As mentioned in my previous post, all our future releases (and re-releases of older series) will feature editions. Let me take a few moments to explain what editions are in case you don’t know.

Before we get to editions, I’ll first explain how chapters work. There are two types of said chapters: “normal” chapters which link to a specific time spot in the same file, and “ordered” chapters, which can link to a time spot in a different file. This is what we did with To LOVE-Ru and some other series. We cut out the opening and endings, made them become separate files and then made sure there’s a chapter entry in every main episode that linked to those openings/endings.

Now, editions group a bunch of chapters together (either “normal” or “ordered”). You can have multiple editions, so you can have multiple groups of chapters. This comes in handy if, say, you want to choose whether to watch our anime with hardsubbed or softsubbed karaoke. We’ll just release two versions of the OP/EDs, one with hardsubbed karaoke and one with softsubbed karaoke. You then just open the main episode and can choose what you want your media player to play back by picking your edition of choice. Real world examples:

Picture #1 and #2: Open the episode, right click on the white icon in the task bar (it’s called Haali Media Splitter), select your edition! (click to enlarge)

Alternatively, in MPlayer you can just use this command: mplayer file.mkv -edition 0 (edition 0 loads the first edition, edition 1 the second one, etc…)

Specify the edition (set of chapters) to use, where 0 is the first. If set to

-1 (the default), MPlayer will choose the first edition declared as a default,

or if there is no default, the first edition defined.

We also included a third option that will make sure no OP/ED gets loaded at all.

These screenshots also show that each edition contains its own set of chapters (e.g. “No OP/ED” won’t load the Opening and Ending chapters).

We can still turn this a notch higher by creating a fake file that just has a few frames of black screen. The file itself weighs in at just a few kilobytes. By making use of editions, though, we can make it load all six episodes of TLR OVA for example. All episodes will get loaded, resulting in a timeline of more than 2 hours.

Picture #3:

Hitsuji did a really nice job on Haruhi in 2007 by the way, and they’re pretty much the only ones I know of that went this far with matroska and editions (yet!). Saizen did some nice things with their Soul Eater releases (which got dropped after episode 4). They joined together the normal TV airing and the late night show by using editions.

If i got you interested now and you want to know more about what is possible with Matroska, I highly suggest you check Fluff’s post on his blog.

Speed
Many people actually rate groups higher on AniDB when they release anime shortly after the episode aired (with “reasonably” good translation, timing, and all that stuff). People are more forgiving about inconsistencies or errors they spot in such releases “because they’re fast”.

So if you’re fast, you’re allowed to have some errors in your release and (nearly) nobody will blame you. On the other hand, if you’re slow, people just expect high quality subs. The rating of that group will go below the one of the speedsub, even if the quality would be the same. Shouldn’t the groups have the same rating then? Apparently speed matters. (Rating. There are rating systems like the one on AniDB of course, or it could also be just your personal rating of a specific group. The above statement is confirmed by Kalessin’s comment.)

Consistency
This is actually regarded to as really important by many, many people. Consistency simply means that you stay consistent with editing, timing, etc. throughout the whole episode, the whole season, or even all anime the group has ever worked on.

Let me show you a few examples. Consistency means…

…you watch out that particular people in the anime always use the same particular phrases when they say the same phrase in Japanese.
…the styling is the same across all episodes.
…to always use the same timing style, even if different timers work on the project.

People can be more forgiving about strange translations if they’re just consistent. For example, some group translates Nee-san, Onee-san and Onee-sama all as just “Sis” (Nobue-oneechan –> Sis Nobue) — while this may not be perceived as the proper way to translate them, if the group always sticks to “Sis” many people will not care since it’s at least consistent.

Reputation / Market Dominance / Prominence
Overrated. Many groups have a high rating on sites like AniDB, people commonly acknowledge them as quality groups, but in reality they’re not really that good. Especially people who have been involved in fansubbing know that plenty groups which fit that profile do exist. The point is, they do not produce quality (nearly) as high as people say they do, but they have a large fanbase thanks to e.g. releasing fast with seemingly good quality while it’s not really quality (simplification of the translation etc.), and so the group will get praised for doing average releases.
The same goes for “quality groups” (when they take too much time). For example, a group was a quality group some years ago, but now most of the core staff disappeared and the rest of the staff desperately tries to keep the group alive, even though if that means that the releases would be “just” average or subpar.

Underrated. Reputation can mean something else too. If your group is known for fast releases you’ll be called a speedsubber and many people will avoid you because they have that common attitude that speedsubs equal bad sub quality. Even if your releases are close to or already considered quality subs, there’s a high chance it won’t get acknowledged (any time soon at least).

Quite a lot of people don’t even compare groups and just go with the one that has the best “reputation” or rating on AniDB.

