This post was written by Dark_Sage. He is Dark_Sage.
Acchi Kocchi is as bad as I expected.
Release format: MKV (253 MB, 10-bit), MKV (340 MB, 8-bit)
Japanesiness: Honorifics.
English style: American English.
Group website: http://hatsuyuki-fansubs.com/
Encoding details: http://pastebin.com/u4ivuwV7
8thsin’s translation critique: http://8ths.in/spring-2012-fansub-comparison-reviews#Acchi
Ji-hi’s screenshot comparisons: http://www.ji-hi.net/sp12/acchi/
Table of Contents
Visual Quality
Karaoke.
Opening. The screenshots really don’t show what it’s like. I like faggotry in my karaoke. I like it when there are sparklies and things flying everywhere and k-timing and hardsubs. I’m unabashedly wapanese. But this… This was too much for even me to handle.
Hatsuyuki, this is an intervention. You need help.
Ending. This is all right.
Typesetting.
As you can see, they typeset nearly everything. Good job!
Other.
Why do groups think shitty fonts are a good idea?
Script Quality
Karaoke.
“Because if you are going to grin, your back will slowly become moldy”
I don’t fucking get this. The OP/ED were fine except for this part, which threw me for a loop.
Main Script.
Masochist-san. Good job, Hatsuyuki. Your bar-setting skills would make you a hit at a midget limbo contest.
“I’m your present.”
It’s not “I am you are present.”
“Actually, I already prepared this love potion for you, Tsumiki-san.”
“Whoa, you’re making a cake?”
This is more a “best practice” than an actual mistake. “Whoa” is the preferred word to use.
“On Valentine’s Day”
“G-G-Good morning.”
Not really sure why they would capitalize this. You’re not writing a letter here, guys.
“As expected from a family that runs a cake shop.”
A family is not a person. Though it is made up of people, think of it like a corporation. A corporation is also made up of people, but you wouldn’t say “Ah, that’s a Wal-Mart who moved into town.”
Place the comma in the right place.
“Giving chocolate with frog in it to an innocent boy isn’t exactly the nicest thing to do, now is it?!”
“Even a girl like me may get the wrong idea.”
The “Even” part means the “too” part is unnecessary.
“By the way, were these all supposed to be romantic chocolates?”
With the use of “these all”, you need to make “chocolate” plural.
“for everyone who was nice to me day in and day out.”
“everyone” is singular.
“about everyone playing with lots of piled-up snow!”
Timing Review
Incoming. When “Timing Critique” is added as a category, you’ll know this part is done.
Results.
Watchability: Watchable.
Timing Grade:
Visual grade: A-
Script grade: C+
Overall grade (timing results not factored in): B
Hatsuyuki’s typesetting definitely carried them a far way through some stupid script errors, a silly main font, and crazy karaoke. It wasn’t a bad release, but it could have used some tender QC loving.
You actually survived two episodes of Acchi Kocchi? That is some incredible endurance right there.
I’ve edited four episodes of Acchi Kocchi. It usually takes me a minimum of 6 hours per.
Come at me, bro.
Acchi Kocchi is a killer. I started this review at 10 AM and finished it at 2 AM. By the end, I was making excuses to go to the Google homepage for entertainment. Like, “Hey, I should check Google. Maybe they have a doodle. Oh, they didn’t have one at 11:10. What about 11:20? Surely 11:20 will have something.”
I have no idea how anyone could like this garbage.
Watch Shiawase no Pan. That’ll give you a whole new understanding of garbage~
Should have played Zerg Rush
Hatsuyuki always has shitty hardsubbed karaoke. It’s why I never download anything from them.
>It wasn’t a bad release, but it could have used some tender QC loving.
It’s funny you mention QCing. Hatsuyuki usually has 3-4 QCers write up QC reports each episode. This release is pretty much as good as you’re going to get from Hatsuyuki.
But all of that is pointless if the guy putting the script together at the end doesn’t do anything with it. I should know, I once worked for them.
>getting your QCers to write reports instead of just fixing the problems themselves
What? What kind of asshattery are you suggesting? QCers should be blanket-banned from touching scripts.
Mm, I think that’s only true in two cases: 1. You view QC as an entry-level position where you see QCs as not necessarily the best decision makers. 2. You have multiple QCs and having them create reports/have the reports applied would take less time than having each apply their changes to the script before sending it off to the next QC. In my opinion, the QC should be the project lead; I don’t see the position in a negative light at all. I can go more into why I think this when I don’t have to type on my phone if you want.
