It’s hard to explain what happens behind the scenes of a game translation. Most contents are bound by NDA and cannot be shown, and many tools and procedures are obscure even for translators.
Therefore, I decided to take the plunge and translate a small flash game, recording and explaining each step in order to build a video tutorial. For this task I will mostly use the memoQ translation environment, but also some other tools (links provided)
In order to play, click on the picture above or click here.
Like something out of the book 1984, The Republia Times is a small flash game that forces the player to become complicit in a government’s attempt to increase public loyalty.
You are a news editor tasked with choosing stories that best function as propaganda. You need to assure the world that Everything Is Okay in the country of Republia. And just to make sure you do as you’re told, your wife and kids are being kept in a “safe location.” So don’t you go getting any funny ideas, now.
Stories taking up more space on the paper will gain more attention, and the trick is finding a balance between frivolous stories the public is interested in, and political stories that highlight positive things about Republia and its government. A good mix will give the public what they want while also raising confidence in Republia’s disastrous regime.
Despite being a somewhat simple game, The Republia Times is easy to pick up and provocative—especially once you get to the twist.
Step 1: Planning with the style guide
How is the game? What people expect from it? What are its main hooks and how we can serve them well? A little bit of preparation always goes a long way!
Step 2: Importing inside memoQ
How do you get from a bunch of XLS files to a memoQ project? Let’s do it together!
Step 3: Crafting the terminology
Glossaries are vital for giving flavor to the text and keeping consistency across files and translators. Let’s automate it with terminology extraction!
Step 4: Sharing and assigning
Let’s check wordcounts, share files and team work with memoQ, Louandu* and Dropbox*.
(*No content is stored on Louandu itself *Dropbox transfers files through SSL with military-grade AES-256 encryption. We use it specifically because it’s safer than plain emails or ftp.)
Merging translations and reviewing them with the help of the DSpeech text reader.
Step 7: Quality Assurance and delivery
Let’s hunt down any possible mistake with memoQ, Xbench and Word before the final delivery.
The localised game
In order to play… come back in a couple of weeks!
Now that the translation is over, we are ready to play! Well, almost.. The game runs great, but I need a bit of time for tweaking (adapting video elements, replacing one of the fonts, testing…). But work needs me!
Please come back later on and I will add a little coda about this, the translated game and -Why not?- even a little kit so that you can make your own localised version!
This past 17th of March we held the second loctalk event. Ten people attended, mostly freelance translators, but also one project manager, one game developer and one student. Most of the event was held in English, but we had a couple of questions in Japanese too.
All the participants sat at the now usual large table and chatted along. A list of topics was provided, but conversation was more free flow than last time, also due to a couple of more general questions from attendees.
Some topics covered:
“Videogame translation industry in general. Fan translations.“
The question was raised about the main differences between fan translations and their professional counterparts.
In general terms, if game localization is a young field (15/20 years old for most languages), fan translations have surged even more recently, with digital distribution tying developers more and more closely with their online communities.
Paradoxically, this is the inverse process of translation as large, where the surge of full time professionals only came with the XX century.
One point raised was the different drive of fans and professionals: where the former only work on what they like, the latter are bound by their client’s requests.
From a fan perspective, this can lead “translators for hire” into being superficial and inaccurate.
This is partially confirmed by some translators, who admit dealing with material they don’t personally enjoy for sheer business reasons, like staying in touch with a client they are interested in.
A different perspective is that translators, for being invested in the translation and not in the product itself, are free of emotional baggage and thus can serve it better (for example through an MDA analysis)
Another point raised is that translation is very time consuming, and thus amateur translators can have only a limited “career”, both in the short (hours spent in a day, week, month) and in the long term (less and less free time due to professional and family life). This, in turn, limits their experience and thus the quality of their output.
Obviously, the topic is far from closed and it would be interesting hearing counterpoints from amateur translators.
Finding clients as a game translator
One attendee raised the question of how getting started in game localization, as even positive tests and interviews don’t seem to lead to much work.
One point raised was the existence of a “circle of trust”, where most companies tend to rely on a common pool of tested translators, rarely taking the plunge on new offers.
One reason for using a recommended resource is that it saves the time, risk and cost of vetting new translators.
Another reason is that, while there are no standard training courses for game localization, there are some general standards and customs to be followed – a professional training most clients simply do not have time for.
One suggestion for escaping this Catch 22 situation was for aspiring localizers to offer a fan translation as a sample with their CV (possibly in a form easily accessible online, like a youtube video). The growing “indie” scene could be a practical starting point for such a project.
