46 thoughts on “Fansub Review: [Underwater] Kill la Kill (Episode 02)”

1. It was a nice release. I’m guessing that with your jizzing over the typesetting the visual grade would be higher without factoring in the karaoke?

   Reply

   • Correct. Significantly higher.

     Reply

     • Maybe you should just have a separate “karaoke grade”?

       My primary concern with song styling is that its readable. A lot of the time, fancy effects or styling will work directly against that. I’d also prefer to spend my time appreciating the actual OP/ED animation instead of having constantly attention-demanding karaoke effects bothering me. You seem to be quite the opposite. As such, I’d say it makes sense to split it into its separate thing. I’d also recommend to have less weight on it, because honestly, karaoke is probably the last thing I’d use to decide which group to follow, right behind honorifics/no honorifics (seriously, unless two different releases are pretty much exactly the same in every other aspect, these two things should pretty much not affect what you pick).

       (By the way, the OP style here isn’t “orange”, it’s pink and red, which show up quite a lot in the OP. The ED colors aren’t “random” either, as the ED is rather predominantly tinted towards light blue/green.)

       Anyway, let’s talk about visual things in general. Ignoring video quality (which should at least be passable), the most important thing visually to me is the main dialogue style – after all, it’s what you’ll end up seeing the most during the actual content. Good font, good size, good colors, good margins. Luckily enough many groups have at least decent styling these days, though a whole bunch of them still needs to learn the importance of horizontal margins. Way too often I still see subs that are almost hugging the left and right edge of the video at the same time. Another thing that I wish groups would learn is the importance of manual line breaking. At the very least, break lines that would be two-liners anyway nicely because VSFilter is dumb. Preferably I’d like them to split even more, but realistically speaking I probably can’t make people do that unless I can almost completely automate it (which is something I’m working on). Then there’s also a bunch of other micro-optimizations, but their importance is not that high (but are nice to have, obviously).

       Typesetting is the next thing. Good typesetting can really enhance the visual experience, especially if there’s a lot of signs. Depending on the amount of signs, it can even be potentially more important than dialogue styling, because at least dialogue is relatively easy to restyle on the user end.

       And then there’s OP/ED styling. They’re just 3 minutes of content that repeats practically every episode. Any relevance they have to the main content is in the animation and possibly in the lyrics, so those should really be the priority, not sparkly karaoke effects. And even then you’re likely to only watch them once (unless you really like the songs), so their importance in the big picture is even smaller.

       Reply

       • “I’d also prefer to spend my time appreciating the actual OP/ED animation instead of having constantly attention-demanding karaoke effects bothering me.”

         Exactly my thoughts. I remember when i watched the OP/ED of vivids Coppelion for the first time, and while the styling effects definitely looked cool, after it was over I was like “Wait, what even happened in the opening?”. I did focus more on the lyrics effects than on what was actually happening in the opening, so i had to rewatch it right after.

         “And even then you’re likely to only watch them once (unless you really like the songs)”

         Yep, i usually only watch the OP/ED once to check out whats going on in there and appreciate some cool animations / art style and see if the song is nice (in which case I usually download it later). After that I hardly ever watch the OP/ED again, unless the song is REALLY good (and since i download it anyway if I like it, theres no real need urge to watch the shorter 1,5 min version of it)

         Reply

         • What I(a.k.a a typical leecher with no fansubbing experience, if you will) prefer is of course clear translations, and as Diaz says I would not download an episode just for its karaoke. In more than 50% of the shows that I’ve watched, I’ve simply forwarded the OP/ED after the first episode.. Everyone has their own opinions; just thought I’d share mine. That, of course does not mean that I’ll not appreciate a good karaoke; just that I feel that it is not as important as typesetting or the translation.

           Reply

       • 1. I do have a separate karaoke grade. And I actually even have it for both OP and ED individually.
         2. If you don’t want to read the song translation, then turn the subs off.
         3. The color you used is not a color that’s found in the OP. lrn2colorpicker. Your ED was also hard to read. Try watching before you release.
         4. Timing/encoding/styling work on a pass/fail basis. To imply main font styling is more important than karaoke and typesetting is rather laughable. Readability is all that matters.
         5. The only one who’s ever thought manual line breaking is important is you. We’ve been over this before. You go ahead and cite the study you like and I’ll provide humorous examples of why it’s a bad idea. yada yada yada let’s not waste time.
         6. Why the fuck are you talking about good typesetting being useful? I don’t believe I was arguing contrarily in this post.
         7. Who the fuck said I wanted everyone to have sparkly kara effects? Stop putting words in my mouth. Yeah, I like ’em, no I’m not gonna drop your score for not having them. I gave this: http://www.crymore.net/wp-content/uploads/2013/10/Commie-Golden-Time-02-F2B01C6A.mkv_snapshot_22.56_2013.10.13_02.22.43.jpg a “Good” for example.

