I’ve been thinking about ways to get translations done quickly for at least a year or two now, but I’ve kept them to myself because I didn’t want to get anyone’s hopes up or have any ideas interpreted as promises. So just keep it in mind that these are ideas that may never happen. However, with VH5 coming out and it being clear that 1) I translate/edit/release slowly, 2) there are lots of games I’d like to see translated, and 3) no one else is translating them, it might be worth exploring some of these ideas, or at least seeing what people think or if you have better ideas and suggestions.
I should mention now that I’m not doing this for St. Monmusu Gakuen, but for other projects that would otherwise go untranslated.
The first general idea is for me to simply play through a game and record myself translating the lines in real time then have someone else handle editing. The biggest hurdle here is that I’m not fluent in Japanese, and even the best translator is going to stumble over lines that don’t translate easily into English. Also, while people do offer to help with editing, who knows how enthusiastic they’ll be to edit 10,000 lines. So what do I do with the recordings…
1. Speech-to-text: Install Dragon Naturally Speaking or equivalent and put text directly into the game as I go.
Easiest for me to edit out garbage on-the-fly.
Easy to pause to think about a sentence or look up a word I don’t recognize.
Easiest for other people to edit.
Can release unedited versions.
Hard to get sounds, onomatopoeia, etc. correct. Same with names and game-specific terms that the software won’t recognize.
More work on my end.
2. Audio recordings: Simply release the audio to editors
Minimal time usage for me.
No problems with sounds.
Easy for me to explain things, add notes, etc.
Matching up lines might be confusing for editors.
More work for editors, and more likely they’ll screw up something that causes bugs/crashes.
Half the audio could be me correcting myself or giving 4 versions of the same sentence.
3. Let’s Play videos
Similar to audio, except people could watch the game being played so matching up lines in the text files would be easier.
Getting pulled from YouTube for pornographic content?
Similar disadvantages for audio.
More work for me than audio.
The second idea is something that Rogue himself came up with a long time ago: creating a translation site that allows anybody to contribute translations of lines for a game. Here’s how I envision it working:
Step 1: A parser that takes the game’s lines and formats them for the website.
Step 2: Build a website where people can view the lines and contribute translations for them.
Step 3: A compiler that takes the translated lines from the website and builds the game’s script file back.
Steps 1 and 3 will be engine dependent, but I don’t think terribly hard for Kirikiri games or RPGMaker games when combined with already existing tools. Either or both steps could be done with manual intervention if necessary too.
The idea here is that there are plenty of people who know Japanese but don’t want to spend hundreds of hours translating and editing tens of thousands of lines. Even beginners to the language can handle easy lines, and people who know no Japanese can still edit the English. It’d have a feature to enable users to flag lines for review and rate difficulty of lines so more experienced translators could take a look at those. Throw in some basic security and version control to prevent/revert vandalism and bad translations, of course.
Unsurprisingly, the reason this has never gotten off the ground is because it requires work…
Anyway, just throwing out some ideas.