“Do’s and Don’ts” on localization projects
What are the main recommendations from translators to project managers in order to ease the localization process? Three main ideas were raised during the event
The first and strongest request is for context. A simple note explaining who is talking to whom and about what in a specific string can save countless headaches.
The second is using a simple formatting in documents to be translated. Interestingly, many Japanese game writers work directly into Word and Excel, and thus enjoy a rich layout with multiple tabs, colored columns and so on.
However, this may become a hindrance during translation, slowing down even simple tasks (like replacing a term) and leading to inconsistencies and mistakes. A simple Excel file with one tab and three columns (string id, source and notes) is usually the most practical and common format for translators.
For developers, the best solution is probably building a central database to import and export the strings as needed, allowing both writers and translators to use the Excel layout they prefer (and building a single storage point for reference and reuse in the process)
The third recommendation for a smoother translation is to avoid stitching. Given a series of strings like “Program failed to save” “Program failed to load” “Program failed to update” and so on, developers often choose to store string as “Program failed to %s” and then task the software with slotting in the necessary word.
This sort of system is fine and elegant when dealing with only one language, especially a very “modular” one like English, but becomes a real headache when you try to shoehorn other languages -with wildly different grammar rules- into it.
In practice, this often leads to dozens of extra testing hours spent to “hack” in the new language, with results that are often unsatisfying. As inelegant it may seem, just storing all the possible messages in a long and repetitive list is probably the most effective solution in terms of time, budget and final quality of the translation.
There have been cases where code has been developed to deal effectively with stitching in accordance with the grammar rules of all the target languages. However, this is a big commitment, one that should be weighted against simpler alternatives (like using shorter modular sentences or icons and symbols).
Movies use shortened subtitles to ease reading. Why games don’t? Should they?
The general feedback was that this is an accessibility issue, not a localization one.
In other words, if subtitles are too long to be read, it’s the product itself that should be amended, either by making the subtitles shorter or by other means – for example Star Ocean let players choose between having subtitles disappear automatically or only after the press of a button.
Unfortunately, accessibility is a low priority aspect for many developers and such solutions are bound to be rare, like closed captioning for deaf gamers or special UI for color-blind ones.
Also, it was highlighted how movie subtitles often sacrifice too much content for sake of brevity, and that games should develop their own codes -more fitting for an interactive media- instead of always looking up at what other fields have done.
Midway through the event, one free “Translator Pro” license for the memoQ translation environment was drawn and awarded to one of the participants. Again, a huge “thank you” to Kilgray and Béatrice Compagnon for the giveaway!
One week after the event, we shared an anonymous feedback form with the participants, 40% of which responded.
In their opinion:
The base of the event is good, it just needs to be developed further (50%) or may even stay as it is (50%)
It would be nice having more people, but it’s not essential (75%)
Feedback on the location is predominantly negative, with most attendees suggesting to find a new one (50%)
The idea of giving out promotional licenses for localization tools is enjoyed unanimously
Most people enjoy the idea of recording the next event and posting it as a podcast (75%) but this might also discourage some from attending (25%)
The next event should be either three hours (50%) or two hours long (50%)
According to all interviewees, the next event should held in June
Did you find this report interesting? Are you based in Tokyo or know someone who does? Then why not attend the next event? To know the dates, keep an eye on this website or that of IGDA Japan, or enter your email address in the form below to be notified directly.
You know what is the worst situation to be in? It’s not being surrounded by nasty, mischievous people. I mean, it’s not great, but at least you can shout and fight back. No, being surrounded by good people with the best intentions, whose good rights hole you down in a messy, uncomfortable position – THAT’S much worse.
Case in point: one of the most common requests about this blog is that I add a portfolio.
And it makes perfect sense, after all, what have I done to be here pontificating all day about game localization? (Well, that specific question would actually remain, but a portfolio would make it slightly less tragic)
On the other hand, I’m a freelancer, a goon for hire taken off the internets, and most clients simply don’t want me to share project titles – simple as that.
Developers don’t want to see their multi-million franchises attached to someone they haven’t even chosen. Fair enough.
Agencies have to stay in business and, like it or not, the sheer information that I did project X and I didn’t suck at it is a business asset. Fair enough.
So here I am, surrounded by good smiling folks, and deep deep in trouble.
So, we could potentially have a “shock and awe” kind of portfolio – we churn at least 50 main titles each year, multiplied by four years of business, it’s sheer law of probability that we net some good ones.
Instead, I’m stuck with this:
Which is the list of games that have my name in the credits and thus can be mentioned without ruffling anyone’s feathers.
