2010-08-29, 13:16 | Link #81 | |
Far out, man!
Fansubber
Join Date: Jul 2008
Age: 40
|
Quote:
|
|
2010-08-29, 16:11 | Link #82 | |||
Excessively jovial fellow
Join Date: Dec 2005
Location: ISDB-T
Age: 37
|
Quote:
Quote:
Moving on to the actual topic at hand, however... Quote:
Also, get a real computer. Re: the entire "you have to take time to make a good fansub" debate: it's bullshit. The vast majority of time spent between airing and release of most fansubs is spent doing absolutely nothing: waiting for people to show up, waiting for people to get around to doing their thing, just waiting in general. This time does not improve the eventual release in any way whatsoever. If you want to look over your work again after sleeping on it, fine. That shouldn't delay the release that much since other people can do other stuff while you're sleeping. If you need to spend more than, say, 12 man-hours on an episode, it's either an extremely fucking annoying episode or you're making up pointless tasks that are doing nothing to contribute to the actual subbing. Furthermore, there is no reason that getting those 12 man-hours of work done should take more than 24 wallclock hours. If it does, you're being inefficient. Some points regarding this: If you have more than one person that is supposed to do to essentially the same thing, you're doing it wrong. If your translator isn't good enough to deliver a decent script on his first try, why do you have him in the first place? TLC's encourages people to be sloppy since "it's somebody else's problem, someone else can fix it". Same thing with having a second editor; if your first editor can't be trusted with doing his job, why do you have him? QC is mostly unnecessary. If you have a good editor, you'll get an acceptable script straight from there; have him rewatch the entire episode if you want but outright QC is just a good way to delay the release in return for catching some typo nobody gives a shit about, if that. If you need to do it, do it collaboratively in a google spreadsheet or something, with the TL and editor present. I've been in some "quality" groups and I see a lot of "enterprise" mentality there: everyone has their own little niche that other people aren't supposed to intrude on; QC'ers can't fix errors (they must be submitted upwards in the hierarchy for evaluation), everything must be done in order, lots of red tape everywhere. QC catch a timing mistake? Oh I could not possibly do anything about this trivial error, it isn't MY job after all. Shuffle it back to the timer and wait 12 hours for him to get back from work and fix it. This doesn't contribute to quality subbing, it contributes to inefficient subbing. This is why most "quality" groups take so long to release shit. They aren't actually spending that much more time on improving their release; they're mostly using it sitting around doing nothing because everything that needs to be done needs to go five layers of bureaucracy first. That, and they spend a lot of time doing the same thing over and over and over (like letting three people QC the same version of the script once each and having each of them point out the same errors and then doing it all over again on the next version of the script). The keys to efficient, quick, relatively painless and good quality subbing are relying on the individual and using a small team. You shouldn't ever have more than 5 people in a team (TL, editor, typesetter, timer, encoder); you can and should usually get away with less (having a dedicated typesetter is usually unnecessary, since someone can double as it; same thing with encoder and timer). The more people you involve the more make-work you create, the more you have to wait for people to stop being afk and the more inefficient the entire thing becomes. With a small team you have to rely on each person to do shit right from the start. Also, anybody should be able to fix any other person's mistakes, and do so if they spot them. This way you rely on every person's individual effort rather than falling prey to the Somebody Else's Problem mentality that most enterprise-style "quality" groups seem to have. TL;DR: a good fansub doesn't need to take a long time to produce; the reason it frequently does is that fansubbers are terrible at managing human resources. Or to put it differently: fansubbing sucks, you should expose as few people as possible to it and spend as little time as possible on it.
