cosmicspear
@cosmicspear

(Edit: Hey, future-me here to issue a correction. Turns out based on what others have noticed there's a good chance FM2 isn't an MTL at all but rather a more conventional type of inexcusably bad translation: https://twitter.com/HardFeelings21/status/1711678658421911776)

When I heard that there were going to be remakes of the first three Front Mission games coming out on Switch (and later other systems), and that they'd all be available in English, I was excited. Particularly for Front Mission 2—that one had missed an English release back in the day and so would be new to me. Not to mention pretty much everywhere I've looked considers it one of the best games in the series—and yes, the "no official English release" and "one of the most liked games in the series" riders both apply to FM5 too, but FM2 hurts more for me because it's a gap in the series, y'know what I mean? Learning that Square Enix weren't directly publishing it themselves but rather leaving it to some company I'd never heard of called Forever Entertainment made me a bit worried, as did the state of their remake of the first game in the series once I played that, but their FM2 remake didn't seem to have the same graphical failings of their FM1 remake and that was my biggest issue with that one, so I figured it couldn't be too bad.

The fact that I'm writing this says it all, doesn't it?

Except no, it doesn't, because I have to explain how it's bad and why this is a big deal. Because what Forever Entertainment have done is machine translate the game. More specifically, what they've done is something called a machine translation, post-edit (MTPE for short) where they throw the script into a computer, then hand what it spits out to an editor to touch it up. And if you know your stuff in this field that's all I need to say, but most people don't so I need to go into more detail.

Machine translations always suck. They can't not suck, because they're being done by a computer program that doesn't understand or even have the ability to understand things like context or colloquialisms that you need to understand to be able to translate things effectively. These days you have stuff like deepL that talks a big game about using AI to be better at this and can actually produce complete grammatically correct sentences often enough to fool people who aren't looking carefully into thinking it's better than the old Babelfish-style stuff, but that just replaces one kind of garbage with another. What that kind of translation software actually does is it outputs what is effectively a weighted average of how pre-existing translations of similar text wrote it, which means it completely flattens character voices and invariably gets things catastrophically wrong because, again, it can't tell the context of its source data.

Having a human editor sounds on paper like a way to minimize these problems, but it doesn't work because, again, the machine translation killed the original text's voice and completely changed the meaning of a bunch of lines. Which means that editing the text effectively requires you to look at the original source text in order to correct everything the program got wrong. I've seen a fair few translators on Twitter mention that this effectively requires you to throw the program's output out entirely and translate it yourself to get something that actually reads well, and I see no reason to assume otherwise. Which means it's almost always gonna suck, because if you're replacing your translator with a computer you're doing it because you want to save money, and with game localization being a dreadfully underpaid field as it is you can probably guess that MTPE rates can't possibly be enough to convince someone to care. So the end result is that almost every MTPE reads identically in ways that are extremely easy to notice once you've seen a few, and that makes it really obvious that the people responsible just didn't care. A MTPE, then, can effectively be read as the execs flipping off everyone who actually cares about the text and the artistry thereof.

Obviously this is always a bad thing, but Forever Entertainment's take on Front Mission 2 is even more insulting than most MTPEs because it's a MTPE of Front Mission 2. This is a game from probably the most respected era in Square's history, that's part of this huge sweeping military sci-fi epic full of intrigue, drama, and realpolitik. It's incredibly respected and considered one of the best parts of the series. It finally getting an official English release after over twenty years should've been huge, finally bridging a gap in the English releases that everyone expected would never be bridged.

And Forever Entertainment machine translated it.

How do I not see this as one of the worst insults ever leveled at a video game series and the people who love it? This is Forever's management saying that, to them, the extremely detailed long-term plotting of the Front Mission series doesn't matter and can be safely thrown into the MTPE meatgrinder, its writing transformed into the most generic drivel imaginable, all in the name of saving a few bucks. It's a disgrace, honestly.

It's especially absurd that they thought they could get away with this considering their version of Front Mission 1st used the DS script, which is a real localization and thus has actual character to it. Which makes their FM2 script's worthlessness even more obvious since a comparison that shows what it could've been is right there.

