ROM Hack witch's wish undubing

  • Thread starter Thread starter shineek
  • Start date Start date
  • Views Views 9,762
  • Replies Replies 23

shineek

Well-Known Member
Member
Joined
Jul 14, 2007
Messages
256
Reaction score
2
Trophies
1
Website
Visit site
XP
202
Country
Poland
can anyone help me with undubing this game?? or rather rescripting it , cause it seems that [someone lazy] cut out the voice acting again to a minimum , and i don't know if adding them back in the us version would work like in bleach 3 . i appreciate any help .
 
yea,, I think it would be best to put the English text into JPN rom since the US rom missing the voices. But then, when I did that, only the 1st letter of English text displayed..
 
well yeah , you probably forgot to change the font too , it worked for me altough i probably didnt find all files with the texts , because some stuff like item descriptions where in japanese
 
if someone is going to work on this..

kindly change the names too (by just romanizing the names).. cause they were totally changed on the english version.. xD
 
changing the names is quite a lot of work.
names aren't "variables" linked to in the files, they are written out.
so to change the name of a character you have to go through ALL script files and change it for each and every sentence this character says.
 
dont care bout the names , i just want to hear those VO from the japanese version , and it seems they only cut out the magic chants and some other stuff , probably because they found it to embarrassing to dub
 
@Shineek
By changing the font.. do you mean just replacing the jpn font w/ the eng one or actually edit the font?
Sory, pretty noob at this.
 
This is really complicated to undub. Several people have tried but none are able to do it. I myself have also given up but goodluck trying.
EDIT: Also I was thinking of translating the script from the US version to the JAP version. But I thought it will take too long as I'm not sure how long the actual game is.
ohnoes.png
 
i currently testing it , ive rescripted it and os of now everything works , and is in english , well not everything , the start menu is in japanese cause it was graphic based . ill try to make a patch later and upload it so you dont have to download both the JP and US versions .

and still nothing , the first two dialogues boxes are normal , than i get only one letter per dialogue box and nothing more . can anyone help us out here ??
 
yeah , but it seems simple script and font exchange from the US version to JP , dosent work , why does it show only the first letter and not the rest of the text??

come one anyone throw me some meat here , i cant do this alone my hacking skills arent good at all . why are the script files showing only one letter ??
 
my best bet, the japanese version reads hex 00 as end of sentence.
the US version has 00 between all letters.(ordinary space)
the japanese version however has 83 (between letters in a name, appears like space) and 82(between letters in a sentence, appears like ,)
you could perhaps try replacing any space with 83/82 manually, but that would take ages.

note: these "spaces" (00, 82, 83) don't appear in the game.
 
Well I tried doing a small test to see where putting hex 83 in between each letter for vickys name heres what I ended up with. For English title copy the data\2d\title from the English rom, I copied the entire 2d folder since Im assuming that there are other graphics based files like the instructions that pop up.

Code:
Hex String in original script
03 00 00 00 56 00 69 00 63 00 6B 00 79 00 00 00 05 00 00 00
Hex String after modification
03 00 00 00 56 83 69 83 63 83 6B 83 79 00 00 00 05 00 00 00
01.jpg


Edit: Did more fiddling with the Game file and I got this. 2nd image is after doing more fiddling and I got 1 script file out of 1371 done.
02.png
03.png

Used the JPN font file for this. W
 
thedicemaster said:
my best bet, the japanese version reads hex 00 as end of sentence.
the US version has 00 between all letters.(ordinary space)
the japanese version however has 83 (between letters in a name, appears like space) and 82(between letters in a sentence, appears like ,)
you could perhaps try replacing any space with 83/82 manually, but that would take ages.

note: these "spaces" (00, 82, 83) don't appear in the game.
Those are because the Japanese text is written in Shift-JIS encoding, which uses two bytes for each character, while the English text is in ASCII, which uses one.

QUOTEEdit: Did more fiddling with the Game file and I got this. 2nd image is after doing more fiddling and I got 1 script file out of 1371 done.
Hmm, it's a pity you lose the variable width font in the process, but I suppose there's no avoiding that. I'll bet there will be problems with line endings and so forth, too... but we'll see.

Did you just update the hex codes to the shift-JIS codes for each English letter?

EDIT: Also, if you're doing this, I should point out that there is some text used by the game in arm9.bin; but of course you can't just swap that (apparently both the font-managing code and the sound-managing code that causes a 'simple' undub to leave out all the removed voices are in there, sadly.) So if you're really going to try to do a complete script transfer, you'll need to edit the text in arm9.dat, too.
 
I put in ASCII for the text and remove the 00 between each letter. At first when I finished rewriting the line I put 00 and left the rest of the text and rewrote next part where it started, then decided to fill it with 00's.

So the hex string for Vicky in my screenshot would be
03 00 00 00 56 69 63 6B 79 00 00 00 00 00 00 00 05 00 00 00

Notable hex's I discovered are
06 - Clear the speech box for new text
07 & 0B & 04 - Close the speech box
0A - New line

Notes
  • You'll need to count the amount of characters put in before you need to start a new line, as the font size is different to the English room. Although since each character has the same width you just need to count how many characters you put in before starting a new line. (If you look at my third screenshot from my last post. The 'h' in "with" is the 28th character)
  • Wrong amount Text noise(the beeping as the text is appearing) eg. for a line that goes "..." you'll hear enough beeps to fill a line.
I was considering modifying the script for most of the game but 1371 script files is too much to go through.
 
Hmm, if you describe the process you went through to change your first script file, and perhaps provide a sample before and after of the first script, I bet I could write a Perl script that would make the changes automatically for the remainder of the script files, assuming they're reasonably similar.
Let me know if you're interested. I have more spare time as of late, and it might only take a few hours to write a script that handles this.

Also, this is my first post, hello all.
 
shineek said:
i doubt this can be done automatically
No, PokeyStick is right, it can probably be done automatically with a cleverly-written script. I was thinking about it myself, but I've been really busy.

There's a few complications, like the fact that you'll have to account for line endings (but that's made easier by the fact that this turns the font to fixed-width, so it's just a matter of counting characters.) But the main difficulty will be making rules to determine what's text and what isn't. Still, it doesn't look that hard to figure out.

The bigger worry to me is that some of the game's text is 'hardcoded' in the arm9.dat... that can be changed, too, but unlike the script files (which we can probably make larger or smaller as necessary) if there's not enough space to fit the English text in there, you'd have to use ASM hacking to make more. I think. I'm not a huge expert on romhacking specifically.
 
i do believe there are specific hex values(like 05) that determine things like end of message.
just checking for those, and groups of 4 or more times "00" should make it possible to automate the proccess.
 

Site & Scene News

Popular threads in this forum