Forums

I've been doing some reverse engineering work on Spellcaster (http://www.smspower.org/Development/Spellcaster-SMS), and I've been able to dump the US and JP scripts.

I think it might be possible to re-translate the JP script back into English to get higher quality dialogue if somebody could be bothered to do that work. To generate interest, I need to get the script into a format that can be viewed and manipulated, which I guess means unicode.

The JP script consists of a lot of bytes that more or less directly map to tiles in VRAM, with the exception of some accent marks. What I'm unclear on is how to map those to unicode. I'm assuming the ordering of the tiles in VRAM and any standardized layout for Japanese text are not the same.

Is there any easy way to do this? Or would I have to manually look up each character, find the matching unicode value, and make a mapping table that way? There's "only" about 120 tiles used so it's not impossible, just really boring.

Conversely I could use the tiles and print out all the text graphically in some monster sized PNG, but that doesn't sound particularly handy. If anybody else has worked on a translation project I'd love to know how this problem is tackled.

They are usually in some order, probably the standard 48 kana order, If you can post the 120 characters it should be easy to give you the mapping.

Oh! I've done a bit of work with this, let me get my stuff together and I will make another post soon with the goods.

EDIT: Aha! Attached an excel sheet with Japanese character table. Not sure if 100% accurate. Not checked since I last worked on it.
EDIT2: found my own script dumps, Shift-JIS encoding.
EDIT3: ..and the table used to do the dump (windhex tbl format I believe)

kuj-table.xls (21 KB)

Kujaku char map

kujaku-script.zip (20.73 KB)

Shift-JIS script

whex kujakuou.zip (743 B)

tbl

Translations usually use a ".tbl" format for the lookup, and (generally) horrible 90s encodings.

The tools for that format are so nasty and outdated, almost unusable on a modern OS. I would love someone to come up with some new general ROM hacking/translation tools. I have great ideas for them but I am too crap of a programmer to realise them.

Something about these dumps I remember: There is a single byte at the start of speaking lines that represents an in-game characters name, usually 01 02 03 I think, translation will probably not make sense if this isn't taken into account.

If you're still working on this, get Cartographer and Atlas from romhacking.net. They make inserting/extracting text ultra easy. Cartographer dumps scripts using a .tbl file and a pointer list and also formats it for you (using the game's control codes and how you set it up). Atlas inserts the translated text back in and modifies the pointers to point to the new lines which is a godsend (although if the pointers aren't in nice blocks at the start of text banks and instead interspersed through the code then it becomes much harder).

They're both command line tools so and take a bit to learn, but cartographer is fairly easy and you really only use it once when you've found all the text. Atlas you just make some .bat files for each block of text in the rom. Once you get them both set up it makes extracting/inserting ultra easy.

kingofcrusher wrote

If you're still working on this

Nah, I totally gave up when I found out everything I did had already been done years before. :\

But those tools sound really interesting, so I'll check them out for sure for some future projects.

Charles MacDonald wrote

kingofcrusher wrote

If you're still working on this

Nah, I totally gave up when I found out everything I did had already been done years before. :\

But those tools sound really interesting, so I'll check them out for sure for some future projects.

LOL, yeah that happened to me once too. Good times; good wasted times.

Really? I thought I had hardly gotten anywhere. There's still a project in this, surely.

Author	Message
Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	Spellcaster retranslation Posted: Sun Jul 31, 2011 5:08 am
Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	I've been doing some reverse engineering work on Spellcaster (http://www.smspower.org/Development/Spellcaster-SMS), and I've been able to dump the US and JP scripts. I think it might be possible to re-translate the JP script back into English to get higher quality dialogue if somebody could be bothered to do that work. To generate interest, I need to get the script into a format that can be viewed and manipulated, which I guess means unicode. The JP script consists of a lot of bytes that more or less directly map to tiles in VRAM, with the exception of some accent marks. What I'm unclear on is how to map those to unicode. I'm assuming the ordering of the tiles in VRAM and any standardized layout for Japanese text are not the same. Is there any easy way to do this? Or would I have to manually look up each character, find the matching unicode value, and make a mapping table that way? There's "only" about 120 tiles used so it's not impossible, just really boring. Conversely I could use the tiles and print out all the text graphically in some monster sized PNG, but that doesn't sound particularly handy. If anybody else has worked on a translation project I'd love to know how this problem is tackled.

