Passion and Endurance: How Videogame Fan Translations Get Made

Whenever there’s a Nintendo Direct, the lead-in has some people wondering if this will finally be the Direct that announces Mother 3’s English-language release. We don’t need Mother 3, though. That’s not a referendum on the (excellent) game itself. It’s that we already have Mother 3: if you want to play it in a language besides Japanese, you can. And we can thank unofficial translators for that. Through painstaking work on their own time, these translators took the original Japanese release apart and rebuilt it for a new audience by creating a patch for the Mother 3 ROM file. As the translators themselves said, you’re on your own for finding the ROM, but the patch to make it playable in English—along with instructions for patching it—are there, and have been for 16 years now. 

The patched version of Mother 3 isn’t a direct replacement for an official release, but sometimes these do-it-yourself affairs are all we’ve got. This translation work is a form of preservation that also fills in gaps in the industry’s history and our understanding of it—think of how long, for instance, it took for something as ambitious and experimental as Live A Live to be widely known outside of Japan. That title was released for the Super Famicom in 1994, but it wasn’t until a 2022 remaster that a worldwide audience got a shot at this fantastic piece of Square’s history, one described by Jackson Tyler as “the beating heart of RPGs.” The reason many of us knew this to be true before it finally got that 2022 re-release is because it also received its first completed unofficial translation way back in 2008. 

“Translation” is a bit of a misnomer for what this work is. “Localization” tends to be the industry term nowadays, for reasons that make sense when you recognize what actually goes into translating these works from one language to another. As Paul Starr, a translator and editor who currently works on the weekly translated Shonen Jump manga, Me & Roboco, explained to Paste, “The practical answer is that ‘localization’ used to largely be professional jargon, a term of art that described a certain kind of translation of a certain kind of media, i.e., translations of software and games where the fact of the translation was meant to be largely invisible to the user/player. As information about how games are made has become more and more available, players have gotten more opinionated about what they do and don’t want to see in a translation, and the term ‘localization’ has become contested territory. As a linguistic descriptivist, I would define ‘localization’ as the term of art used to describe the professional field of translating software and games. It’s a specialized discipline that needs some kind of descriptor.”

That bit about “invisible to the user” is a vital part of the translation and localization experience. You’ll sometimes see calls for something to be directly translated from the original text, with the defense of this being that it maintains the original intent of the artist in question, but that obscures that things can be literally lost in translation: a joke that lands in one language, for instance, might not land in another. It might be because of what the reference is to, it might be because of some societal norm that’s not quite the same everywhere, it might even just be because what was a play on words in one language won’t present as such when translated into a different language. Something like this being left in would ensure that the translation work was not invisible, as it would be clear that something, to reuse a phrase that exists for a reason, would be lost in translation. 

“When I said ‘invisible to the user,’ what I was getting at was a situation wherein the fact that the translation was performed at all is meant to be invisible,” Starr went on to say. “Consider something like the Japanese localization of the Windows operating system—clearly there’s a ton of text that has to be translated there, and thousands of tiny judgment calls that someone could potentially dispute, and they’re all ultimately in service of an experience that’s meant to make Windows feel like it was authored in Japanese to begin with.

“I think this was once typically the goal of game localization, but when your audience knows and cares about a game’s specific origin, the ‘invisible’ approach becomes undesirable. The localization needs to honor the source text, and the difficulty there lies in balancing what the creators, the translators, and the players each might consider an honorable localization.”

from the unofficial translation of Mother 3

This balance is not something new to the process of localization, either. Clyde Mandelin is a professional translator who also happens to be at the center of the unofficial translation of Mother 3. He’s the “Mato” referred to on the patch page, and, in addition to his work in the industry as a professional, has also authored books on translation. In addition to This be book bad translation, video games!, which looks at translation mistakes in videogame history, and Press Start to Translate, which studied the state of machine translation back in 2017, Mandelin has also authored multiple titles in the series Legends of Localization, which are deep dives into specific games and their localization process. The second book in this series covers Mother 3’s predecessor, EarthBound, and is over 400 pages long. That is, in part, because EarthBound has extensive dialogue that was localized from Japanese into English for its North American release: a straight translation from Japanese to English wouldn’t have worked, if the 400-plus pages of explaining the decisions made for an entire game’s worth of dialogue and text changes didn’t already alert you to that. 

These localizations were enormous undertakings 30 years ago, and in this era of even lengthier and larger games, they’re no less enormous even with the more streamlined approaches to the work that now exist. Which is one of the reasons you still don’t see publishers throwing games that might not be a hit in other markets at the wall to see what’ll stick, and why some classic games, through services like Nintendo Switch Online or the ProjectEgg series, are released as-is instead of localized. Nintendo didn’t end up releasing all three of its Fire Emblem titles on the Game Boy Advance to international markets in part because their localization team was already overloaded with RPGs that would release worldwide. Some classic RPGs, first released in Japan, would take literal years before they arrived elsewhere, translated, which in turn made these hugely innovative games seem behind the time once new audiences got a look at them—part of why Dragon Quest lags behind Final Fantasy, popularity-wise, is due to early differences in the release schedules and localization work behind the two series. 