Well, I see how it can work, but you would need the QCer to be the best person in the group for that, and that’s unlikely in most groups. The issue I have with QCers touching scripts is the possibility of them adding errors in – which as an editor, is then reflected on you rather than the QCer who messed about with it. I’ve been in both types of groups (still am, in fact) but yeah, the end script is only as good as the QCer doing it if you let them change the script.
Or let me rephrase that: the project lead/QC should have a lot invested in the group and really want to make that group as good as it can be, for the releases to be any good. I haven’t seen that in any of the current swathe of groups about (though admittedly, I haven’t been looking that hard).
As the sole QC/co-editor (read: I do a lot of editing work on the edited script) for SS, I generally don’t touch the script itself. It has its advantages and disadvantages.
On the plus side, having only one person touching the script tends to keep the writing style consistent and often encourages back-and-forth dialogue about changes I propose. On the other hand, the editor is the final decision-maker, and pride sometimes gets in the way of improving the script. It’s also a lot longer to write up a report than to just change the script directly.
I disagree. I find QCing a lot quicker than editing, generally-speaking, because I’m looking out for mistakes more than improvements. These usually fall in the camp of contextual errors as well as grammar fails, but when editing, I look to spin lines so that they sound better – which as a QC you don’t really do. Or you shouldn’t, because that’s just backseat editing.
QCing is a lot quicker than editing. But if your QC is allowed to directly modify the script to apply his changes, rather than having to write up a report, it’s a lot faster. That’s the point I was trying to make.
But then, how do you get around differences in style between editor and qc, not to mention anything that might be considered contentious?
By letting your editor do their job. If the QC is going to go through and reword everything they personally disagree with then they might as well be the editor in the first place.
I don’t change subjective stuff when I QC, just obvious errors.
Yeah, I do a lot of editing in my QC passes for a variety of reasons, but optimally, your editor does his job and the QC just cleans up any technical mistakes that were missed (which tends to happen when the same person reworks the script a few times without any other input).
Well, I guess I don’t leave enough technical mistakes in my edits to see the benefit of this particular method of working. I, for one, much prefer to have feedback from QC than just fast-tracking it through error-control. It makes for a better overall release.
No, you just need better QCs. You want people who have at least basic knowledge of editing, timing and typesetting and are competent enough that you can trust that there’ll be less errors in the script than when they started.
I didn’t say they were easy to find, but when you have someone who you can just hand off a finished script to and have them handle the rest, it really does improve your efficiency a lot than having to go back down the chain.
I’ve QCed plenty of episodes at this point and I’ve fixed plenty of problems and issues (both minor and major) across all of them. I’ve yet to add an error of my own.
Basic knowledge? lol. I think we’re talking at cross purposes here because for a start, when a script leaves me, there are very few technical errors that need fixing (save a stray comma or two, or an errant typo). Most of what qcs pick up on my scripts are contextual issues or (as much as it pains me still) the odd Briticism I accidentally drop in – some of which are contentious, all of which require a competent editor to mull over the possibilities and perhaps completely re-edit a whole line or section.
As I said above, both can work and I’ve been in groups where both *have* worked, but for it to work well, the person fixing the script has to be the most competent. And more often than not, that person does have an extra pair of eyes helping alongside and pointing out bits that have been missed.
Though, this only really applies to groups that strive for quality. I hope you’ll forgive me for not taking advice on QC from someone in Commie >.>
Low blow, man. I don’t see how it should be any different from TLC, where generally speaking you want your TLC to be a better translator than your TL who simply doesn’t have the time/inclination to do the whole script themselves. That way you can trust them not to add errors.
And yes, I don’t believe QC should be an entry-level position, but quite the opposite. I wouldn’t even want to work for a group that didn’t even trust their QCs enough to fix anything, and I hope you’ll forgive me for that as well.
So wait, the only requirement is “basic knowledge” but it’s not entry-level? I’m not quite sure how that works…
It’s not about trusting/distrusting QC – I’ve been in a position before where I’ve made some stylistic choices as an editor that QC has viewed as “errors” and changed. As I say, what you’re proposing only really works if you have a *really* good QC – as in, they have to be on the same tier as the editor themselves, so that they can spot when an edit is a stylistic choice or a *true* error.