A more up-to-date example would be a.f.k. subbing Haruhi S2. Approximately half the people who were watching Haruhi when it was airing waited for a.f.k. to release their sub because they did S1 and they were reliable. This also leads me to the next term:

Reliability
Doesn’t matter if the group is the best or the fastest. Many groups just drop anime out of the blue, promise to pick up anime and then release the first episode 2 months later just to tell everyone that they lost interest in working on it and are going to drop it… So reliability can indeed be something very important. When you know that a group promises to not drop any anime and finish it no matter what, like Yabai, that’s a big plus for some people (like Hamlon).

But mainly people understand under the term “reliability” groups that release at the same time every week in the usual quality.

Selection
When a group does a single genre of anime, they earn an image of doing only that genre. Others, who select a variety of anime from different sources, such as shoujo, mecha, or even sport, can develop and expand the group’s fanbase. Is it only the responsibility of those on staff to choose the series? And what about the series that fall through the cracks? Should we cooperate more as fansubbers? How important is it to offer a variety of different anime?

Softsubs VS Hardsubs (Hardsubs meaning that karaoke and TS is hardsubbed, dialogue being softsubbed, though!)
“They’re no good cause they hardsub their TS and karaoke.”
“Even though they’re better, they hardsub, so I’m gonna choose this other group which does full softsubs.”

These and other arguments are often mentioned. Doing softsubs has become a new trend. Quite a few groups switched from hardsubs to softsubs, new groups often do only softsubs right from the start. Surely one positive thing to mention is that you can skip some parts in the subbing workflow which ultimately results in the group releasing the anime faster.

Viewers often state they prefer softsubs because they can e.g.
– change the subs if there’s something they don’t like
– manually fix editing errors or timing errors

There’s are some big buts too, though.
– softsubbed typesetting is very limited compared to After Effects (e.g. smoothly moving signs and zooming, 3d effects, special effects)
– softsubs can be very CPU-intense, dragging down the overall performance and making the video unplayable in the worst case

I’m not intending to start a flamewar Softsub VS Hardsubs here. It’s just another factor many people  regard to as very important. Softsubs? Hardsubs? Or a mix? What’s the best solution to make as many people as possible happy?

I guess just as you can hype movies or more recently even computer games, the same can apply to anime — and fansubbing groups. Though sometimes the hype turns out to be true, too. My question to you is quite simple,

What do YOU think is most important in a fansub from the abstract point of view? What’s maybe equally important or the 2nd most important thing?

Of course, it’s more convenient to have someone synchronize them for you—also called timing—so that the subtitles appear right at the time the person is actually talking. Checking the original translation and an additional editing stage wouldn’t hurt either. Nobody is perfect. Mistakes are bound to happen. Have more people work on it and the amount of errors will be minimized.

Here’s what I want to know:   What’s the most important thing in a fansub for you?

Translation/Translation Checking/Editing involves a person fluent in Japanese who translates the dialogue/signs into English. Someone else checks the translation for errors,  a 3rd person will try to make the English script as error-free as possible and additionally try to make the “flow” of the English perfect.

Timing
Perfect timing means you won’t even notice that there are subtitles—you’ll think of them as a part of the video. Bad timing interrupts your pace of reading and can be very annoying.

Typesetting/Karaoke
There’s nothing wrong with having fancy karaoke or advanced AFX typesetting if it fits into the anime. If it does not stand out, or if you think it’s part of the video even, then and only then, is the typesetting done well. If a karaoke can make a horrible song sound really amazing, then the karaoke is good! :P

Quality Check
Even the best in fansubbing make errors—and not as few as one would think. In order to avoid releasing error-ridden episodes there are Quality Checkers. There’s (nearly) no people who can be good at spotting “all” kinds of errors. Some are good at spotting timing errors, others at spotting editing errors etc. So basically, there should always be at least 2 or 3 people QCing the episodes.

Encoding
Watching crystal clear anime on a 42″ display must be really awesome. It sucks if the encode is bad. If the encoder knows what he’s doing, you’ll receive the best video and audio quality with the best compression ratio possible.

There’s also other stuff like subtitles styling which deserve to be mentioned. If the font is not readable—especially when it’s hardsubbed (fused with the video)—it can be a pain in the ass to watch the episodes cause there’s no way to change it.

I’ve talked to many people, read many comments or reviews on AniDB, and I’m also speaking from my own experiences. Everyone seems to have their own definition of what’s “good.” What I’m interested in is what the majority of you think is the most important in a fansub.

I’m interested in your opinion. Take part in the poll! State your opinion! Why do you think X is most important? Don’t forget to state your reason!

Edit 21.10.2009
Additionally, what do you think about karaoke? Is karaoke fine no matter if it’s harsubbed or softsubbed? Does it have to be fancy or should it just fit into the video (no matter how simple the effects are)? Or would you prefer no karaoke effects at all and just timed lines so you can sing along? Or just cut out the opening and ending and release two versions — one with karaoke and one without effects (ordered chapters).