hikarisomnium

I decided to try my hand at a project that I find truly interesting. I've been eyeing the game Maitetsu - Last Run!! for a long time, and I finally got around to working on it. I've always considered Kirikiri to be one of the easiest engines for translation, but this game surprised me. Before committing to the project, I made a "basic test MTL translation" for the very beginning of the game to see how everything would look. However, I discovered that I couldn't repackage the modified game files back into XP3 in a way that made everything work properly. I usually used Garbro or KrkrExtract, and it always worked fine. In the worst case, I would simply delete the original XP3 files and replace them with folders of the same name containing the extracted content. But this time, I hit a dead end. Archives packed with Garbro result in the game failing to launch. KrkrExtract, on the other hand, simply refuses to pack the archive, throwing a "can't read" error during the process. As for the "universal patch," I don’t really understand how it works. I tried different approaches, but none of them worked properly. I tried looking at how the translation was implemented in the sugoivisualnovel translation. It seems they used the "universal patch" and added a ProjectDir folder to the game with translated scenario files. However, with this approach, tips simply don’t display, and I’m sure there are other negative aspects tied to this method. I want to take part of the translated resources from the "standard" eng version of the game (if that’s possible) and add them to the final version, while translating the remaining content using GPT-4o (or possibly DeepSeek v3). If anyone has any useful advice, please share.

Hello everyone! I decided to try translating the game "Mugen Kairou 2 Rasen" using ChatGPT4o, and I’ve currently finished the MTL part of the project. I’m now proofreading, doing manual corrections, and also fixing display issues. So far, the best display has been achieved using VNTextProxy. I placed the sjis_ext.bin, ddraw.dll, and a font (the filename doesn't match the actual font name) into the game folder (for example, the font's name is Tsunagi Gothic black but the file is named TsunagiGothic.ttf). I don’t know why, but this method (out of all the ones I've tried) prevents the display of Latin characters with significant gaps between them. The problem is that occasionally, in some dialogues, there are display issues like this – https://freeimage.host/i/2L8UHa1 (Don't mind the brackets at the start of the name and line; this is a special marker that allows me to check the text in the game more efficiently and make corrections in the files). In fact, the line doesn't have that many dots, it's just an ellipsis – https://freeimage.host/i/2L8kVBS It looks like the game engine is ignoring what's specified in the translated text files and inserting punctuation marks that were in the original. If anyone has any suggestions or ideas, please share! UPD I couldn't solve the problem in the end. Overall, such "defects" don't prevent the game from being completed, but it's a shame that I spent so much time and got nowhere. I ended up testing the game and completing it 100%, but I skipped proofreading. Maybe I'll come back to the project later, tweak it a bit, and release it in some form, but for now, I have no desire to do so. In the end, if you translate using VNTranslationTools, there will be an issue with text display—not critical, but still annoying. If you use "selfmade tools," it seems like the problem can be solved, but then it's unclear how to fix the issue with the large spacing between letters.