         Reply

         • Well, as it turns out, there’s actually two issues here when it comes to colors. One is in the release, one is on your end. First of all, your rendering is fucked. This is what the OP/ED should look like color-wise:

           http://puu.sh/4PSmZ.jpg
           http://puu.sh/4PSnU.jpg

           I’d suggest you take a look in your GPU settings, look for options regarding video, and turn absolutely everything you find off. Alternatively, use madVR.

           However, there was also a problem in that the release script was incorrectly tagged as BT.601 when it should have been BT.709 because the last person who touched it had “Force BT.601” on in Aegisub. As a result, the colors got fucked, so even if your settings weren’t wrong the colors still wouldn’t be right. I’m going to make sure that never happens again, because absolutely fucking no fansubber should have that setting on in this day and age anymore.

           I’m not going to bother with any of the other stuff because we’re probably not going to get anywhere by arguing about it (let’s just agree to disagree, though while you certainly do have separate karaoke grades, you should probably keep it away completely from the final “visual grade”), but you should really fix your playback. Meanwhile I’m going to make sure the message about “Force BT.601” goes through to fansubbers far and wide.

           Reply

           • They’re still shit “effects”. Why are you here just because you got a lower grade for colours/style? Do you care for Dark_Sage’s score that much? You’re cute

             Reply

           • Your OP looks fine there. Too bad that’s not what I saw.

             My graphics card is set to its defaults (I assume AMD Catalyst is the “culprit”?), much like I’m using CCCP as default, because my test environment is intended to reflect what most people are going to experience when they watch subs. You can’t expect your viewers to change their system settings just to play your release perfectly. That’s insanely arrogant.

             I still don’t exactly see why karaoke should be a non-factor. And it seems to me if you really believed that, you’d just… not do karaoke at all. Go ahead and set a bold example for us with that philosophy then. Let’s see how well it goes for ya. See? It matters. We can disagree as to what degree it should matter, though. But I’m not dropping it down to 0%.

             Reply

             • >My graphics card is set to its defaults (I assume AMD Catalyst is the “culprit”?)

               Yes, it is. AMD enables a ton of video “enhancements” by default which do nothing but fuck shit up, and you should disable all of them.

               Reply

               • I have an ATI card so I took a look in Catalyst to see what was set and by default everything did seem to be turned off, related to color anyway. The only options turned on were automatic deinterlacing and enforce smooth video playback. Dunno if it’s something the manufacturer enforced (these drivers were acquired through them) or if it’s because it’s old enough that it’s still called ATI, not AMD, and they started messing with things after that.

                 Reply

             • Seconded for AMD Catalyst. I’ve always disabled everything since I discovered this, but I’m not sure I’m comfortable trying to uninstall it entirely.

               Reply

               • What do you mean you don’t want to uninstall it? Sure, games, and pretty much everything else will look worse, but at least Daiz’s karaoke will look perfect.

                 Well, you know, if he decides to have someone QC the next episode.

                 Reply

                 • Wow, I’m surprised at your lack of reading comprehension. Nobody wants you to uninstall your gfx drivers and disabling the video “enhancements” doesn’t affect your games (or anything else than EVR for that matter) at all.

                   Reply

                   • He’d ask me to store my data on tape if it made his subs look better.

                     The problem here is that I’m not changing up my test environment for one group.
                     Here’s my super-unfair set-up:
                     CCCP
                     All drivers for my hardware
                     Everything fucking updated

                     I’m not changing anything for one group’s colors to look slightly nicer due to their own inability to handle multiple environments, especially not when said group originally tried arguing they didn’t even matter in the first place.

                   • Look, D_S, advising you to turn GPU “enhancements” off has not really anything do with my karaoke, as far as the colors go they were fucked up just by the incorrect colormatrix tag.

                     But those GPU “enhancements” are seriously nothing but fucking stupid garbage that fucks with literally EVERYTHING you watch, tampering with the intended looks of every series with completely detrimental filtering. AMD is really terrible with this, and I’ve been telling people to disable that shit elsewhere on the internet long before this comment thread too, you’re not some special case.

                     So please, for your own good, turn those “enhancements” off or at least use madVR which isn’t affected by them. And keep things that way for everything you watch.

                   • Oh, if it’s actually fucking with *everything* then sure. I just didn’t want to make an exception for one group. I’ll fix it before the next review.

                   • Not sure if serious.

                 • D_S, just turn off the video enhancements in catalyst.

                   Reply

                   • I think you’re all misunderstanding the definition of test environment.

                   • it’s ab it of a bad test environment since, the integrated intel graphics cards and nvidia graphics cards do not turn on such “enhancements.” It’s only AMD that does it..

                   • Ooh, integrated graphics and nvidia. You’re right, I forgot this was 2006. My bad. Let me go to the junkyard and find a more suitable test machine.

                   • You’re just bitter because we nvidia folks didn’t have to hack our emulator before playing Danganronpa.

                   • Dark_Sage, the thing is if you turn of AMD’s stupid video “enhancements” you get the same results that everyone who doesn’t have an AMD card.

