Hacking God Eater 2 Translation (Looking for translators!)

[Coldbird]

If by story you mean cutscenes, just search for files named event_scr_text.

If you mean actual NPC dialog from inside the cradle then I don't know for sure yet, I'm searching myself (didn't have too much time to check into the dump contents myself too much).

 

[devnoname120]

Good job!

I've made a branch of the main project and uploaded what I've done into it -- would it be alright to make my branch public as well?

I was busy this weekend, but I will set up the project on Transifex.
You should then be able to do it there, which is more convenient, because there are tools such as a dictionary that help us to keep the names consistent across the translation.

 

[Masaki Kenshi]

Thank you very much Coldbird. I'll let you know if I'm having problems. Now i can help somehow to translate this game!!! OMG i cant w8 for the end results T^T. so hyped up xDD

 

[Masaki Kenshi]

Sorry for Disturbing Coldbird but, for ingura_poly_text are the words suppose to repeat?? Because when i translated the first 3 texts, it is the same for the rest ... and is it okay to save the translated files as rtf?? if not wat type do you want me to save it as?? And can u plz tell me how to send it to either you or Noname?? I apologize for asking lots of questions and Thank you for your time.

 

[Coldbird]

You open the file in notepad++ and save your edits.

Do not change the file format! You will only end up giving me extra work if you do.

And yes, duplicates might exist, but that ain't our fault...
It's just the way the game stored the files.

 

[Swing]

Dammit I think a year or so ago I found the files as asked if people were interested in helping me and was ready to extract them, then my fiance packed up all her shit from my house and told me never to talk to her again, kinda threw a kink in my interest as I was getting good and finding the right files to translate...

 

[Scorp]

Coldbird, do you know about this? https://translate.google.com/toolkit
It could save/load in getText format (that .po, as you seems know), so ideal for collective translations of said format. So no more messing with Notepad Also get auto-substitutes for already translated thing, auto-google translated part, etc...

 

[Coldbird]

Swing You should have pulled it off then, too late now. Instead of crying over how you didn't finish the analysis back then, you should instead try and contribute to the cause over here. We all want the same thing after all.
Also... the file format used in GE games is a pain in the ass anyway, its a fully linked-list multi-storage file system...

So for unpacking it you need to find iterate all linked list items until you reach the leaves of the tree, and when repacking, you need to roll the tree back up from the leaves to the root, storing the data into the right storage files (as indicated by the storage flags, there are 3 storage files, 4 if you include local resource storage).

Furthermore, even if you get that far, you need to still unpack the blz2 (yet another zlib compression with header information) files to get to the actual data (in our case the tr2 table files).
And even then, you still need to take those apart in a way you can hand the files out to translators.

Just be happy the shitty work has been done and the real fun can start now.

 

[Scorp]

Just be happy the shitty work has been done and the real fun can start now.

Depends By me most fun is in digging out the data and creating tools As really when you have all tools ready - remains a tedius, slow and boring translation work... Maybe that's only because I never translated a game? Who knows...

 

[Coldbird]

Haha, not all the tools are fully done yet.
Some of my repacking code is still glitched and needs reworking, also I constantly finetune my TR2 dumping code, which is why a new (third) dump will be put on github in a few minutes.

As I want to make translation as easy as possible and available to everyone no matter how tech-illiterate they are, I need to make sure my files are fully compliant with all those cloud translation tools out there.
Also, due to cloud translation we should be able to have a easier time translating the actual content, being able to cross match repeating data over multiple files, etc.

Another upside to GE2 when compared to other games is that it actually uses the firmware's own text rendering engine in combination with a fully international PGF font, this saves us yet another hacking step we would have to do otherwise.
Other than that, the only thing left to be done hacking-wise is patch up the sceIoOpenDNAS call to use sceIoOpen instead, and substitute the PGD encrypted data.pkg file with a unencrypted + english patched variant of ours once we repack the translation.

So yes, once all the files are on the cloud translation hub I think that translating this thing might actually be a fun thing to do.

 

[Scorp]

Also, due to cloud translation we should be able to have a easier time translating the actual content, being able to cross match repeating data over multiple files, etc.

Which service you're speaking about? Google? Or something else?

