HomeForumsWhat's newResources 
 
 
Any docs on trans/text extraction?
GaijinPunch - Mar 3, 2004
  GaijinPunch Mar 3, 2004
I see a lot of docs here on developing games for the Saturn.

I know enough C/C++ to do some really lame stuff on a terminal, and that's about it, so that's out for now.

I would like to tinker around w/ some files though, suck out some text and whatnot. I've done a search here, but only found a few threads (and some dead links) so am back to square one. A simple finger in the right direction (and preferrably no the middle one) would be appreciated.

Oh, yeah -- 'sup SSDDR? I finally made it.

Ciao

GP

 Djidjo Mar 3, 2004
See the Cyber Warrior X's Junk Pile... and look at the "Saturn Hacking docs". Everything here is explained on how to extract text and hack games.

Hope this helps

Djidjo

 madmalkav Mar 3, 2004
CyberwarriorX documents are great. Anyway, if you are unfamiliar to game translating, take a look at the documents in Romhacking Repository.... Is always good to know the basis.

 GaijinPunch Mar 3, 2004
Thanks for the prompt reply.

I'll check that out!

EDIT:

madmalkav posted just as I did.

I'll check out the Romhacking repository. I'm familiary with the translating side...not the extracting/inserting.

 GaijinPunch Mar 5, 2004
I've taken a quick gander, and I do have a question (that I'm afraid I know the answer to)

In the example at Cyber Warrior X's page, he talks about stumbling across a file that has the font layout and whatnot (Magic School Lunar), and says that it's in SJIS. I looked at Lunar: SSS and found something similar (as well as the FMV sequences script written out in Japnaese... kinda cool).

Anyways, what would one do if:

1) It's not so easy to find out if it's in SJIS or not

2) It isn't encoded in SJIS?

I've read on places and they basically abandon a project that's not in some type of standardized font encoding.

One of the games I was poking around was Black Matrix (b/c I'm playing it right now). It's only got a few dat files (15 or so at most). Couldn't you in theory search each file for the hex values of say... a characters name and go about it that way?

I've still got some more reading to do, but those questions just kinda popped into my head.

Finally, it seems w/ (snes and nes) roms, the emulator is the key to dumping info (and a LOT easier to attain/setup than the CAR method). Any comments on SSF or Satourne's ability to play such a role?

 madmalkav Mar 5, 2004
Current saturn emus are totally unable of playing that role.

About what to do with the games that don't have the text in standard encodings: do exactly the same you will do on a SNES/NES/Genesis/... game. Of course, if the text is in SJIS is easier. Is also easier if the txt is already in a Word document with instructions for the translators but, hey, we are talking about the pleasure of reverse engineering!

 ExCyber Mar 5, 2004

  
	
	

		
		

			I've read on places and they basically abandon a project that's not in some type of standardized font encoding.


These projects most likely never would have been completed anyway for lack of a skilled low-level hacker. If you look at good finished translation projects, they usually have an assembly/machine code wizard on the team, and that's primarily whose job it would be to figure out a complex encoding. Neill Corlett, for instance, had the unenviable job of figuring out the encoding in Seiken Densetsu 3, which had its text in some kind of twisted... multilayered compression code. Later, however, that person would likely need to do other things like hacking code/tables that keeps track of how long strings are, modifying the layout of screens where you name characters, and in some cases rewriting some of the display code to make text windows bigger, character names longer, etc. Translation generally isn't just a copy/paste operation.

 CyberWarriorX Mar 5, 2004

  
	
	

		
		

			Anyways, what would one do if:



1) It's not so easy to find out if it's in SJIS or not



2) It isn't encoded in SJIS?


1. Pick a sentence from the game, and generate the appropriate code in shift-jis for it. Now search the entire disc(using your tool of choice). Chances are, if you don't find any matches, the text isn't in shift-jis. See #2

2. If it's not in shift-jis it's usually time to cry/swear/break things/whatever suits your fancy ;P It could just be compression or a type of encoding(see [insert random konami game here]). With those type of games you basically need to know sh2 assembly and find the function that decodes the text.

It could be too that they just wanted to cut down the size of the font and decided to use a cut down font. Of course the text format for games that use a cut down font depends on the order of characters in the font. For the older systems using a cut down font you usually just had to come up with a fairly small table file(see romhacking.org for more info) and use a generic dumping program to dump the script. The problem with the newer games is that the fonts can be so bloody large that it can take forever to come up with table for it. Anyways, what you could try doing is first of all find the font. See what order the characters are in. Then try doing a relative search based on the ordering.

Cyber Warrior X