As most of them start to be kinda old, I must explain and contextualize them a bit, and thus I must write a personal post – something I am pretty uncomfortable about. So, here’s the deal: I will write this long post about my career, doing my best not to make it devastatingly boring, and then we move on and never mention it again. Deal? Deal.
My parents had a toy shop, and from the early days of Atari VCS down to the Super NES era, me and my brother played pretty much every title they sold. The shop even hosted the first videogame tournament in my city (Genova).
For my geek creds, I can say I used to cut out cheat codes from magazines and glue them on a massive notebook for future use. I liked how some of them allowed to peek “behind the scenes”.
At least now I have the excuse of being paid for this kind of things.
As for my studies, I have a Master degree in French and Spanish plus an ESOL Certificate of Proficiency in English and an IOL Diploma in translation. My final thesis was on the humour and language of the comic book “Astérix and Cleopatra”. (Which proved oddly useful for a JRPG adaptation a couple of years ago).
I was hired by Take2 interactive/Rockstar Games in 2003. The entrance test required a review of a recent game and, as I came from three months of complete unemployment, I had no console nor cash. So I bought a new PS2 with Max Payne, played non-stop on a flatmate’s TV, then returned the lot one week later just in time to get a refund.
Undeterred by my application letter (which ended with the embarrassing line “I want to go to the other side of the screen, I want to meet you”), they asked me to come for the interview. At the end of which my future boss asked me if I was really interested in the job, or if I was just trying to keep unemployment benefits.
But I didn’t sound like Spud, I swear
State of Emergency PC (2003), Global Star Software Inc.
This was my first real project for the company, which at the time still went by the hip name of “Take 2 Interactive Europe Quality Assurance office” and was located in the brownest office the East Midlands town of Lincoln could offer, haunted by the ghost of the game developer Tarantula studio, shut down in the very same premises.
Very brown indeed
There I joined the rest of the localization staff: one (1) Frenchman by the name of Antoine Cabrol. For most of the first year he reviewed all the French and German games, while I took all the Italian and Spanish ones.
Our main tasks were proofing the translations coming from agencies (usually Babel or Synthesis for Italian) and then checking them again inside the game (mostly for spacing or context issues).
The level of freedom changed with time (and faith in our actions). Initially, we had none. Text files were edited by the developers only, and all changes had to be confirmed by the original translators.
This taught me two precious lessons about editing:
You cannot change something just because you would like it differently. If you can’t explain objectively (and convincingly) why something is wrong, it’s a good sign it should stay as it is.
Editors cannot make mistakes, as this negates their very purpose.
And I learned that the hard way: when the translators reviewed my very first list of edits, they found about 25 typos… Ouch! As the british version of Word came with no Italian spellchecker, I ended up hacking in the relevant DLLs taken from home.
Space Colony (2003), Gathering of Developers
Firefly projects were strange. For most of the time you tested the demo and then, as soon as the demo was published, all the missions suddenly appeared in the testing build and you had to check them all in a couple of weeks!
For testing this, I had a spreadsheet with all characters/items/levels of the game and tested each single combination, which lead to one of my favorite bugs: when you sent the two punk girls Hoshi and Kia to the shrink robot, they would stay there forever and starve to death. Talk about problematic cases!
Nice project though, also thanks to the producer Sajjad Majid inviting us all in London for an end of project dinner!
Rockstar Games Double Pack: Grand Theft Auto (2003), Rockstar Games, Inc.
When Rockstar published Vice City, it had no Italian localization staff yet, but at least its translator came over a couple of weeks for a quick check.
But when GTA III came out before it, strings from the external agency (who had never seen the game) were simply added as is. So, when GTA III and Vice City were re-published on Xbox, I was tasked with reviewing both the Italian and Spanish versions of GTA III. Antoine did French and German finally had its own full time specialist: Chris Welsh.
Localization wise, I remember fixing mostly missed idioms, like a “greaseballs” wonderfully translated word per word in Spanish (“bolas de grasa”), just like Antoine had several instances of cat (as in “bloke”) translated literally (as in “cat”). In any case, at the time we were still required to open a detailed bug for every single edit and wait for the developers to fix it, so I had to limit myself to major issues and focus on playtesting. Actually, it’s the work project I enjoyed more as an actual game.
Drakengard (2003), Square Enix Co., Ltd.
Why Square-Enix chose not to publish this and flipped it to Take 2 instead? Because it was barking mad! Implied incest, giant flying babies, child soldiers to be killed in order to get a special sword… The list goes on and on…
Localization wise, we were quite excited to work with Square’s famed translators. Unfortunately we never received a text file with English and Italian side by side, which lead to some mistakes to be found only when it was too late to fix them (in particular, I remember some very odd stuff in weapon descriptions).