Even in the mid ‘90s, Square would routinely leave some killer titles in Japan, because they didn’t carry the Final Fantasy branding and weren’t considered worth the effort. The mainline Xenoblade Chronicles games are all reliable million-plus sellers now for Nintendo, and their primary non-Mario role-playing games; when the first one released for the Wii in 2010, Nintendo of America didn’t have any plans to localize it themselves. And when they finally did in 2012, it was using the European voice actors and localized script from Nintendo of Europe, suggesting that Nintendo of America still didn’t believe it was worth the time or resources to fully localize it for their region. 

To this day, companies as large and successful as Nintendo still need to employ third-party translation companies in order to keep up with the massive workload of localization (a fact that they’ve made “invisible to the user” in a different, worse way). 

So how, then, if this is so difficult and time consuming, do unofficial translators even go about their work? It’s important to recognize that these localizations may be unofficial, they may be DIY in that sense, but they’re performed by professionals, by real people with real skills. They are doing this in their own time for the literal love of the game. Vanessa Giselle Villeda Mendez, known as “AnonymusAxolotl” on Twitter, is one such unofficial translator. She’s best known in these circles for her work on Ys vs. Trails in the Sky: Alternative Saga, a crossover arena fighting game that falls somewhere between Super Smash Bros. and Capcom’s Power Stone as far as gameplay goes, featuring characters from Ys Seven and The Legend of Heroes: Trails in the Sky, released only in Japan for the Playstation Portable. 

Given that, at that point in time, developer and publisher Nihon Falcom wasn’t even releasing its early Trails games internationally—the worldwide release of Trails in the Sky wouldn’t occur on the PSP until nine months after Ys vs. Trails in the Sky, and the Windows release was three years after that, a full 10 after its original Japanese release—it’s no surprise that they, and any other possible international publisher, said pass to localizing a crossover fighting game. As the Trails games have grown in popularity, however, and Ys, too, has had a more mainstream revitalization, there was suddenly a noticeable hole where a translation could be. Which is where the work of those like Vanessa “AnonymusAxolotl” Villeda, as she’s been credited, comes in. 

As she told Paste, her interest in unofficial translations first came about by learning about the efforts that were going into the Project Exile translation of Fire Emblem: Thracia 776. It escalated from there when she began to play Trails games, however: “As someone who was playing in order, I kept being told to ‘wait for Geofront’ [a translation team dedicated to Falcom’s many untranslated releases), who at the time were working on Trails from Zero. Seeing Geofront members talk about the process publicly (namely, the late Scott Tijerina, also known as @KillScottKill) piqued my interest. Talk about style guides, character voices, etc., was all fascinating for me as someone both into literary writing and language in general (considering I’m ESL).” 

Villeda echoes Starr in that this work is meant to be of a quality where it feels like no work was even done at all. “To me, some of the best localizations are the ones that don’t feel like they’re there. The ones that make you feel like a game was written in English rather than translated from a different language.” She would get her first chance to put that belief into practice on a localization project when a friend considered translating Ys vs. Trails from Chinese to English: “I spent a few weeks poring over existing Trails in the Sky scripts to learn as much as I could about the character’s speech patterns. That little translation was born on the same server as the previously mentioned Project Exile, so we surprisingly had the people and tools to make it a proper thing, despite starting out as a casual project.” This project got the attention of Geofront, who had their own interested parties on board, and the two joined up to complete the project as one. 

from Geofront’s translation of Ys vs. Trails in the Sky

Localizing is not as simple as seeing an untranslated game and deciding you want to translate it. As Villeda put it, “For fan localization, broadly speaking, you need a few things to start out first. Full dumps of the game text usually come before assembling a team, then comes translation and editing, script insertion, the necessary graphics and video editing that accompany the game, programming and bug squashing, and all of those steps enter a long rinse and repeat process till the game finally works and reaches a desired state, or you give up, basically. Fan localizations are nothing if not a labor of love, passion, and endurance.”

If none of that outside of the “you need endurance” part makes sense to you, don’t worry. Unofficial translation is exceptionally technical work! Even the people being paid to do said work run into trouble: to bring up Dragon Quest once more, Dragon Quest V was outright skipped over by Enix America to instead begin work on its sequel, because the localizers kept breaking the game when they attempted to change the code. Which, if it wasn’t clear, changing the code is a requirement for translation work. 