__________________
Last edited by TheFluff; 2010-08-29 at 16:29. |
|||
2010-08-29, 19:11 | Link #83 | |
Florsheim Monster
Fansubber
Join Date: Aug 2006
Location: UK
|
Quote:
The death of the QCer is highly exaggerated. Even the best editor makes mistakes from time to time, either grammatically or typographically, and sometimes you need someone who will be able to actually pick these up (there are very few encoders who a) rewatch the whole episode before release once they're happy with their encode and b) are good enough at the editing side of things to pick up subtle mistakes). Yes, the tendency nowadays is to have less QC, and it really shows. Admittedly, a release with lots of QC may never be "perfect" or even completely accurate, but it's more likely to dodge the silly mistakes than a release with no QC. You may also say "Who cares?" Well, I do, because what's the point of investing my time if I'm not gonna get things done properly? I fear I may have wandered a tad off-topic... Back on topic: A common mistake made by fansubbers is thinking that their opinion actually matters. |
|
2010-08-29, 19:26 | Link #85 | |
Florsheim Monster
Fansubber
Join Date: Aug 2006
Location: UK
|
Quote:
EDIT: You added some after I'd posted. I use "properly" in the sense that I'm outputting the best I can do with the resources available to me (for example, I might read the manga before I start subbing a series if I feel it'll be useful to the overall output, but it doesn't mean I would spend a ridiculous amount of hours tracking down interviews with the creators and the voice actors). QC is a pretty easy resource to acquire and it has enough merits to make it worthwhile in reaching this goal. |
|
2010-08-29, 19:28 | Link #86 |
uwu
Fansubber
Join Date: Dec 2005
|
If I wanted to translate something "properly", I would wait until every episode aired and watch/read every side material before even translating the first episode so that I can "properly" translate any foreshadowing, implications, lines without context that require interpretation to work in English and so forth.
So if half-assed means being more reasonable about not knowing what's going to happen, then yeah. |
2010-08-29, 19:46 | Link #88 | |
uwu
Fansubber
Join Date: Dec 2005
|
Quote:
We're recruiting new editors this season. One comes from a group with QC. I will probably have to talk with her + whoever else we recruit at the beginning of next season because there is pressure on their position they might not be used to. |
|
2010-08-30, 03:57 | Link #89 |
Excessively jovial fellow
Join Date: Dec 2005
Location: ISDB-T
Age: 37
|
Eclipse has never had QC either. At least not a dedicated QC; the "QC" usually consists of the encoder proofwatching the encode before release. Then again Eclipse hardsubs stuff, so groups who don't will probably not have the encoder around to do this.
__________________
|
2010-08-30, 04:44 | Link #90 |
Translator, Producer
Join Date: Nov 2003
Location: Tokyo, Japan
Age: 44
|
By my own estimation based on errors reported to me through CR or that I've discovered myself, shows which I work on completely alone on have an error rate of around 0.05% per line.
I.e. I tend to have a typo get through 1 out of every 2000 lines. At an average of 300 lines an episode, that's one mistake every 7 episodes or so. I consider this reasonable enough accuracy for simulcast translations on the cheap. A secondary editor or qcer can probably cut that down by another 50% or so to an error every 13 episodes or so. Getting it down farther than that requires at least 2 independent qc passes and I'd consider only necessary for DVD or blu-ray subs where mistakes cannot be fixed. The amount of qc you need or want is entirely dependent on the quality of your original staff and the quality you want to release. It's not a "mistake" to have too much or too little qc, it's simply a conscious choice about how to work. Just as it's not a "mistake" to release in avi instead of mp4. That's a choice that depends on the audience and venue you are trying to reach. (On the other hand, doing h.264 in avi is a mistake ).
__________________
|
2010-08-30, 06:52 | Link #93 | |
uwu
Fansubber
Join Date: Dec 2005
|
Quote:
Given the quality of the English in the old days, it's not wrong to say there was no QC. When we considered doing Seto again, I re-opened a script from ep 18 or so labeled "QC3" and found a pretty obvious error in the first few lines. ( ´∀`) The way "QC" on Geass R2 worked was seriously dumb. I was so happy that the last episode was just me, Xabin and grunty. |
|
2010-08-30, 07:46 | Link #94 |
Member
Join Date: Oct 2007
|
QC as a separate step is pretty useless because the last person to touch the ep should be watching most/all of it anyway. Whether as an editor checking for flow, an encoder checking their hardsub, or even a timer making sure everything snaps to scene.