Bock Site Admin Joined: 08 Jul 2001 Posts: 8649 Location: Paris, France	Posted: Sun Jul 31, 2011 6:32 am
	They are usually in some order, probably the standard 48 kana order, If you can post the 120 characters it should be easy to give you the mapping.

DMEnduro Joined: 27 Apr 2005 Posts: 420 Location: Australia	Posted: Sun Jul 31, 2011 9:41 am Last edited by DMEnduro on Sun Jul 31, 2011 10:15 am; edited 2 times in total
	Oh! I've done a bit of work with this, let me get my stuff together and I will make another post soon with the goods. EDIT: Aha! Attached an excel sheet with Japanese character table. Not sure if 100% accurate. Not checked since I last worked on it. EDIT2: found my own script dumps, Shift-JIS encoding. EDIT3: ..and the table used to do the dump (windhex tbl format I believe) kuj-table.xls (21 KB) Kujaku char map kujaku-script.zip (20.73 KB) Shift-JIS script whex kujakuou.zip (743 B) tbl

Maxim Site Admin Joined: 19 Oct 1999 Posts: 14740 Location: London	Posted: Sun Jul 31, 2011 9:57 am
	Translations usually use a ".tbl" format for the lookup, and (generally) horrible 90s encodings.

DMEnduro Joined: 27 Apr 2005 Posts: 420 Location: Australia	Posted: Sun Jul 31, 2011 10:05 am
	The tools for that format are so nasty and outdated, almost unusable on a modern OS. I would love someone to come up with some new general ROM hacking/translation tools. I have great ideas for them but I am too crap of a programmer to realise them.

DMEnduro Joined: 27 Apr 2005 Posts: 420 Location: Australia	Posted: Sun Jul 31, 2011 10:29 am
	Something about these dumps I remember: There is a single byte at the start of speaking lines that represents an in-game characters name, usually 01 02 03 I think, translation will probably not make sense if this isn't taken into account.

kingofcrusher Joined: 15 Aug 2009 Posts: 74 Location: Portland, Oregon	Posted: Tue Mar 20, 2012 4:25 pm
	If you're still working on this, get Cartographer and Atlas from romhacking.net. They make inserting/extracting text ultra easy. Cartographer dumps scripts using a .tbl file and a pointer list and also formats it for you (using the game's control codes and how you set it up). Atlas inserts the translated text back in and modifies the pointers to point to the new lines which is a godsend (although if the pointers aren't in nice blocks at the start of text banks and instead interspersed through the code then it becomes much harder). They're both command line tools so and take a bit to learn, but cartographer is fairly easy and you really only use it once when you've found all the text. Atlas you just make some .bat files for each block of text in the rom. Once you get them both set up it makes extracting/inserting ultra easy.

Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	Posted: Wed Mar 21, 2012 12:49 am
Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	kingofcrusher wrote If you're still working on this Nah, I totally gave up when I found out everything I did had already been done years before. :\ But those tools sound really interesting, so I'll check them out for sure for some future projects.

kingofcrusher Joined: 15 Aug 2009 Posts: 74 Location: Portland, Oregon	Posted: Wed Mar 21, 2012 5:43 pm
	Charles MacDonald wrote kingofcrusher wrote If you're still working on this Nah, I totally gave up when I found out everything I did had already been done years before. :\ But those tools sound really interesting, so I'll check them out for sure for some future projects. LOL, yeah that happened to me once too. Good times; good wasted times.

DMEnduro Joined: 27 Apr 2005 Posts: 420 Location: Australia	Posted: Wed Mar 21, 2012 11:01 pm
	Really? I thought I had hardly gotten anywhere. There's still a project in this, surely.

Forums

View topic - Spellcaster retranslation