Max Payne 2: The Fall of Max Payne (2003), Rockstar Games, Inc.
The whole office was quite excited to work on this, after all it was the first new Rockstar title in a while (and the sequel to the very title I did my first test on!)
Localization wise, there were two main problems.
The translator really enjoyed using the semicolon (;) but the font didn’t display it at all – forcing a lot of strings to be rewritten.
Also, the story was narrated through a virtual comic book, and the game split words at the end of the line automatically, without following any grammar rules. If any programmer is reading this, please ensure that your font supports all standard characters, and never ever allow software to modify words. Never. Ever. Please.
Interestingly, the first couple of sentences of the Italian dub had a very strong hard-boiled tone, which vanishes pretty soon. I wonder how the game would have come up if they kept it throughout.
As the office still had no Spanish specialist, I was still doing double duty as a Spanish tester too, which lead to some very long days…
Manhunt (2004), Rockstar Games, Inc.
First announced by the press as “Rockstar Games’ survival horror”, it was a surprisingly subversive title. Like wikipedia says:
The game's story follows a supposedly executed death row inmate who is forced
to participate in a series of snuff films for former film producer and now
underground snuff director, Lionel Starkweather.
Which meant that players were forced to kill enemies in the most gruesome ways in order to appease the director.
As players would gradually become more and more skilled in such killings, they would actually start to enjoy executing them, finally leading to a disturbing realization: the real monster in this game is you, the player, dipping in your primeval violence as soon as soon you were given the (virtual) chance.
At least, this is what would happen in normal conditions, playing a couple of hours, once in a while. We worked on it the full day, plus overtime, for months. We didn’t dip in primeval violence, we drowned into it!
My personal situation was even worse as I was still doing double duty for Italian and Spanish, which meant that I was assigned two playtesters. One would play the whole Italian version with me on the side to check all the translations, while the other would videotape Spanish. As soon as Italian hit the end screen, I would take the Spanish tape and start over…
One of those evenings I went to my favorite pub and the neon lights suddenly broke: for a split second I thought there might be a killer behind me!
In the meantime, the staff had grown considerably. This meant the welcome end of my Spanish testing duties, but also moving to a new building, in a business park far far away from pretty much everything. As far as I know they are still there.
Mafia (2004), Gathering of Developers
Similarly to what happened later with Vietcong 2, this was a port of a previous PC game and thus most of the localization was already recorded and couldn’t be changed.
Luckily, the Italian version was quite good, despite its huge size. A feat that probably cemented the relationship between Take 2 and Synthesis for the coming years.
Wings of War (2004), Gathering
A budget arcade shooter set in the first world war. I remember it for two main reasons.
The first one was the Italian translator complimenting us for the quality of the proof. Apparently one year of having each single edit vetted by translators was bearing fruits!
The second was being joined by an Italian colleague: Stefano Moretti, a former Sony tester who moved over with the main motivation of getting into game design.
A motivation that became immediately clear in this project. For balancing purposes, we were tasked with playing the title through, noting down our scores for each stage and eventually providing some feedback.
It quickly turned into a grind, recording numbers for numbers’ sake, when our boss suggested a more proactive approach. Stefano went in full overdrive with ideas, and soon the ”suggestions” column massively outgrew those with the scores.
We sent it to the developer… and were thoroughly scolded! “Where are the scores?!” they wrote “All the other forms are perfect, but this one is awful. It sticks out like a sore ass!” Ouch!
Vietcong: Purple Haze (2004), Gathering
A very strange project. In retrospect, I think that there were no major language projects and the management tried to use the localization team mostly as normal testers.
Not much to do on the language front anyway, as most of the dialogues were already recorded and could not be changed.
At least the producers were very kind, making custom beer bottles with the game title and the caption “Made with the blood, sweat and tears of…” followed by all our names. And I still have the god-awful promotional t-shirt!
Grand Theft Auto Advance (2004), Rockstar Games, Inc.
I loved GTA:GBA! Quite surprisingly, we didn’t use some arcane and madly expensive proprietary hardware for testing but… a freeware PC emulator (Visual Boy, I think)!
This meant that I could start a level, quicksave, try any sort of mad stuff and then quickload back. Which lead to one of my favorite bugs of all times: continuously spray-painting your car slowly ate the memory of the game. Bit by bit, everything on the map would start to disappear: pedestrians, cars, buildings… Finally, you ended up floating in the middle of nowhere, until the whole game crashed with a black screen.
Not much to mention translation wise, but we had our moment of glory when we sent the game for approval by the notoriously picky Nintendo offices: first time pass, not a single issue found!
