CatBus

Project files have been updated to version 8.0 (original post has been updated as well). Please PM me for the temporary download links until the files are available in a more permanent location.

The goal of the 8.x series is to provide complete subtitle parity, in terms of both quantity and quality, with every known release of the 2011SE worldwide.

I was not able to complete everything I wanted to do before I felt obligated to release what I had in its current form, so there will be a smaller point-release or two down the road--but in its current state, we already offer a complete superset of the subtitle languages offered in every 2011SE release (that I'm aware of), and are very close to matching them in terms of quality as well. Congratulations to everyone who has contributed anything to this project so far!

Rough summary of changes:
- Added four new languages: European Portuguese, Hebrew, Slovenian, and Icelandic
- Changed quite a few languages that were partially or completely based on fansubs to official translations, including: Mandarin/Simplified, Russian, Turkish, Romanian, Croatian, Greek, and Bulgarian. Candidates for similar changes in the future include Arabic and Korean.
- Operation Eyestrain (a joint effort between Sadako, Feallan, and myself) is converting our graphical-only subtitles to text-based subtitles, and this work is ongoing. The status as of this release is: Mandarin/Traditional (complete), Cantonese (complete), Japanese (SW & ESB only), and Thai (ESB & ROTJ only). The result of this work will be editable/correctable subtitles in these languages that also happen to look nicer than our current graphical ones.
- Minor "correctness" fixes and cleanup to documentation and file naming conventions. The subtitles formerly known as "forced" are now known as "titles".
- Instructions are included for how to mark certain subtitles as forced, if that's what you really want to do.
- The version of Muxman included with this project has been modified (again), so that it no longer uses incorrect language codes for the following languages: Indonesian, Javanese, Hebrew, and Yiddish. No plans for Javanese or Yiddish subtitles, but never say never...
- Minor changes to Polish and English subtitles.
- The project now includes a HELP_WANTED.html file, as a reminder to people who use our subtitles that we can use feedback on our existing subtitles, and could always use subtitles in new languages.

Thanks to Feallan, pittrek, Sadako, Teesel, and lexsanor for their enormous amount of help, hard work, and raw materials used in many parts of this release.

As always, I'm sure I've forgotten a thing or two, but those are the big changes.

Nearly all subtitles are now of high enough quality that I feel the categories of "verified" and "unverified" subtitles have outlived their usefulness, so I will no longer make this distinction. The only fansub that I feel I can safely say needs some serious attention at this point is Arabic, and I'll keep working to address that. As always, any of our subtitles may contain errors, including errors duplicated from the original translation, but they can be perfected over time as people report the problems.

The shift toward using official translations in this release doesn't mean I think official translations are necessarily better than fansubs--simply that without the ability to determine the quality of any particular translation, the official ones are the safest choice, and also more likely to be consistent between films.

Lastly, if anyone out there knows how to read and type Arabic text, and would like to help this project provide top-quality Arabic subtitles, let me know. I already have complete Arabic translations in graphical form, but the OCR results from multiple programs just don't seem to be very usable without heavy correction, which I doubt I can do by myself. You don't even really have to know the Arabic language, just be able to read and type the characters accurately. Qualified parties please apply within--otherwise, Arabic subtitles will require a minor miracle to reach parity with the others.

TV's Frink said:

I don't see how someone could post screenshots and video previews of their doubts.

That's the Special Editions in a nutshell, really.

While using Google Translate to help sync the official Simplified Mandarin translation to the GOUT, I got my favorite mangled translation so far.

As our heroes are escaping Mos Eisley, C-3PO says: "I forgot how much I hate Star Trek."

FWIW, there appears to be a Cantonese and Thai dub on the Hong Kong Blu-ray set.  That plus the official Korean subtitle translations make that set pretty interesting to me.  If anyone happens to be sitting on this set, send me a PM and you can be my best friend forever.