I’m not suggesting that QC is solely an entry-level position (though it can be that in a lot of groups, as it’s the perfect way to see how other people operate within the group, and learn how others go about their respective jobs – i.e. it’s a good learning tool). I occasionally QC too. I can see the benefit of good QC and I do rely on it to bring up issues I never even considered on my editing pass (I sometimes leave a line with a note in parentheses for QC to discuss!).
But it also breaks the feedback chain. If an editor makes an error, or if a QC feels that a section works better another way, how does the editor learn from it if the QC makes the changes directly? The role of editing is one where the editor is constantly learning and reusing things already learnt – if you sever the feedback loop, a vital part of that process, how do they ever get any better? :S
No, read what I said again.
Basic as in, they know how to fix things like timing or typesetting errors when they see them, even if they’re not an experienced timer or typesetter themselves. For example, I have very little actual typesetting experience and can’t do complicated moving signs or rotations, but I’m still able to do things like copy & paste recurring signs that were missed or typeset basic static signs that weren’t done. That makes me a lot more useful as a QC than someone who has editing experience but doesn’t know how to use Aegisub.
Like I said before, the QC should be fixing obvious errors that can be objectively judged to be wrong. For more subjective stuff I would either ask the editor if they were around or just leave it because something subjective isn’t worth holding up the release for.
Obviously your staff should talk to each other, but I consider the culture where QCs can only write “reports” that may or may not get acted on by someone else at some point instead of being responsible for the actual quality themselves to be harmful. If I QC something then I’m the last person to touch the script, period – I even do the final mux myself and make sure the chapters are there and all the right fonts are included. If there are errors in it after that it’s squarely on my shoulders. It’s all about having people who take initiative instead of waiting for other people to act.
Sorry, what? Editors that don’t use Aegisub? Do they even exist? I know there are some like Collectr who edit hard copy first and then do the second pass on Aegisub, but there are editors who forgo the most basic of fansubbing tools altogether? Wow.
“Or just leave it.” My, my, we have a very different philosophy on subbing. I’m not even sure this conversation is all that worthwhile because we seem to be two opposite ends of the scale. Mind you, I guess that’s the speedsubbing mentality?
I don’t see how that way works any better. As the editor, when I mux, I know all the other jobs (save encoding and translating, obviously) so I can generally make reasoned decisions on what needs fixing. But even then, you’ll sometimes step on the timer’s toes because you’ve cut something that he wanted lead-out on. The QC *are* responsible for the quality of the release because their reports will get looked at and all the relevant errors will be fixed when pointed out. There’s as much chance of a report not being looked at as there is of muxing in the wrong script into the final release, so… yeah.
And wow, do all subgroups work like you nowadays? I’ve heard of the leader/project leader sometimes doing the QC and changing the script before release (which I understand, it’s their group/project – but then, I make sure I work with groups I’m comfortable with making those changes if they’re gonna do that), but not *all* QC people. I think it’s just asking for trouble.
Uh, yeah? We usually edit before timing, which is done directly on a pad unless it’s a simulcast script (and even then it’s often just done in notepad or whatever).
And yes, “just leave it”. If it’s a completely subjective thing that barely even matters, it’s not worth delaying the release for. When there are people depending on you to provide them with subs so they can actually watch the shows they’re following from week to week, I can’t imagine very many of them being thankful that they had to wait an extra 12 hours for subs because QC wanted to ask the editor over a subjective difference of opinion in one line.
And don’t ask me, I’ve only worked with two groups so far. Underwater works the same way except that everything gets QC’d before release (usually by Daiz) and the QC itself is carried out in the way I describe.
…
When did fansubbing turn into this? OTZ
I would have thought simulcasting would have taken the impetus out of speedsubbing and headed back in the direction of quality subs, but it looks like that hasn’t happened. Shame, really.
Good luck finding a simulcast for any of the shows I’m working on this season.
Speed and quality are not mutually exclusive. Subs can still be quality even if there’s one typo in the script that we didn’t spend an extra week on QC to fix. The fact that we, in my completely subjective opinion of course, are able to do a better job subbing this anime in ~30 hours than Hatsuyuki do in a week should be proof of that.
>Hatsuyuki
Aiming the bar high there, I see.
So much more goes on in that week than typo fixing. In fact, those bits are the easy parts. The bits that don’t even need thinking about. Some of the other bits are more time-consuming to get right.