Red Dead Revolver (2004), Rockstar Games, Inc.
A special project, both on the testing and localization side.
Testing was complex due to the adventurous history of the title. Red Dead was a canceled Capcom game, which Rockstar took over and reinvented when they bought its developer (Angel Studios) and turned it into Rockstar San Diego.
This lead to some oddly empty locations (especially the towns, which I guess played a much larger role in the original design) and some understandably unpolished code (in the early builds you could even plug a light gun in the console and trigger an half finished shooting mode!)
As this made development rather complex, halfway through the project we were asked to step up for localization. We still had to record each edit as a bug and revise it in the following version but, for the first time, we were handed the source files and requested to implement all the changes ourselves.
This made the process much smoother and, most importantly, was a strong sign of confidence in our work. So, despite it being a pretty straightforward translation, I remember Red Dead pretty well for the sheer pressure of NOT MESSING UP THE CORE OF A TRIPLE A TITLE.
Outlaw Golf 2 (2004), Global Star Software Inc.
A surprisingly memorable title, for several reasons.
The first one was how complex the translation was. A translator once told me that the the work of famous authors is sometimes so rich and beautifully built to be a breeze to translate, while “simple” popular literature can be so reliant on a specific context and culture to be almost impossible to convey.
Alas, I can’t vouch for famous authors, but the sheer number of puns and references indeed made this one a real nightmare.
The second was system messages. While the developers made a great choice by allowing us to browse system messages with the press of a button, saving countless hours of testing, they made the terrible mistake of building them with stitching.
Vital messages like “Insufficient free space on memory card” or “No memory card inserted”, which have mandatory match-or-be-damned translations, were stored in chunks and built on the fly by slotting in the necessary bits. Which works well in English but is a complete nightmare for romance languages.
The third and last reason is a wonderful bug I found out. If players wrote %s or %d anywhere in the game, it would morph into a longish hexadecimal string like 536f206e6572647921 (%s and %d were the codes used to store variables in the game).
So I decided to test it on the online rankings, naming myself %s%d%s%d%s%d%s%d… And it destroyed the layout of the rankings to the point that they had to ring Microsoft and get our (testing) data reset. Xbox live destroyed, achievement unlocked!
Grand Theft Auto: San Andreas (2005), Rockstar Games, Inc.
A huge project, that redefined everything.
First of all, it had to be tested 24/7. My Italian colleague Stefano Moretti and the other “junior” testers worked by day, while the senior staff was tasked with nights.
We would start working at 9pm in the evening, have a briefing, stop in the middle of the night for “lunch”, then finish at 9am in the morning with a debriefing (often eating the leftover pizza, promoted to the rank of “breakfast pizza”).
For a whole summer, the office never locked its doors – and my friends in town started thinking I had returned to Italy for good. It wasn’t too bad, but it sure didn’t bring a relaxing atmosphere, or a good smell, for that matter.
Translation wise, we had to do a lot with very little.
Almost every day we would receive a new version of the source text, with several changes and edits placed through more than 16 thousand lines.
On the other hand, our work document was just a very long TXT file, with headers marked with square brackets!
Considerably concerned that deleting even one by mistake could potentially break everything, I started collecting versioning utilities online, which would end up being used by the whole office.
On a strictly linguistic level, the constant rewrites meant this is probably the Rockstar title I was most involved with as an “added translator”. One choice I am particularly proud of (although it required a lot of effort) is how we tried to keep subtitles short, a bit like in movies, as a response to players who complained struggling to read them while playing.
Testing-wise, we had to check each single string on screen, all 16 thousand of them. This means all the cutscenes and all the system messages, but also all the different results of a dozen of different mini-games, down to the twenty or so random sentences a character tied on your bonnet would shout in that specific mission and so on…
I’d say the most challeng part was keeping the necessary focus and discipline in the long term, as mistakes are always around the corner!
FireFly Studios’ Stronghold 2 (2005), 2K Games, Inc.
Like Space Colony, this is a game that substantially went overnight from a tutorial and a couple of levels to a rich and complex campaign, to be tested very fast!
Translation wise, I remember it for the dubbing genius of Pietro Ubaldi whom, given a list of “medieval” multiplayer insults (marrano, gaglioffo, cicisbeo, etc), ran with it, creating some hilarious voice overs.
I also remember it for being the game that allowed my Italian colleague Stefano Moretti to (briefly) realize his dream. Using the in-game editor, he created a couple of levels and used them as a portfolio, landing a game design job at Gizmondo. Stefano passed away in a car accident in July 2006, I’ve been told his English years were his happiest.
