Jump to content

Recommended Posts

Posted

Hello! I'm new here and I just wanted some pointers. I'm a guy with one main goal; fan translations.

Issue is... I don't know where to start. I don't know crap about code and even where to start. My goal is to MTL first, then line by line heavily edit each, using the mtl and my somewhat-okay Japanese to make a good translation, and use my good friend as a proof reader and editor.

As someone completely new to everything, I was wondering if anyone could give me pointers? Tools, videos, tutorials, forums, anything is useful! I'm open to longer discussions too!

I plan to start with very short games and make my way up to longer games!

I'm using a acer laptop that runs things pretty well, and have experience playing Japanese games with textractor

Thank you in advance!

:choco:

Posted (edited)

Welcome!

Don't mind Zaka, they are just prickly regarding MTL. Even VNDB started to accept MTL a while back.

The traditional method to translate VNs is to extract the assets, translate the scenario text, and then put them back together. Garbro is the most notable program to extract assets. Sometimes, like with kirikiri, it can put them back together too. Otherwise, you usually have to search for game engine specific tools on Github or posted on odd forums. VNDB sometimes lists the game engine, the software used to create the game. The main upside is that this gives a native-like experience when playing the game.

There is also a newer method to translate text that Mtool and LunaTranslator, site, use which is to extract text from memory while the game is running, and then overlay a translation. The main downside being the program to extract and overlay text is always required to view the translation.

MTL can be used in either of the approaches. The fastest JPN->ENG MTL that is still mostly legible right now is the model used in SugoiToolkit. For non-English MTL, there are lots of models available on huggingface.co (the transformers library website). The best quality MTLs are AI translations and DeepL. There are some questionable attempts to quantify how much "better" each AI model is at translating than other ones.

Edited by Entai2965
Posted
32 minutes ago, Entai2965 said:

Welcome!

Don't mind Zaka, they are just prickly regarding MTL. Even VNDB started to accept MTL a while back.

The traditional method to translate VNs is to extract the assets, translate the scenario text, and then put them back together. Garbro is the most notable program to extract assets. Sometimes, like with kirikiri, it can put them back together too. Otherwise, you usually have to search for game engine specific tools on Github or posted on odd forums. VNDB sometimes lists the game engine, the software used to create the game. The main upside is that this gives a native-like experience when playing the game.

There is also a newer method to translate text that Mtool and LunaTranslator, site, use which is to extract text from memory while the game is running, and then overlay a translation. The main downside being the program to extract and overlay text is always required to view the translation.

MTL can be used in either of the approaches. The fastest JPN->ENG MTL that is still mostly legible right now is the model used in SugoiToolkit. For non-English MTL, there are lots of models available on huggingface.co (the transformers library website). The best quality MTLs are AI translations and DeepL. There are some questionable attempts to quantify how much "better" each AI model is at translating than other ones.

This is all super helpful!

About the MTL the other person mentioned, I get being skeptical, I usually avoid MTL as well but my goal is to mix my knowledge I have ( I'm better at speaking and listening to Japanese than I am reading, as I'm still in the middle of learning that ) while also comparing it to the translation I get from the machine until I get a perfect sentence. I have 2 friends helping me out too but their Japanese is worse than mine. I'll only be using MTL as a "base". I've also been a writer for 6 years, so I have confidence in not making the TLs sound lifeless. In the future, I plan to no longer use MTL, but currently that's where I'm at even if I wish I wasn't 

:oyasumi:

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...