And some of the best groups I’ve ever worked in/watched subs from have still had the odd typo in. But appreciation of a series can be so much greater when time is actually spent on a show, rather than just pumping it out for the sake of speed. Admittedly, this can be a quick process all the same – you just have to have a few more hard-working staff who are around lots. But releasing something because you want to get it in the hands of the leechers quicker, before you’ve had time to discuss a potentially disastrous point… well, each to their own, I guess. You’re the “popular” ones. I just lurk about on forums, starting arguments :D
>Project leader
What a fucking useless position.
Honestly, any time I see the phrase “As expected from” in a fansub to me it’s a symptom of shitty editing. That phrase ranks just below “It can’t be helped” on the annoy-o-meter.
It’s still better than Hadena’s version, which had “As expected from a Cake shop’s son.”
If you think their typesetting is good just wait (´∀｀)
I downloaded this episode just to watch the faggotry in the OP karaoke.
There wasn’t enough faggotry. I was disappointed.
4koma adaptations are always boring. Why hasn’t Japan learnt this yet?
Easy marketability?
Man. I JUST finished editing Episode Four. It took me a grand total of five hours, I think.
/sobs
Oh hey, I didn’t think I’d ever see my own translations being evaluated at Whiners.
Regarding to the weird ED karaoke translations: Somehow all the changes after the first version got omitted in the script. Sadly I noticed this after the release of episode one and two, since I backed up the changes on my PC. Fixed translations for the opening and ending for the upcoming episodes. (And apparently we just love them hardsubbed fancy karaoke)
As for the translation/QC: We’ve been switching staff for this series quite a lot. First episode had me stumble upon the script for hours (this was my first translation project), the second episode had a second TL but his English could use some improvements (like the your ==> you’re error). I guess this line didn’t get changed in my 2nd pass translation and was skipped in the QC/Editing script. Sad story.
Next episodes will have another, better translator and me appointed as TLC (Exams are coming up next month so I need more time), so maybe it’ll be better from then onwards. But truth to be told, this series is really hell for typesetters.
Five/six hours to edit? Man, you guys are doing it wrong. Or need better transrators.
I also see nothing particularly wrong with the font, but then, I *am* an editor, not a typesetter. As long as it’s readable, it’s fine to me.
If it’s an Arashi script, then five hours sounds about right. Just about every line needs to be edited or re-translated because it looks like jibberish.
Arashi is… an exception. His scripts are garbage and you’d be better with a guesslation from the editors themselves.
In terms of editing in general? A three-pass edit should take 1.5 hours for an average editor. A good, thorough edit should not be taking more than 2 hours. 5 hours is insane. You’d be much better off having 2 editors look at a script for 2 hours each than 1 doing it for 5 hours. I can’t even imagine how many passes they’d be doing. 10 passes? Holy shit.
There’s no way I’d have that staying power and I have admiration for people who do. Having said that, I try and get away with as few passes as possible when I edit, so that’s not saying much :D
Usually just ~2 passes for me. I just spend a long time on each line since I rewrite most of them and it takes a while to think of something decent (I also tend to play the audio a few times per so I can get a good feel for the mood and length)
I’m probably just especially slow though, since I’ve never had an original translation take less than 4 hours to edit. Only time I can manage it in 1.5-2 is if it’s something like a CR edit and I barely change any lines – even Ozma took like 3 on average.
Four hours minimum? Holy shit. I’d kill myself if it ever took me more than 2 hours of actual work.
Here’s my process:
Pass 0 (only for shitty TLs): Script-only. Fix as many stupid mistakes as possible. Capitalize sentences, add punctuation, and run the Aegisub spellchecker. Then take the entire script and toss it into Microsoft Word and get rid of the stupid errors that way.
Pass 1: Load up the audio/video and start watching. When I get to a line that needs fixing, I’ll fix it. If a line seems like it’ll take a while to figure out, I make a note of it and move on. Really tricky lines I will ask for help in the staff channel on. (“Hey, TLC-kun, what does this line mean?”) I like to put {}{}{} for those lines so I can quickly come back to them. [Target: 45 minutes]
Pass 2: Script-only. Now that I have context for the entire episode, I can better fix those pesky lines I couldn’t figure out earlier. Mistakes that were introduced by my editing in Pass 1 are eliminated, most errors are fixed, and flow is improved. [Target: 15 minutes]
Pass 3: A more thorough version of Pass 1. It goes faster because there are fewer errors and allows me to make sure that everything fits perfectly together. [Target: 30 minutes]
Speeds vary depending on script difficulty. A decent CR script would take me about an hour to get done.