(Stefano Moretti and Gabriel Bienzobas Mauraza)
Back in Italy
I’m credited on a few other Take 2 / Rockstar titles, but either they contain very little text (Motocross Mania 3), or my involvement was minimal (Midnight Club 3: DUB Edition, Serious Sam II, Top Spin, Kohan II: Kings of War).
After about two years and a half of career, I had learnt two main lessons.
The first one is that localization is a war of attrition. Take too many risks or burn your energies too fast and you WILL make mistakes, which players will sneer at way more than they will ever appreciate your prowesses.
The second one is that translation, far from being the end of localization, is just the beginning of a very delicate and complex process. Any effort a translator can make to ease it is probably more than welcome.
I moved back to Genova in 2005 and started working as a freelance translator, first for car maintenance manuals and later for video games. Suddenly I had a much larger role in localization and, thus, my name would disappear from anything I worked on (for the reasons detailed above)
The area where I worked and lived. Hint: not inside the stadium.
Two Worlds (2007), TopWare Interactive GmbH
As the Italian publisher Newave Italia was searching for a translation team, fellow translator Francesca Pezzoli contacted me, Raffaella Brignardello and Mirella Soffio and offered our services through Marco “Cavaliere Ombra” Macciò.
Working directly for the publisher, we had the honor to appear in the credits… as co-authors of the manual (Italian taxation is much lighter for writers than for translators)!
I also remember this title for being invited to the official presentation (and being quite bemused by the questions of the press). Speaking of which, reviewers seemed to enjoy our work:
"(Two Worlds) remarkable localization and equally excellent interface offer playability
at top-notch levels"
"Dialogues are still in English, but subtitled in an exemplary Italian. You will not
miss the sketchy localization of other titles."
Iniziopartita.it - Two Worlds
Nintendo, Square and Blizzard are quite different from pretty much everyone else in terms of focus and procedures, even in the smallest details. Take 2 asked me a game review in order to apply, Square-Enix asked me this:
Creative writing assignment
Topic: You are a character from a Square Enix game who has been magically teleported
to Japan. Write a short story about your adventures in this new land.
As Kotaku summarized it, “Want To Work For Square Enix? Write Fan-Fiction”! Anyway, I didn’t really need to make up much, as Drakengard obviously had one ending based around that very premise, in which you flew your dragon around Tokyo Tower and fireballed Migs (never, ever think too hard about Drakengard).
After passing the test and subsequent phone interview, I became the Italian translator for Valkyrie Profile 2: Silmeria through the Unicon Products agency.
As I worked from home, they actually went through the effort of mailing me a Brady Guide, a copy of the English game and one of those insanely expensive development PS2 to play it on. Too bad I received the lot only a couple of days before translation began!
Translation wise, I tried three little experiments.
First of all, I kept subtitles short and movie-like, a bit like in GTA: San Andreas, although this time the aim was to make things more cinematic.
Second, as an anonymous fan had ripped and posted all the cutscenes on Vimeo, I based most of my work on the actual footage.
Finally, I tried to make the translation easier to implement by limiting stylistic elements that I knew would bog down testing without giving much to the title.
I also remember this game for a very odd bug: we had to ensure that each single item had a different translation because if even one name was shared, for example between an attack item and the equivalent magic, the game would remove it and misalign ALL the other strings!
Hopefully, we worked hard and were rewarded by some good reviews in the press:
"Valkyrie Profile 2 (...) comes fully translated in good Italian, dubbed in English
but with an effective subtitling."
"A masterpiece enhanced by a solid Italian translation."
PlayStation Magazine - Valkyrie Profile: Silmeria
The Last Remnant (2008), Square Enix, Ltd.
The odd thing of being limited to the titles I am credited for is that I need to leave out some major projects and give prominence to companies like Square – which I only did two projects for.
But I can’t deny that The Last Remnant had a tidal impact on my life and career – something I’m pretty sure very few ever said about that forgotten title!
Working on this project was a real culture shock. First of all for being an in-house assignment, in Tokyo, for six month.
(Yes, in the same building as Pfizer. No, we didn’t get freebies.)
Second of all, for the complex dynamics of a company with hundreds and hundreds of employees, with wildly diverse backgrounds.
Finally, for Square’s unique localization style. Some things were brilliant and became part of my own process, like collecting huge amounts of reference before starting, some were a bit bizarre, like having a dozen of translators (none of which had any artistic background) providing feedback for the new system fonts.
In the end, I think we did a pretty decent job, which the press seemed to agree with.
When the project finished, I stayed in Japan with whom would soon become my wife, working as a freelance from our small apartment in the north of Tokyo.