Another upside to GE2 when compared to other games is that it actually uses the firmware's own text rendering engine in combination with a fully international PGF font, this saves us yet another hacking step we would have to do otherwise.

It is usually not really hard to change that to load separate font. Usually enough to replace NID for sceFontFindOptimumFont with NID for sceFontOpenUserFile (if it was not imported in game), and replace the execution with custom routine which load font. Well, anyway I think you already know all that

Other than that, the only thing left to be done hacking-wise is patch up the sceIoOpenDNAS call to use sceIoOpen instead, and substitute the PGD encrypted data.pkg file with a unencrypted + english patched variant of ours once we repack the translation.

Does it uses PGD when load game? I do not have a game to check, but some games (Persona 2 for example) do not load PGD when just do a load, it uses PGD to do an install.

 

[Coldbird]

GE2 uses 3 archive sets, package.rdp (base game data), data.rdp (often used parts of the game) and patch.rdp (this one gets loaded from dlc and game update files).
If you do a data install of GE2, it will write data.rdp in a PGD encrypted variant, to the memory stick. It basically is a 390MB space waster inside the ISO file, as it keeps a PGD encrypted copy of that file and a plaintext one (for reading from the UMD).

So yeah, unless you install the game to the memory stick, you will never ever have to worry about the PGD file anyway...
But we are perfectionists, aren't we? So let's do things the right way and substitute that sceIoOpenDNAS call later so that idiots that feel like having to install the files even though they play an ISO can still play the game and not face awkward translation issues / crashes, ok?

Also... I've uploaded the latest gettext compatible dumps to github.
On top of that, I wrote a curl-script that is currently synchronizing the github repository into our transifex translation page.

So... translators, in about 30-40 Minutes (when the script is done synchronizing the data), you can start collaborating on the translation over at https://www.transifex.com/projects/p/god-eater-2-translation-project.
If you wish to contribute, just register at transifex and send us a call there that you wish to contribute, we will add each and every willing person to the team there so they can contribute.

 

[Scorp]

So just disable install option It is useless for ISO players anyway. But, of course, if you want to go by the harder way - no worries Also not too much of a hassle I think. But as I never did such thing, would be nice if you would share results in a armips file (I believe you use that for asm hacks?).

 

[Coldbird]

Actually I go a different route for ASM hacks.
I use the same routine I used back in my CFW crafting days: payload injection.

I extend the ELF file section size, and create a fixed load address code dump (written in C), which I then inject into the newly made code-gap...
Then all I need to do is place a single hook / jump point to the payload, and from there on after I'm in my managed code and can program whatever I want to fix / patch the game up.

No extra PRX file, no(t a lot) ASM hacking gibberish, just plain simple and readable C code.
I know its not everyone's favorite way of doing such things, but I do enjoy having managed code to rely on rather than a shit ton of complex hard to understand ASM code.

 

[Kitsu-neechan]

i would hardly say patching the pgd encryption is the "hard way".
takes somewhere from 2 to 6 bytes of edited data. Depends a bit on the game.

 

[Scorp]

i would hardly say patching the pgd encryption is the "hard way".
takes somewhere from 2 to 6 bytes of edited data. Depends a bit on the game.

Yes, but this in case if you have old PDG unencrypted, instead of reusing the one which is already exist. So wise way would be also to patch the name, to take unencrypted version. Which would take more than 6 bytes

Seems we're doing too much tech in the translation thread btw

 

[Kitsu-neechan]

maybe so, but i find it easier to just relink them on the iso TOC and leave a 0kb "install.dat" in the filesystem.
saves me from asm patching anything

 

[Scorp]

Oh, never liked things like relinking iso TOC, as pretty always this is pain to test things and also for patches to apply (unless you do xdelta sort of patch). For me much easier just to patch path to file with the path to decoded file in asm one time and forget about that Everyone prefer his own ways

 

[Kitsu-neechan]

i suppose, i just never stop finding it hilarious when people take the iso apart later and go all "WTF why is it 700mb bigger after i rebuilt it!?"

 

[Masaki Kenshi]

How can i join the enslish team in the transifex?? I can review the translations and fix minor errors.... I saw few of them already