Noah Steam, known for patching the 1996 Sega Saturn title Sakura Wars and currently patching its sequel, went into detail to explain these technical challenges to Paste: “The goal of course is to translate the game so that it’s playable by non-Japanese speaking people. But the question is, how much of the game actually needs to be translated in order to play it? For a game like Sakura Wars, the story text, battle menus, and system menus are the most essential. But what about load screens? Optional Minigames? Adding subtitles to movies? Credits? Miscellaneous minor images throughout the game that contain text? I’d say these are not necessary, but by patching them you can deliver a more professional game. My goal with Sakura Wars was only to fix up the bare minimum needed to play the game, but everyone within the team was so passionate that we ended up trying to translate basically everything we could find and in that we were successful. Sakura Wars 2 has the same goals.”

Not every game requires everything to be translated in order to work, but even if the translators aren’t tackling everything, it’s all still connected, and you’re limited to some degree by the original systems and boundaries the game was created within, as Dragon Quest V reminds. “To start, you have to figure out if the parts of the game within the minimum requirements are even patchable,” said Noah Steam. “For Sakura Wars and Sakura Wars 2, that’s the story text. So the first thing is to figure out how the game stores and renders its text data. Where is it stored? How is it stored? How is it rendered? Is it compressed? Is it encoded? Does modifying it affect the rest of the file it’s stored in? As in, does it result in shifting data around? Does the game support non-Japanese characters? Is there even enough space to re-insert English text back into the game files? Old consoles have a very limited amount of memory, and English is a much more character heavy language than Japanese is. For example, in English, we need to insert nine characters for “Commander,” while in Japanese it can just be a single kanji character.”

This isn’t just a hypothetical issue, either. “In both Sakura Wars and Sakura Wars 2, the format for the text had to be changed so that it could all fit into memory,” Steam explained. “As English is a much more character heavy script than a symbolic script like Kanji, once the text is translated, it ends up taking a lot more memory than the Japanese text. In order to deal with this, the format for the text data has to be modified. This requires fixing a lot of game code scattered all over the game’s code. Anytime we run into a bit of code that wasn’t fixed, the game will either print out junk data or crash.” 

You hear about single individuals creating unofficial translations, sure, but there’s a reason that teams like those working on the Sakura Wars games and Falcom’s titles exist. There’s being able to cross check progress on the writing and programming sides, but there’s art to consider, too. Sakura Wars, to make it all look more professional, had artists rework various image screens throughout the game to “match the look and style of the original.” Per Steam, “With Sakura Wars, people might not have realized that the intro logo was actually done by us because it mimicked the style of the original so well.” (Now that’s “invisible” localization.)

Even Ys vs. Trails, despite being more fighting game than RPG like its inspirations, required, per Villeda, “two graphics editors, three text editors (two handling main scenario text and one handling system text), one translator, one video editor, and two/three programmers,” who worked as a group for years, plural, to localize a game that had already been finished for its intended audience. 

Something like a mainline Trails title somehow requires even more, considering there are now 20 years of established character behaviors and speech patterns to keep in mind as localization occurs. All of this has to be kept track of, and it’s stunning that anyone finds the time to manage it without also receiving a paycheck for it to keep them going—there’s that “passion” and “endurance” again. Though, as Villeda pointed out—and as was the case with the eventual unofficial localization of Dragon Quest V—the fact that these are unofficial works on their own schedule means they’re free to aim for perfection, and not as constrained by timelines. She’d add, though, that “Unofficial work is also a matter of fighting you and your team’s perfectionism to a degree.” At some point, you’ve got to say the project is good to go. After all, there are other games that require unofficial translations.

It’s a tribute to the quality of the Geofront’s unofficial work on Trails from Zero and Trails to Azure that, when NIS America secured a deal to officially publish these in North America, the Geofront’s work on those games were used, and some of their members put to work finalizing the official release. Villeda, too, has now performed professional work on a Falcom title, in the form of the upcoming re-release of Tokyo Xanadu eX+ on the Switch, which will feature both new gameplay and a brand new localization, to boot. And it’s easy to see how that could happen, too, this switch from amateur to professional: the skills required to do this job right are significant, and doing anything besides scooping these folks up for this kind of work is the wrong call, given just how much localization work is out there to be performed.

Localization is intense, time-consuming work, performed by those with an array of skills that leave them up to the task. This is true whether it’s official or unofficial, but regardless of the origin, localization gives the rest of us more and more options to play, more of videogame history to sink our teeth into, leaves us with fewer gaps in the timeline and a broader understanding of games as a whole. It would be something if all of these titles were given official releases, sure, but that has never been the approach, and it’s unlikely to be the case in the future, either, not in a corporate industry scared of anything that isn’t a sequel. Luckily, we’ve got unofficial translators to pick up the considerable slack—remember that the next time you consider complaining about Nintendo not releasing a game that’s been left in Japan, as, if you just put in a fraction of the effort of translators on your end, then you’ll be able to play it, anyway.

Marc Normandin covers retro videogames at Retro XP, which you can read for free but support through his Patreon, and can be found on Twitter at @Marc_Normandin.