Kingdom Hearts: Final Mix Translation Patch (In progress)

Just map L2 and R2 to the right stick in pcsx2 with pcsx2 bonus (it allows you to have different configurations for every game that apply automatically)
http://pcsx2.net/download/viewcategory/43-frontends.html

 

Awesome, now the real translating side of things can happen :).

Found a bug in your ARDread program, though. As a test, I used dh01.ard (which contains text from the first cut scene in the game, including "So much to do," and "so little time..." in the English version). It extracts 3 text files containing 4, 0, and 99 strings each. For the second one (0 strings), it gives a red error instructing me to submit the name ;)

This was the output when using the Final Mix version ARD, but the same error occurs with the US version at different offsets. It didn't extract the area of the file that contained "So much to do," in the US version (which is in the section beginning at 0x3ED80 in dh01.ard, the 7th offset in the initial header). I belive this bad entry resulted in the <1KB file named "dh01.ard-1.txt"

How are you determining if an ARD section contains strings? I noticed when I try to extract strings from another ARD (di01.ard, containing first Destiny Islands cut scene), the cut scene strings are again found in the 7th section of the ard (err, pointed to by the 7th offset in the begging of the file), but this time your tool successfully extracts the strings to di01.ard-1.txt.

I think the cutscene text from dh01.ard is supposed to be extracted to dh01.ard-1.txt, but your error catching routine discards it as bad data. Oddly, the part that contains the strings in dh01 does in fact end with the "Obtained Postcard." line.

 

Well ****, seems things aren't as hard-coded as I thought. I was literally hard-coding entry numbers (like entry #6; 0-based), as it seemed set-in-stone (something like 10 files matched). Apparently it's not... Wonderful. I must of picked test files that happened to use the same layout... -.-.

That being the case, I again am not sure what to try. The game has to know somehow what everything is...

Also: This invalidates my current ARD tool, don't use it for anything that's not testing.

 

From what I've seen it goes something like this:

Let x = 7th section offset in ARD file's header.
Goto offset x.
Let y = 2nd offset from section beginning at x.
Goto beginning of ARD file.

So x+y would give you an offset to a table of strings. This seems to be the pattern in several files, but I might be wrong.

 

dh01.ard from the US version, the table shows up different.
x = main_header[8] //9th file; x = 257408; this seems to normally be the 7th file
goto x
y = this_header[1] //2nd file; y = 1536
goto x+y
z = this_header[2] //3rd file; z = 276
goto x+y+z
//KMB starts here; off = 259220

The fact is uses file 9, not 7, is what messed my extractor up.
Also, a heads up about my currently-bugged extractor: If you see the .txt file show up, that is really what it read; the red warning doesn't touch the output. ;)

Attached the 2 ARDs I'm using, tag files, and the program that can read the tag files. I think what I labeled it right.

 

hey . did u translate anything eles ?

 

Geez Sora. Lol, stay patient.

 

He/me/us/them/everyone is currently tring to figure out how strings are stored in the ARD files. I don't think a consistent pattern has been found yet.

 

He may start giving us a percentage when he can. So he may update what he has when he wants or when necessary. It gonna take a good while. It won't be done before HD remix comes out I can say that right now. So just sit back for a bit. ;)

 

That's all right. It just means that the patch will be localization accurate

 

Yeah, there was always meant to be an accuracy update anyway.

 

Hey, this project seems pretty amazing. It doesn't matter if 1.5 is coming out with 1FM, I just think it would be awesome to play this on a ps2 and see it get done! I'd be glad to be a beta tester for this or something, (I'm not much help coding :/) but hopefully you could use me at some point!

 

any good progress since the beta 2 trailer that u put on youtube for us ?

 

Hey, I'm just making the tools; I know no Japanese. The journal parts were easy, everything lined up with the US version. Subtitles I can already see don't line up as close, and Google doesn't help much. :p
---
ARD progress: There's no way to tell which entries are strings or not, afaik. My guess is another entry somewhere else tells the game "entry 31 here is the text to use".
So.... I changed my extractor to just attempt on every entry. This means it will extract a lot of garbage, but then the human translator could decide what's what. There's also a command line argument to tell it to extract only certain entries, for when you know what it has. I'll consider this done until another bug\problem shows up.
And that leaves the repacker. It sorta works right now, in that it can change the text. But only if it's the exact same size as the original, otherwise the game starts doing some funky stuff. I'm probably just forgetting to update some number somewhere...

Edit: Might have the repacker working good. Doing some final test-runs, of whether the game takes them or not; assuming it does, I'll release a new version.
[DOUBLEPOST=1373329081][/DOUBLEPOST]
Extractor instructions:

Spoiler

Drag an ARD onto it to extract. Multiple txt files will be made containing the string table (assuming all is right, 1 string per line).
Warning: Some text files WILL be trash. I can't figure out how to find the string tables only, so I try to extract everything. Delete the trash txt files, you don't need them.
The text files have a header in them, the part like this "OFFSET=259220;LENGTH=7088;CHARS=7073". DO NOT TOUCH THAT EVER. That's used for recreating the ARD, and nothing you do will change those numbers.
And finally, the command list: "ARDread ARD_to_extract [-l 1,2,x]". The -l (L) argument tells it to only try those sections, instead of all 32. So "ARDread ew36.ard -l 6,18,31" would extract only from sections 6, 18, and 31 from ew36.ard (those are the only 3 text sections).

Insertor instructions:

Spoiler

Drag the original ARD onto it to make a new one. The new ARD will be named "<orig_name_here>.new".
This program will try to read the 32 text files (default names), and only replace those it finds. If the new size is different enough from the original (~16 characters), it will also update the header tree for all files after this one.
Tested with all strings consisting of "00" (smallest possible size) and using English equivalents (larger then original), and both seemed to work fine.

(Funny... the insertor is so much easier to use :p)

And

I'm going to take a crack at menu.ard now; others can give it a go as well. ;)

Edit: I'm dumb, menu.ard has no strings. lol

 

awesome job Man you're very good programmer, congratulation

 

Good job crazycat. Maybe now Sofa can stop asking questions and help contribute to the patch he wants :p

 

So why is that text misaligned and seemingly random? The tables not matching?

 

Well what ever you are doing. You know what Im saying.

 

Pretty much. The FM table is actually longer then the US one too, it has more entries (at least in the file I was testing).

:p
Edit: ---
Can someone translate these for me? I'd rather a human translate, then crappy Google.
子犬を助けたお礼にプレゼントをもらった!
ミを手にいれた
を手にいれた
手にいれた
ミのアビリティを覚えた
ミのパーティアビリティを覚えた

 

Ok so I'm using your tools (amazing, by the way) to translate the very first cutscene where sora wakes up from his dream and starts talking to kairi on the island. Translated the first 15 lines of dialogue and I now want to pack it all together to see what I've done in game. But when using the ARD inserter I get this:

Code:

ARDedit di01.ard
 
Start
32 files in archive!
l:18103
pad:5
l:0
pad:12
l:18103
pad:5
l:0
pad:12
l:10417
pad:11
l:0
pad:12
l:0
pad:12
l:18103
pad:5
l:0
pad:12
l:4691
pad:9
Updating #6...
wp:1724568
need to update size!
Updating #7...
wp:1839508
need to update size!
Updating #8...
 
Unhandled Exception: System.ArgumentOutOfRangeException: Non-negative number required.
Parameter name: count
  at System.IO.BinaryReader.ReadBytes(Int32 count)
  at JScript 0.main(Object this, VsaEngine vsa Engine)
  at JScript 0.Global Code()
  at JScript Main.Main(String[] )