So, what do I think should be done? Well, for starters they could apologize. A public admission from Forever that they fucked up would help. But they'd also need to fix it. Ideally, this would involve throwing the entire machine translated script out, hiring a real translator-editor team to redo the whole thing from zero, and paying them enough to care about the result. If they're willing to do that, then they can probably fix things.

But that's a pretty big "if"


gosokkyu
@gosokkyu

I've seen virtually nothing of Forever's FM2 remake but I have no trouble believing it's entirely or substantially machine-translated because they have an established reputation for abusing MTL in Japan, and I'm constantly baffled by the fact that the likes of Sega and Square-Enix continue to allow them to release remakes of Japanese-made games in Japan with such shoddy Japanese.

One of their more notorious gaffes was this trailer for the first Front Mission remake, which they quickly pulled off youtube after people roundly clowned the goofy Combat Echizen-esque narration:


iiotenki
@iiotenki

I will simply add to all of this as a professional Japanese game translator who rejects MTPE offers a couple times a year that, yes, the rates are insultingly, unsustainably bad. All of the ones I remember getting have been for less than half of my starting rate when I was brand new back in 2015, an amount of money I would describe as "barely better than your average eroge visual novel translation" and I suspect with the way things are headed, those amounts are on the high end these days.

Quite frankly, the reality is, even if you pay us premium rates, the actual cost of a basic, low-frills text localization is going to be fairly inconsequential compared to building the rest of a game like this from scratch. Not only that, the fact they're just using the existing Japanese script as a base means that they're already saving tons of time and money generating that content from scratch, meaning there's even less reason to cheap out in the manner that they have. I get that this is a release that's probably only going to sell so well even under the best circumstances, and I don't know what their revenue/dev team size looks like, but from my napkin math, I have a hunch that they're probably paying more in engine licensing than a respectable localization would cost. And even if our work did cost more somehow (that script would have to be pretty damn gigantic, though), so what? At least you're showing proper respect to the source material, the rights holders, the original developers, and, of course, people in the localization trenches in general.

Incidentally, while poring over the credits of the first remake to figure out what engine they are using (it doesn't explicitly say, but I think we can all safely bet it's probably Unity), I saw that despite using that DS localization and crediting the original Japanese development team, they couldn't be bothered to credit even the people responsible for translating that material originally. Or, you know, giving them the slightest shred of decency in exchange for legally not having to pay them to reuse their work. đŸ€·


You must log in to comment.

in reply to @cosmicspear's post:

This sucks. Doing MTPE to a game with simpler dialogue is still a declaration that the dev leadership thinks their audience will be happy with slop and jangling keys, but the entire reason I would consider trying out something like the Front Mission franchise is gonna be the quality of the writing.

Considering Square had been saying "hey if these sell, you might get a remake of later games in the series too" - I feel like that probably won't happen. It sucks that this still happens in localization, in fact it's probably gotten even worse with the advent of DeepL and the like. Been a fan of these games for a while but this puts me off spending money on even the FM1st remake now.

The fantranslation has been incomplete for over a decade. Apparently text insertion has been an issue. I was really hoping this would be a full version of Front Mission 2 that the fans have been hoping for...but Forever's pretty much consigned themselves to churning out mediocre remakes.

SAME

i'm just glad there's precedent for a game studio responding to feedback for a crappy loc, throwing it out, and having a new one made from scratch (Ys VIII: Lacrimosa of Dana), though given how Square-Enix seems to be fumbling financially it's not a guarantee here

The whole damn series is. It's just been getting kicked while its down since the shitshow that was Evolved.

Front Mission 3 was my first exposure to mecha and everything post-FM1DS has been garbage. Fuck Forever and fuck Squenx for approving this.

in reply to @gosokkyu's post:

in reply to @iiotenki's post:

so someone on my discord server got curious and decided to cross-reference and we ended up with likely someone who doesn’t know english nor japanese and may or may not have depended on some machine translation.

there’s some mistranslations only humans could make (easy example: どうかăȘ translated as what do i do instead of something like hmm i wonder, making another line read like nonsense).

take that what you will, but it sure sucks to see how bad front mission 2 tl got.