(Here. No, not on the highway. And not on the lettuce fields either)
Street Fighter IV (2009), Capcom Co., Ltd.
Oddly enough, after about four years as a full time translator, I was offered this brief gig as a pure localization tester, in Osaka.
Back in the shoes of an humble tester, I didn’t have much say on style and wording, but I did my best to polish everything.
Still, I remember this project for two main reasons.
The first one was receiving all the in-game strings printed each morning, like a phone book sized manual. Ever wondered why Japanese files often come on countless tabs with a complex layout? That’s probably why.
The second reasons was finding a 100% crash bug on my second day of work. As I had to check all defeat sentences, I set the rules in order to be beaten in a matter of seconds. To my surprise, the game would crash if you did that many times in a row. I know, it’s just a natural talent I have… (I mean for causing bugs, not for losing)
Naruto: The Broken Bond (2008), Ubisoft, Inc.
A very special project, as it was my last big title as a single translator, and it laid the foundations for team GLOC.
In a nutshell, I took the rich and detailed approach we used for Square’s translations and tried to streamline it until it would fit within the much smaller budget and timeframe of freelance jobs.
In the case of Naruto, my efforts went into three directions.
The style guide was simplified and rationalized, so that it could be compiled in half a day using both direct and online resources.
The core terminology was extracted on a statistical basis and carefully checked against the anime, as some reviewers critized some departures in the previous episode.
Finally, the whole cycle was closed by an internal QA with Xbench, ensuring that style and terminology were applied consistently throughout the project.
Its was complex, as I was testing a brand new workflow while trying to give this huge franchise the translation it deserved, but I was very happy with the results.
In a way, it’s the most personal translation I ever did, as it coupled the rigor of the “team” framework, with the clarity of a single voice (mine).
An opinion that seems to be shared by the critics, which gave some good reviews.
While working on Naruto I thought that the whole style guide/terminology extraction framework would have been ideal for working as a team, but I only took the plunge one year later, moved by three factors.
While games were getting bigger and bigger, deadlines seemed to get shorter and shorter. Like it or not, titles you could translate alone were a dying breed.
Also, games started to come with lots of small ancillary texts for website, updates and DLC to be translated very fast, making my 8 hours time shift with Europe increasingly problematic.
Finally, Lehman’s bankruptcy lead to a very slow 2008, so it was really time to try something new.
In January 2009 I joined forces with Antonio Vaccarino, which had been my project manager for Naruto, and Matteo Scarabelli, which I knew for reviewing his (pretty good!) entrance test for a client, and we embarked on our first team project… a shockingly awful budget game about astrology!
Still, the client was satisfied with our work and we started from there. About one year later Antonio left in order to work for Ubisoft and was replaced by Elisa Di Fiore, former editor at PlayStation 2 Magazine – leading to the current team.
The rest is recent history: in 2010 we were invited to present the mechanics of the team at Localization World with our “Joe Freelance VS the Mammoth Game Translation” speech , which also lead to the creation of this blog – and our cartoon personas.
A couple of months ago Team GLOC entered its fourth year, celebrating more than 1,000,000 words translated each year with our own translation system “Louandu”.
And a couple of days ago I reached ten years of career, leading to this endless post. A big thank you to everyone and hear you again in 2023!
Like many translators, we often wanted to see the Pegi ratings explained in a brief and clear way. And finally we took action ourselves with this brief infographic, summarizing all you need to know about how games are reviewed and rated all across Europe. Enjoy!
Video game subtitles are increasingly important. As more and more titles aim for a cinematic experience, which techniques can we learn from the movie world? To find an answer, let’s review the “15 decluttering tips” article by specialists Bianca Bold and Carolina Alfaro de Carvalho in a game perspective.
Hi there! No localization post this week, as we are taking some time off to rest and hang out with our families, but we will be back in January. Best wishes of a Very Merry Christmas and Happy New Year from Alain, Elisa and Matteo
How the Last Guardian ended up in our background, you ask? No idea! But apparently he’s waiting for a certain someone at Eurogamer.it …
This last Saturday the 17th we met in Hatagaya for the first Loctalk networking event with the Tokyo videogame localization community.
How did it go? I’d say pretty well! Ten people attended, mostly videogame translators (both in-house and freelance), but also three project managers and one game developer. Most of the event was held in English, but we had a couple of questions in Japanese too.
All the participants sat at the same, large, table and chatted along, making their own question or picking one of the “rants”, humourously aggressive statements meant to spark controversy.
