Kelebek1

Does this help? http://seiya-saiga.com/game/0verflow/shinydays.html

Because there isn't usually much point. If the script is written by an American then it just sounds awful, because it's written the way Americans speak (or the way they THINK British people speak), rather than how we actually speak. There have been countless games where that's been the case, and I'm sure it'll be the same with anime as well. The script should just be read by the nationality of whoever wrote it, otherwise it just comes out a lot worse than just making the character American.

Nyaa, and anime-sharing is a good place to check for new VNs.

I seed my bedroom floor more than I should. I did seed things and always tried to get a 1:1 ratio, but since switching my net over, my upload speed has gone to absolute balls and is capped at 30Kb/s. There's no way I can seed GBs at 20kb/s and aim for a 1:1 ratio. Companies have really been hitting upload speed hard since it's more expensive than download, and I don't live in an area that get very good internet (2Mb/s is the max, no fibre or anything here yet). I still await the day that I have a good enough net speed to stream anything.

Which game is it?

Blade Briders out next month, it's looking good. \o/

If you want difficult translations, a translation of Senshinkan would be amazing, cause that VN's absolutely unreadable.

The game files are actually in DATA.cpk, with the script files themselves inside tar archives within DATA.cpk. After unpacking all the tar files and creating script dumps from the scripts, it came out to be 430MB. 7625 files in 1901 folders. Mental. So here's my extract script anyway: http://pastebin.com/qwvHJiQF - For python 2.7, not 3.0+ It's a ridiculous mess and I don't really recommend you actually use it, but it's available for whatever "reference" reasons. To get all the scripts: Extract all the files from DATA.cpk with whatever program. CriPackedFileMaker didn't work for me and I had to use the QuickBMS cpk script instead. Extract it into some sub-folder. Put the script into the folder, so the filepaths are like: script.py files\_dummy.dat files\BGAE.LBG Open a cmd and run the script as "script.py -unpack files" (or whatever that folder name is that has the files). This will extract all the .TAR archives into an Unpacked folder. Then run "script.py -e2 Unpacked" to extract scripts from all of the script (*.ssb) files. You will then have all the script and can edit whatever you want in them text-wise. You can add opcodes to the script for new text boxes, but I don't know exactly what you'd need to add, testing takes way too long with repacking the CPK so I don't really want to spend ages trying. If you do add to the script though, just add the opcodes themselves, so just copy everything from the right of the ::. Boxes need to end with a ▼ character for the game to wait for input, otherwise it just goes to the next line. A tip for finding the actual text in the scripts, look at the first call-0xd you see right near the top of the script, and then skip down to the line with that number, as that's where the "main" text script part is. After you make whatever changes to the text, you can re-compile the whole Scripts folder by running the script as "script.py -c2 Scripts" and everything will be recompiled back into the .TAR files in a Compiled folder. You can then use those with the other files and build a new CPK with CriPackedFileMaker. Options I used were use filenames, try compression and align to 2048 and that works fine. I didn't try other options or data aligning but they may work too. Put your new DATA.cpk back into the UMD and done. Also you can extract and compile scripts one at a time instead of doing the whole massive Unpacked\Scripts folder. You can straight up extract the .TAR archives from the DATA.CPK with any standard archive tool, winrar, 7zip etc, it's a common format. Then you can copy any *.ssb you want to work with into the same folder as the script, and you can extract it with "script.py -e1 file.ssb" and compile it with "script.py -c1 file.txt" - I recommend you add some extra part to the original *.ssb after extracting it, otherwise it'll get overwritten when you compile, so just call it *.ssb.orig or *.ssb.bak or something. You can then use that and make your own .TAR. 7zip works just fine for that. It's not very nice I know, but there we are. As I said, probably better to find someone who can make a proper program for you and handle all this stuff is a much nicer way. Oh yeah, text should be 2 lines, not 3. Also that font is bleh, at least above native resolution.

Ah, I figured there was going to be a simple solution but I went for a more difficult one anyway. I was able to edit the exe to allow for non multiple of 2 strings (and allows for mixing JP/Eng regardless of position), but it then somehow turned into giving all characters full-width spacing even though they were ascii, no idea why. I fixed that and then found it strips all spaces from the text before even getting to the display loop anyway. What a pain. >: The loop for the text at 0044A9C0 is just silly.

It can read single-byte characters just fine, the only limitation is that the string has to be a multiple of 2.

Change the filename and get rid of all the kanji then?

Weirdly my first one is also the first one on VNDB purely by coincidence; Let's Meow Meow.

What surprises me most is that of the 3 VNs listed with "Dying Sex," one of them is an otome game.

