The Difficulties of Fansubbing: The Process

The simple act of fansubbing can be a long, drawn out process. I've been involved in this hobby since the second half of 2009. I had dabbled a little in it before then, but only a small bit as I subbed performances under the name Berryz Kyuuden Fansubs, and then when I joined Hello!Fansubs' trainee program. But it was from joining ICU-Subs in August 2009 that I really start counting myself as a fansubber.

And in this time, I've come to further appreciate just how complex the whole thing is, from the simple task of getting those translations to appear correctly on screen, to the difficulties in working with people you don't know, this is to be a short three or four part series on some of the complexities of the trade, as it were.

But, first up, I think I should explain to those who aren't familiar with how fansubs are created a little about the process itself. Other groups do things differently, and more on that in part two, but this is how things have worked in KIDS.

Getting Projects

The very first thing that any group does is decide what project to work on. This in itself can be fraught with hazard. When you are working as part of a group you have to be mindful of the other members in that group. This, of course isn't much of an issue when you are working solo, but not every member of your group may want to do the same things, and you try your best to make sure that you choose projects that everyone is eager to do, but there will always be cases where one or several members will say that they simply have no interest in doing a project.

How do you proceed? You have a project that you've been dying to work on, and when you let the others know about it, the translator says "That's too difficult" or "That's not a very good project". Do you make him do it anyway? Do you drop the idea - or at least sideline it? It has happened to me. I love stage plays, and they're the projects I most want to work on. But when I was in ICU, He-Hulk effectively filed the idea perpetually under "to be considered at a later date" and went on giving me projects which, for the most part, I did and enjoyed doing, but ones that were somewhat forced upon me when I hadn't gotten a chance to work on my own ideas.

Similarly, here at KIDS we haven't had much chance to do stage plays. Our translator, Lone!Wolf, isn't too fond of working on those kinds of projects, and the second translator we had in the past, CynicaltheCat, simply refused the suggestions due to difficulty and personal time constraints. It was only when we were approached by an external translator, snoboat, that I finally got my chance to work on one.

I don't blame them for being wary about such projects, and the limited opportunity isn't something that really annoys me, because there are other fun projects that we've done based on my recommendation. In every group there are ideas you love which never get done, and ones you weren't so fond of but you do anyway. Compromise is the basis for all human society.

One of the other major obstacles to overcome in beginning a new project is simply finding a raw to work from. Most of the time this is a fairly simple task and one which goes without any problems. But sometimes sourcing the video proves to be difficult, and there are several projects that we at KIDS have had on our list that we haven't been able to touch for this reason. One in particular has been on there since we started KIDS almost 12 months ago and we've been unable to secure a good high-quality raw for it. A high-quality one most likely doesn't even exist, as it was a TV rip from the early-to-mid 2000s, and at the time, rippers simply didn't think about how different the quality standards would be 5 or 10 years down the line.

Timing and Translating the Script

There are actually two ways of proceeding, once you have your files to work from. The way that seems to be more common among H!P fansubbers is to time the script, making sure that each line will appear and dissappear at the correct moments. But an alternative method is to go straight ahead and translate it, working the timings in later. Both are common methods, and I know that Hyakupa actually did a bit of both, but we generally opt for the first option.

Timing a script is a fairly simple task thanks to modern technology. Using more traditional methods, like the fansubs you might see on a bootleg anime VHS back in the 90s, were much more complicated and expensive. These days it's arguably the easiest job in the whole process.

Using a piece of subtitle editing software, such as the popular Aegisub, you simply match up the start and end of each line of dialogue with the audio, using the intuitive interface and functionality. The only real difficulties creep in when the project itself is difficult and requires intense focus and concentration. To use an example, a solo DVD like Momo-ONLY was a fun DVD and an easy one. You have only one person, and all of the dialogue spoken by that one person, it's easy to follow along, and you can time each line without much thought at all.

Now take an episode of Hello! Morning. Suddenly you can have 8 or 9 people at any given time and it becomes a lot more complicated. With it's fairly spontaneous unscripted nature, it's normal to have 4 or 5 girls all talking over eachother at once, and you have to listen very closely for the start and end of each line of dialogue in that mess - even more complex if your group colour-codes like ICU did!

And then comes the next stage, and the translator has to try make sense of all that mess. 

Where the timer just had to differentiate voices and say "Okay, she started talking here and stopped here," the translator has to say "Right, and this is what she said." It can be difficult to listen and understand 5 people all speaking at once in your native tongue, I'm sure you'll all agree. So now imagine trying to do it in a second, or even third language. Its an impressive feat by any means. There's a reason why many potential translators don't want to get tied down into working for a group like this. It puts quite a lot of pressure on you to keep delivering when, I'm sure, all you really want to do is enjoy it.

