Tom

By Grauw

Enlighted (7549)

Grauw's picture

15-03-2018, 09:10

XSEED, that’s a neat place to work Smile.

By Takamichi

Supporter (12)

Takamichi's picture

15-03-2018, 13:33

Yes I have been away for a long time, though haven't forgotten MSX.

Several Japanese microcomputers supported 1 byte hiraganas as part of ASCII codes. As far as I know, Basic Master Level 3, PC6001 and PC8801 could display them, but they were not standardized. One easy way to tell whether a microcopmputer could display hiragana is to look at the keyboard; if hiraganas are written, it can display Smile
The hiragana codes always occupied the unused parts, 80H-9FH and E0H-FFH like language specific characters do in here. They conflict with shift JIS system which always consider these codes as a part of full width (2 bytes) character and never as an independent half width (1 byte) character.

By Takamichi

Supporter (12)

Takamichi's picture

15-03-2018, 13:44

On the other hand, when it came to 2 byte characters 80s microcomputers used universally same shift JIS codes. http://www.kiwi-us.com/~ohta/pc88/kanji/other.htm Modern shift JIS isn't standardized but seemingly it was then, at least from these small samples tell.

By sd_snatcher

Prophet (2880)

sd_snatcher's picture

17-03-2018, 18:35

I noticed that the Wikipedia article about the MSX character set doesn't have the Japanese character set. Only the international and Brazilian ones. Any knowledgeable candidates to fix that article? :)

Page 3/3
1 | 2 |
My MSX profile