Are you sure? If the file is still encrypted (and actually a PGD instead) then you'd have to decrypt it with another tool first, like PSNPKGDecryptor. I've never run into an actual CPK file that CriPackedFileMaker can't handle. EDIT: Yeah I actually had a look and you're right. Apparently it uses some special type of encryption and doesn't have UTF8 tables. I found a program which can extract the CPK: http://hcs64.com/files/utf_tab07b4_special.zip - use unpack1. It can't pack again though, and the author said he has no intentions of trying to. So I suppose either extensive ASM hacking to load unencrypted files, or explore the source in that zip and make a compiler.

There already is a tool, it's called CriPackedFileMaker.

Alright fixed that problem, looks like there's only .txt and .dat so that's alright: https://www.mediafire.com/?8juirjx61sn4ghn Nice to know the exe works though, since I haven't actually been able to test the full version. And yeah I said about that problem with the text. I really don't know how you'll get around that. Given how I think the script works, it'd require a whole bunch of re-coding how the script engine works, and yeah I can't do that. The "easier" (I think) way would be probably be along the lines of modifying the font or something. Probably have to create your own for it though... Certainly not a nice solution. I've never seen a VN that can't display ascii at all before. Does seem to be a pretty gaping hole in their scripting system, if they ever wanted to use ascii... Some ascii can load, but the strings must be a length divisible by 2. It throws an error when the script loads and then shows up all broken like that, every second byte gone and it took the whole thing was a command or something. But I suppose it, technically, *can* show ascii. Just that the script format isn't taking it as text properly.

Ah, right. Eh that's a real pain. I wish the ExtractData preserved the filenames. Since it cut off the extension for all the .scr files I thoguht it would do it for them all, but it left .dat it seems. I'll have to make a specific list then, and those files aren't in the trial, everything in the trial is a .txt. Can you give me a list of everything that isn't a .txt? Or hell, just sending me the scr.aos itself would be really helpful, then I can test myself. The file shouldn't be too big I hope.

Did you put everything from the extracted scr.aos archive in the Scripts folder?

Alright well try this: http://www.mediafire.com/download/cjkiyb0yh2xuitq/ Extract that to some folder, and put scr.aos in with it, and then in a sub-folder put all the decrypted scripts. So the directory tree would be like: Folder\Compile.exe Folder\scr.aos Folder\Scripts\00_00.txt Then run the exe with the name of the folder that the scripts are in. So in that example I'd use "compile.exe Scripts" (without quotes) and it'll then create a scr.aos.new which you can rename and use in the game with the edited exe which I also included in the .rar. To call the exe with params, hold shift and right-click somewhere in the folder (as if you were making a new file), and you'll have a "Open Command Window Here" option, use that and then type in compile.exe with the name of the folder like I said above.

Alright here you go: https://www.mediafire.com/?4xcbg54gl83j8gd The only files that work with it are the 0*_0* ones, A_*, B_*, C_* and E_* type files. The rest have to stay original. I don't know how many more are added in the full ver, but as long as the file starts with # * then it can be loaded decrypted. You still have to add it into the archive file (scr.aos) though.

Don't have Skype unfortunately.

Well, making Huffman trees was way beyond me, so I went for a different tack and instead just had the game skip decryption for script files. Downside is that it requires editing the exe so you have to distribute that as part of your patch. I also only have the trial to work with, so I can't fix the real exe unless someone wants to send it to me (if you even want to do it this way to begin with). The major problem with it from a translating point of view is that you cannot use ascii for the script, it just crashes the game. The game will only accept SJIS for text. I think what they've done is have the script take in all ascii as script commands. There's probably a special way to get it to work, some sort of escape code or something, but I can't find any examples of ascii in the script to see how.

I don't usually care much about visuals of VNs either, but this one really impressed me. Resolution does a lot for the graphics I think. A bad 720p VN can look better than a good 480p one.

Well I got one working. It's not the most beautiful thing ever, and honestly I don't recommend anyone to actually translate anything with this because it's ridiculously messily bad and the dump is 6MB so it's not very easy to navigate at all, but I'll put it out there anyway to encourage some proper devs to make better tools. Also I've only had the trial to work with, and it's extremely likely that there's going to be more opcodes in the full game, and I can't really do anything about that as I don't have (or really want to, sorry) get the game. http://pastebin.com/tWrzAdLM - For Python 2.7 (not 3.0+) Usage is "script.py <-e/-c> <filename>" -e to extract, -c to compile. Search the dump for "string" to find all the actual strings. Keep them to one line, the game uses automatic newlining, although it ignores spaces. To add a new string you'd need to add code like this (everything highlighted): But again, I'm not really putting this out as something to actually work with, just to hopefully have others make better tools. All the opcode definitions which is all you really need are all defined at the top of the script.

The one in my screenshots is OllyDbg, but you don't need that. Any hex editor will do, I gave the addresses that you need to edit. There's tonnes of hex editors around, HxD or wxMEdit are two free ones. I don't have any SQL editors, it just says at the start of the files "SQLite format 3," so that tells you you'd need a SQLite program to read/write them properly.