Editing and Typesetting

After the two stages in actually producing the script, comes the refinement phase. In KIDS, immediately after translation, script then comes back to me to be QC'ed. QC is short for Quality Control, and it's actual meaning differs slightly from group to group.

During this stage, I format the script. It's not really a necissary part of the job, and wasn't something I even did originally, but over time we have developed a presentation style unique to us, and I spend a great deal of time getting the script to match that. The more important parts of the task, however, is to simply check spelling and grammar, and to work with the translator to ensure that any liberties I take to make with the English script still convey the intended meaning from the Japanese. 

In an unscripted project, like a solo DVD, this can present a nightmare for us. The H!P girls use weird Japanese at the best of times. Couple this with the fact that they usually think as they talk, rather than thinking about what they want to say before opening their mouth, means everything is usually a mess of incomplete sentences, non-sequitors and non-words like "um..."

Typesetting follows up. This is another term that has varying meanings from group to group. In some groups, it refers almost exclusively to the choosing of fonts and colours and such - the typography of the script, if you will - and to the creation of special effects as 'scripting'. 

Here at KIDS, we simply merge these tasks together under the 'Typesetting' tag.

As stated above, the first part of Typesetting is to choose fonts and such. Thanks to the styles function of Aegisub, this is usually completed (for the dialogue, anyway) before the project has even been timed. So the real job is the onscreen text. Sometimes we'll have a project which doesn't have any of it and our Typesetter, Firren, will only have to code effects for the 'title screen' and staff list. And then there are times when we'll have hundreds of onscreen text to work through - all with varying levels of complexity. One might simply fade in and fade out, and one might explode onto the screen whilst rotating and then fade out. 

It's a long, arduous process of staring at the effects of the original, coding an effect to match it, and fine tuning and fine tuning until you are happy that the effects match the original onscreen.*

*If you'll excuse a little rant/tangent here, I feel compelled to point out one of our policies. We NEVER replace the original Japanese text with our own. I've seen a few groups make use of this technique where they felt they could get away with it, and I've always felt that this went against the grain of what fansubbers were. We add to the video. We make it understandable by people who don't speak Japanese. But by replacing the original text, you are changing rather than adding. This was something that Firren wanted to do when we first started out, and it's one of the only times in the history of the group that I've been stubborn - adamant, even - in my own point of view. 

Distribution

Typesetting is the last stage of the refinement process. After that, we're ready for release, and Firren begins encoding it.

Now, encoding is a fairly dull process which I don't fully understand anyway, so forgive me for kind of glazing over this stage, but I can only assume that it's a process that works similar to the Imp-Camera in the Discworld universe. 

Those software developers among you will know that the refinement phase actually continues on past this point in the Waterfall model. Once encoding is done, we test it by sitting and watching the completed video. If there are any mistakes that were missed in the initial QC, these are fixed and the file is re-encoded. 

The idea is that this is done as many times as necissary to ensure a mistake-free release. But some always sneak through. And the process can take so long that often if we do find mistakes we just debate on whether or not to release anyway if it's a fairly minor grammatical mistake (a missed apostrophe, for example), or to do another pass if it's a more major error.

Of course, there is one aspect to this phase which requires some thought. And that's how you want to package your subs. Hardsubs vs Softsubs. Both have their merits, but the one almost universally used by H!P fansubbers is the hard-encoded version (especially these days, where most people have a decent net connection, and the processing power of computers increases). In other fandoms, anime and dorama fansubs in particular, softsubs are often the preferred method, allowing for a faster release with no need to encode, as well as a faster download time as they download a file never more than a few dozen kilobytes. It does rely on them already having the video file themselves, but for people who downloaded or bought the raw straight away no longer have to grab the file a second time.

The biggest downside is that it doesn't offer anywhere near as much freedom on the kind of effects you are capable of. In hardsubs, it's possible to sub any onscreen text properly, and match the effects. In softsubs, the most effective way of doing it is simply to translate it in a note (either at the top or bottom of the screen), or to simply not translate it at all.

After this, the file is released, and provides perfect opportunity for us to sit back and enjoy a breather before completing (or, as the case may have it, beginning) the next project.

Which concludes my first post on this topic. Let me know what you think in the comments, and be sure to join me in my next post where I'll delve into the world of Collaborations and Politics.