“Do translators use Computer assisted translation tools? if so, which one?“
In the majority of cases, attendants did not use any CAT tools, as they are not considered efficient in gaming, where the same term may need different translations depending on the context. Also, the fuzzy matches/Trados split rate pricing system is seen negatively. Two attendants, however, do use and recommend them. One has used Dejà-vu for many years, the other one is Alain, who waxed lyrical about memoQ’s term extraction for a good ten minutes.
“‘Are you a fan?’ Sure, I just wrote a Fanfic, would you like to to read it? that’s how you make a great translation, isn’t it?“
Again, mostly a monologue by Alain. If any of you listened to the Outcast.it podcast, you know how easily it can happen! Anyway, the point was that some clients sometimes approach translators asking first if they are a fan of a specific topic or series. A legitimate question, but one that somehow negates translation as a profession. The translator is not seen as a specialist in the craft of translation itself, but as an interchangeable “fan” whose real skill is simply to be already knowledgeable in the topic at hand.
Some rightly argued that this kind of “fan” knowledge is vital for some titles, like a long-standing RPG series, where the jargon might even go against the usual norms of the language. However, assuming a professional cannot interiorize this kind of styles and terminologies into his workflow opens somewhat disheartening perspectives, like sidelining older and more experienced translators simply because they cannot really be part of a specific fanbase anymore!
Community translations are a different matter. While the idea of a large AAA title relaying on amateur translations was unanimously frowned upon, and seen almost like a fraud, most attendants didn’t mind about micro-niche titles, that might otherwise have no translation whatsoever, being translated by their own audience.
“What is the usual background for a Japanese to English translator?“
In most cases, Japanese to English translators seem to be bilingual that for a reason or another got in touch with game developers. Former teachers, creative writers, programmers… The paths are many but almost never direct, as Japanese to English translation courses and schools are very rare.
There is a higher number of FIGS localizers with an academic background in translation but, even in this case, few have a specific preparation in narrative.
“Why are translator doing this job?”
One of the project managers enquired about the typical schedule of freelance translators, for example if it was better working on long or short projects. The answer was that it doesn’t matter that much, because long projects might give a steady income, but freelance translators still have to accept shorter ones on the side to keep in touch with future clients.
As an aside, a translator working at a major Japanese developer mentioned that’s one of the main benefits of their current situation, as they receive a salary for working in-house, but they still keep the freedom to carry freelance activities on their own.
The project manager sincerely wondered what motivates people to become a freelance translator despite this general feeling of uncertainty, and one of the answers was the perspective of improving conditions in the coming years, brought upon by the growing influence of gaming as a mainstream media.
“And what makes a translator prefer an agency or another?”
This tied to the following question, which was about the main qualities freelancers look for in an agency. Besides the obvious desire for timely payments, one of the requests was for educating clients towards a more collaborative approach towards translations or, at least, some sort of protection against the most disrupting situations.
“Do you usually know who is translating your game?“
While in-house translators usually get mentioned in the credits, freelancers generally aren’t. Provided reasons vary from the “black box” effect of agencies withholding the name of their providers, to the common trend of external vendors being forbidden to link their name to the projects they worked on, to the extensive editing game translations receive before reaching the last version making it difficult to give a clear authorship.
“Why are developer files so messy?”
The localization process often seems to lack planning, starting with poor directions (non-translatables, character limits, context) and finishing with a constant and unmanageable drip of one line edits weeks or even months after the last delivery.
One translator mentioned the interesting paradox of working on a text supposedly “Final, but subject to changes”.
Unfortunately, in the complex world of game development, this really makes sense as a text may well be finished as far as the writer is concerned, but may still need to be tweaked by the legal department. After all, with the spread of iterative and incremental agile software development methods like Scrum, everything is in a state of flux and small, daily edits are here to stay.
After the event, we shared an anonymous feedback form with the participants, 50% of which responded, In their opinion:
The base of the event is good, it just needs to be developed further (80%)
It would be nice having more people, but it’s not essential (80%)
The next event should be more structured, with longer prepared talks (40%) or stay as it was now, with a mix of prepared talks and spontaneous chatting (40% SPLIT VOTE)
The next event should be two hours like this one (60%)
The next event should either be in two months (40%) or in more than three months (40% SPLIT VOTE)
Expenses for the next event should be paid by each participant, up to a fee of 1000 JPY (80%)
Did you find this report interesting? Are you based in Tokyo or know someone who does? Then why not attend the next event? To know the dates, keep an eye on this website or that of IGDA Japan, or enter your email address in the form below to be notified directly.
I am really amazed by the good quality of the translation of all the material and the rapidity in making it. It is perfect! Congratulations! This is the first time we see such a high quality translation. The terminology of the game was really difficult sometimes, but it has been perfectly localized. Really a very good job. — Developer