Phew, kudos to you two doing that much due diligence! But yeah, if those are the sorts of mistakes being made, I agree that's definitely a sign there was more of a human touch in this than initially thought, for better or worst. At that point, in a best case scenario, we're obviously look at someone dropping the ball at least on the developer side, if not beyond that. I don't think it'll come as a surprise to anyone that whenever Square does things internally and hires freelancers, they pay close attention to our work (but, in my experience, will respect your decisions and opinions if they're sufficiently explained), which suggests to me that once money changed hands, Square likely wasn't offering much supervision beyond surface level stuff. (NOT that I'm saying their localization chose to stay out of it; people underestimate just how much they actually handle internally, speaking as someone who's only done more modest fare with them.)

Either way, like you said, a crying shame for one of the last big titles left to localize in that era of their catalog, especially when they've shown such loving treatment to a lot of other holdouts in recent years. I know for a fact their internal staff maintains institutional knowledge and documentation for this specific series, so it really, really didn't have to be like this. I have to imagine that staff isn't terribly happy about this either.

What fries me the most about this is how blatantly it shows Squeenix playing favorites. Subletting it to a royal crapshoot of a contractor is bad enough, but then to not do any sort of oversight or quality control until the damn thing is about to release shows that Front Mission has always been an afterthought to them. Anything that isn't the big fat Final Fantasy machine gets treated like dirt, and considering how much of a roulette wheel FF's been lately, I don't understand the logic of putting all their eggs in one basket like that. I guess the conservative reliability of Dragon Quest makes them feel they'll always have money coming in, but considering the last major DQ game was 2017, even that almost feels like it's not getting doted upon as much as FF.

I was hoping these remakes would do well enough that we might finally get an official outside-of-Japan release of Front Mission 5 someday, but that sure as Hell isn't happening now. This series deserved better, but I swear, the amount of boneheaded decisions SE has been making lately makes me wonder if they're heading for a big fall. And who's going to take the brunt of that fall? Certainly not the executives who are leading it down the drain! Layoffs ahoy for the decent working designers and developers who get tossed out into the streets when the parasitic suit-wearing leeches can't suck any more nutrients from their flesh!

And how's FF doing right now anyway? XVI sold three million copies its first week, pretty damn good. And yet SE has said its sales have still been "below expectations." How high are these expectations anyway? High enough that God itself needs a ladder to reach them?! Whoever doesn't get laid off is going to be forced through the crunch belt for the next big multimillion dollar project, which, by the way, is ALL they're going to be working on after they decide trying to uphold more than one series is a waste of time because Front Mission 2's nuclear hackjob was marched out to die on arrival.

I swear, the gaming industry as it stands CANNOT sustain itself if it keeps going in the direction it's going. I'm not sure if it'll ever face a great crash like it did back in the early eighties, but something bad's going to happen if the people in charge don't pull their collective heads out of their asses and wright the ships before they get sucked down the whirlpools of failure. I don't know what could happen at worst, but at the very least, I think the great streak of games we've been enjoying since 2017 is going to come to a stop as more projects get halted and/or outright cancelled. If it doesn't print CoD-money, it sucks, so throw it into the garbage. That's what I'm worried we'll be seeing the big companies do, and while I wish I could say "indie games will save us," even smaller studios have their struggles, i.e. everything that happened with ZA/UM.

I just don't know how to fix the problems that are going on behind the scenes with my favorite hobby, but I think a dark age is on the horizon and there isn't too much we can do to stop it from happening. Mass layoffs left and right, rumblings of the "Metaverse" being the future from self-important CEOs, AI threatening to derail the worth of artists and voice actors, and even if Riccitiello was ousted from Unity, I'm sure he'll have a golden parachute to cush his fall while everyone else in the company goes down with the plane. This FM2 fiasco only feels like the latest in a long gauntlet of ball-kickings that this year in games has wrought upon the people trying to make a living in it. All I feel like I can do is vent my impotent rage as things just continue to mutate out of control like this. It's been a great year for the actual quality of games, but it's been a terrible time for the industry behind it, and if this is the price of video entertainment... I dunno, maybe I need to take a break from games until the industry gets its shit together. Which, I'm not sure it ever will, at least not anytime soon.

I dunno, fuck capitalism. Rant over. Very sorry.