Having help from a staff channel is invaluable. There are always lines that are tricky to figure out, so allowing people to spend time on those while you work on other things allows you to shave a lot of time off the final product. Fansubbing is a team exercise.
I’m more efficient than most people, sure, but I can’t imagine spending over 2 hours on an episode on a regular basis. I’d suggest aiming for faster times on your work. Otherwise, you may find yourself burning out.
For an original TL it’s more like this.
1.) Wait for TL to translate in pad. Slowly edit as they work. For a long show with lots of signs this can take like 3-4 hours.
2.) Read over again at the end.
Of course, a CR script or something shouldn’t take more than 2 hours. Xythar is just slow.
I just finished editing Hyouka ep2 as Vale translated, all up it took about 5-6 hours (somewhere in the neighbourhood of 3am to 8am with a bit more on either side). I have no idea how I’d do it any faster than this and still maintain quality but maybe I’ll get better with experience.
It usually takes me anywhere between one and four or five hours to go through a script. I tend to be faster when I’ve been drinking, though it takes me longer to fix harder lines ;D
It also takes me somewhere around 3-5 hours. How can you do a good job on THREE passes in 1.5 hours for an average 300 line script? I have no idea how that’s possible. Even simple maintenance on a CR script would take over an hour (fixing their shitty line splits, removing their 100 ellipses, etc).
I could livestream my editing sometime if you wanted to see just how exactly Dark_Sage does it, but it’d be so boring that I’m not going to.
After five years of editing where I always forced myself to do my scripts in less than two hours, I simply became really, really efficient at it.
I honestly thought every other editor did this stuff in a similar timeframe.
I wish my scripts were only 300 lines long :(
Hyouka ep1 was something like 475, I shit you not.
I finished editing Shippuden movie 5 which was just shy of 1000 lines. I think it took about 7-8 hours spread over three days. :(
>A family is not a person. Though it is made up of people, think of it like a corporation. A corporation is also made up of people, but you wouldn’t say “Ah, that’s a Wal-Mart who moved into town.”
According to the SCOTUS, corporations are people.
Also,
>“Giving chocolate with frog in it to an innocent boy isn’t exactly the nicest thing do, now is it?!”
You’re missing an “a” and a “to”. It should be:
“Giving chocolate with [a] frog in it to an innocent boy isn’t exactly the nicest thing [to] do, now is it?!”
Doesn’t seem all that necessary to me. Think about this common phrase you’ll need when you go to Korea:
“Wait, does this have dog in it?”
Everyone knows you mean “dog meat” but you don’t have to be that specific. Similar case here.
Oh, for some reason I thought this was some sort of hollow chocolate with a live frog in it (I haven’t watched the show, I’m going purely off of your screenshots). Still need the “to”, though ^^
Ah, good point. Whoops.
>According to the SCOTUS, corporations are people.
Dark_Sage will revisit this when super PACs start fansubbing.
Just throwing it out there, I am a proud Acchi Kocchi fan. It’s like Lucky Star, which a lot of people like, myself included, but a little different. I think it’s just as funny though, and certainly just as cute.
Lucky Star was really fucking bad.
Lucky Star was great. :/
See, I forced myself to watch the whole series because of the praise it got, but I never got the appeal. The OP was super kawaii no desu, but the actual show was really boring.
“At Valentine’s Day”
WTF?
I’ve been liking Acchi Kocchi so far, honestly. It’s refreshing – reminds me of Lucky Star and Nichijou, but with a more “romantic” twist… and the MC isn’t a derp for once.. plus he is well-liked. Which is nice, since normally the MC is some dumbass that only the show’s heroine(s) like. That being said, the first episodes were a bit slow, I’ll admit – but so were Nichijou’s. First couple of eps of Karin, too, at least in my opinion… and it ended up being a pretty decent romance/comedy.
do a review of HorribleSubs AKB0048 their first ever subbed anime.
2012, it’s happening.
I just want to hear the story about how Hatsu got to the moldy back. That has got to be worth something.
I haven’t a clue. That line should be 背中借りるから (or at least that’s what we had, and from what I could make out from their kara it’s what they had as well) but I have no idea how you get “moldy” from that.
Sorry for the late response.
Before the actual release, we found a pre-airing of the ED on a radio stream upload, but it was in really low quality. It had guessed parts here and there, but I fixed them after the airing of the first episode. Sadly, the person in charge of the karaoke took the older version, which still had errors in it.