It doesn't exactly take much training to notice typos or missing lines, so the only way things should slip through this is if everyone involved is being passive-aggressive at each other re: "not my problem". About the only use of it is catering to staff laziness or scratching the ego of someone who can't even time by letting them feel important anyway. On a side-note: Groups who release karaoked textless OPs/EDs. syduck: |
2010-08-30, 07:54 | Link #95 | |
Ancient Fansubber
Fansubber
Join Date: Aug 2004
Location: KS
|
Quote:
__________________
|
|
2010-08-30, 11:54 | Link #97 |
Pioneer in Fansub 2.0
Join Date: Aug 2007
|
I completely agree with basically everything TheFluff said about bureaucracy. Frankly, it's completely ass-backwards. On top of that, most groups that do this also don't even try to aim to release in any sort of timeframe, which is pretty much the opposite of what actual corporations with bureaucracies like this do. With the combination of "not my problem" and no release timeframe, the result is usually taking weeks to get a single damn episode done. The fansub groups I've been in have pretty much all been like this (save for one), and less surprisingly shit just doesn't get done.
The most recent one I worked in (which I left before we managed to release a single episode) had the "workflow" consist of TLC (they were using a CR script as base), "timing", first editing, second editing, typesetting, first QC, second QC, encode, RC-QC, mistake fixing, potential re-encoding if there are problems in the hardsubbed credits or typesetting. QCers obviously have no permission whatsoever to touch the script, they must all write extensive QC reports, and if they find a timing error, it gets passed to the timer, an editing mistake, it gets passed to the editor, etc. And apparently the one taking care of the "who fixes what" was one of the editors, not the leader, who apparently did... the secondary editing. Go figure. 99% of the time was spent doing absolutely nothing due to either the person who was supposed to do something being missing or "not feeling like working", and of course there was no flexibility whatsoever, NO-ONE HAS THE PERMISSION TO TOUCH SOMEONE ELSE'S WORK. Another experience comes from my friend who also did some timing for this group, and once when he was timing something he noticed a very obvious and glaring typo, something like "the" being spelled as "thhe". He fixed it and mentioned it afterward in the staff channel when he uploaded the timed script. The response? A snarky "hmph, I didn't know timers were supposed to be editors now". I mean, seriously. What the hell? Someone in the group told that the way they do things is "the traditional way of fansubbing". If that is truly the case in most groups, I seriously wish all those groups would just die off. This sort of way of doing things is absolutely ridiculous. It kills pretty much any sort of passion someone might have towards working on something by being absolutely productive-less idling 99% of the time and also results in horribly slow releases. In my opinion, a group doesn't have to even be small to work efficiently, the key thing to releasing is flexibility. As long as you have staff interested in doing something and getting shit done, you should be able to have someone do whatever's necessary to do at the moment, and move forwards without needless waiting for a certain person to come online or stop idling. Though in practice smaller groups probably work better because a certain person must work on it in order for shit to get done, so there's no room for any kind of "not my job/problem"-mentality if the goal is to release something. Also, setting a release timeframe usually helps a lot, because people it motivates people to get things done in the timeframe, and when you manage to keep the timeframe it usually makes you feel good and motivate you further as well. But yeah, because of my mainly bad experience with fansub groups, I work solo 99% of the time (with the TL/timing coming from elsewhere for the things I do). With the TL/timing ready, my work mainly consists of editing, typesetting, QCing and encoding. And some minor time shifting quite often. All in all, it works very well for me because the only person to blame for slow releases is myself, and because I do everything myself there's no lag between the various phases of work since I generally do everything in one long streak. This way I also don't have to hate on other people for being lazy fuckers for not doing their job when they're online and very much available. So, to sum everything up: Less inflexible and huge bureaucracies with no timeframes, more small, focused and flexible groups with release timeframes! It gets things done, and usually with equally good if not potentially better quality, provided that the people working on things are motivated
__________________
|
2010-08-30, 12:31 | Link #98 |
Ancient Fansubber
Fansubber
Join Date: Aug 2004
Location: KS
|
Instead of common mistakes having to do with the actual subtitles we seem to be drifting into group make-up philosophy and how it should be done motifs. Let's get back on topic. Or maybe we've used up everything and need a new thread?
__________________
|
2010-08-30, 14:05 | Link #100 | |
Senior Member
|
Quote:
IMO: Karaoke. Now, I'm fine with simple /k, English at the bottom, romaji at the top. But why hiragana/kanji? All it does is clutter the screen. And don't get me started on those hideous flashy exploding effects. The only way those could be more distracting is if a magic hand comes out of the screen and pokes you in the eye. I don't even understand why karaoke even exists... |
|
|
|