2. Haven’t watch this yet. Thanks for the review. I now know where to get the Kills.

   Reply

3. Eclipse were right, who gives a fuck about karaoke effects.

   Reply

   • Dark_Sage does! But a group can completely forgo kfx and still get a good or an excellent so long as their karaoke visually enhances the OP/ED to a notable degree. I’ve given Commie quite a few high scores for karaoke, and they’re the group that claims they care about karaoke the least.

     Ultimately, if a group decides to go with a chosen aesthetic, that’s fine, so long as it works in practice. Here, it didn’t.

     Reply

     • >”the visual grade would be higher without factoring in the karaoke”
       >”Significantly higher”

       I think it’s unfair to “significantly” lower a grade for a release that has excellent typesetting just because they didn’t style 3 minutes of bullshit that no one except you and a very, very, very minute minority even give a shit about. KFX/song FX are such an insignificant part of a release that knocking a group for not doing it ‘well’ is unfair.

       Reply

       • Typesetting and karaoke are weighted similarly (though the weights change depending on how many signs there are). It’s the visual section, not the typesetting section. If you don’t care about karaoke, then ignore that section of the review. You’re free to take away whatever you want from my reviews, and I specifically format them so it’s easier to do so.

         That being said, I’m willing to change my weights if you can convince me that on the whole that I’m unfairly representing the quality of a groups’ work with my reviews. But you’d better have a really, really, really good argument for it, because I’m quite happy with how it is right now.

         Reply

       • Plenty of people care. If you don’t care about karaoke, you might as well stick with HorribleSubs, since karaoke and typesetting are generally the biggest reasons people avoid them.

         Besides, it seems to count as one letter grade while typesetting counts as the other four since it still gets a B.

         Reply

         • Are you for real? lol people who know what they’re doing avoid HS because CR can’t not shit themselves when they try to encode.

           Reply

           • I guess you’ve never seen what HD TV looks like? (hint: http://www.screenshotcomparison.com/comparison.php?id=43722 your eyes will bleed).

             Reply

4. Dialogue: 9,0:06:49.59,0:06:51.42,Default,,0,0,0,,Thank you, milady!
   Dialogue: 9,0:06:53.42,0:06:55.06,Default,,0,0,0,,Without fail, milady!
   Dialogue: 9,0:19:33.56,0:19:34.27,Default,,0,0,0,,Yes, Milady!
   Dialogue: 9,0:22:22.37,0:22:23.76,Default,,0,0,0,,Forgive me, Milady.

   Hm. I’ve seen it suggested that terms such as “miss” and “milady” should not be capitalized for direct address when used by themselves. Job and familial titles are capitalized, while terms of respect, status, and endearment are generally not. Supposedly. That said, a quick search shows that some books use caps while others do not.

   Anyway, no matter which way you’re going on this, having 2x milady and 2x Milady can’t be right. Of course, this is a rather minor detail.

   Reply

   • You’re right. I neglected to double-check their consistency in that regard. Good catch, it’s definitely wrong. Updated.

     As for standards of capitalization… that’s not a terrible rule to go by. I like it. Unfortunately when it comes to right/wrong critiques, the whole issue is muddy. Gotta love English.

     Reply

5. Episode one was some fun tracking: http://i.minus.com/iL5IYyprPpB58.gif

   Also, I’ve been editing the font to make it appear rougher like the original signs.

   Reply

   • Maybe it was just my computer, but the font in the first episode looked like it had been unnecessarily bolded.

     Reply

     • It wasn’t thick enough to me, so I used \bord to make it thicker. The font in episode 2 was edited to be a bit thicker, but not as thick as it was in episode 1 with \bord.

       Reply

       • During that first week, I kept repeating that scene over and over again because of the kickass tracking and music. You did a great job Juular, much appreciated :)

         Reply

   • That is some epic work there. Great job, Juular!

     Reply

   • Did you use Mocha? Cause that’s some amazing tracking! I don’t think Mocha would be able to track that.

     Reply

     • Mocha can track almost anything if you know how to use it right.

       Reply

6. >Best typesetting of the year
   >Meh karaoke
   >Gets B-tier ranking
   As expected of D_S!

   Reply

   • B-But 3 minutes of KFX drastically affect the viewer experience.

     Reply

     • 3 minutes of meh is not going to bring the quality down for 99% of people.
       Bad subs? Sure, but these weren’t bad subs, just mediocre ones.

       Reply

7. “Wh-What kind of kinky outfit is this?!”
   “get-up” works better than “outfit” here.

   I disagree. Get-up connotes oddity whereas outfit connotes specific purpose. The garb Ryuuko wears is a transformed sailor uniform. Oddity is already implied by her use of kinky.

   Also, Ryuuko may wish to acknowledge that her shreds of clothing have purpose to help maintain a shred of dignity.

   Reply

   • *”purpose, to help her maintain” not “purpose to help maintain”

     Reply