The dubs should be no harder to despecialize than our Ukrainian audio--perhaps easier since we don't have to deal with perfectly syncing the English audio accompanying the Ukrainian overdub.  ROTJ should be a piece of cake, SW somewhat hard, and ESB will be a complete bear.

AFAICT there weren't any dubs in these languages prior to the Special Editions (or else they'd have been on the Region 3 GOUT, I figure), but I'd be delighted to be proven wrong.

team_negative1 said:

There will be better encodes and transfers of the 8mm and 16mm eventually.

Let's also hope for a complete 8mm mono track, which I understand to be yet another ESB audio variation.

The Blu-ray mix isn't really all that different from the 2004 DVD mix overall, AFAICT.

Badness depends on your reference.  So many of us have seen the films so many times that we have at least one of the pre-97 mixes permanently etched into our brains, so that's our reference whether we like it or not.  Maybe if you judge the Blu-rays without that mental reference, they sound okay, but for me it's impossible, because I also know how they could/should have sounded.

So here's what I remember not liking about the Blu-ray audio.  Please keep in mind I'm not really an audio guy, so my terminology or explanations might be completely wrong.  They had a weird combination of muddiness and harshness throughout, which I took to mean overzealous use of noise reduction (muddiness), combined with a hamfisted attempt to bring the reduced details back out (harshness).  Usually this was constant, so your ears could get used to it, I suppose, but occasionally they must have done some major adjustments, and the sound suddenly sounded more broken than normal.  I remember that happening with Vader dialogue occasionally.

Then there's the score, which was pushed way into the background in a few scenes--yeah, it's not like the score isn't the best part of the soundtrack, people--let's just minimize that John Williams garbage so people can hear our crap new sound effects for our crap new CGI.  Bleah.  Bad choices all around, even if you accept the bad choice of the Special Editions to begin with.

EDIT: hairy_hen may be your best bet for more detail. A lot of people who don't like the 2011 mix may have only heard it once quite a while ago, or be actively trying to forget it.  It's not something I hope to re-live.

Well, my first reaction was: God, the Blu-ray is unbelievably ugly.

After that, I guess I could see myself being equally happy with either the workprint or the new colors.  I think the problem for me is that, without a reference, I'm just looking for things that are obviously way out of bounds, and I don't see much in either.

I'm assuming you're making changes based on the "they made the colors of the 97SE good on film before they made them bad on the DVD" theory, using some sort of adjusted 97SE as a reference?  What sort of adjustments?

Red lights are clipped on the Blu-rays (often in more than just the red channel), and very hard to restore to a more natural-looking state.  Unless Harmy has a new trick for ESB, I imagine these will remain, as they did in Star Wars.

Now we're talking.  Thank you Harmy!

I found five:

Lando's aide, Luke's gunner, Wedge's gunner, JOANNE D'ANTONIO, KEVIN O'CONNELL

Most noticeable in the first three.  You are a saint.

Ah, then doing it by hand is the only option, unless you can get someone to re-rasterize the text ;)

Nevermind then--it's not horrible not to care.  I'm just trying to find anything that might signal to people that this is a reconstruction, and anachronistic PC-era punctuation is one of them.  But it's not a big deal at all.

EDIT: Actually if you're re-rasterizing--even if you can't match the form exactly, I think U+2019 would be "a better kind of wrong".  Wrong, but more plausible for the era in which it was created.

For an upside-down drop, you may get lucky with U+2032, depending on how the font you've chosen decides to render it.  Different fonts may render it differently.  It's not a typical form though.

Harmy said:

Is it horrible that I don't really care? Or at least not enough to actually go through the whole thing and replace every apostrophe by hand, which I think is pretty much the only solution.

Heh, not horrible at all.  But if they're "curly" (look like the commas, just up high), then you should just be able to swap out the characters in the RTF file with the Unicode character (U+2019) that looks more like that.  If they don't easily match up to a particular glyph in the font you're using, then yeah, hand-replacing seems to be all there is.  But curly apostrophes should be a piece of cake.  And slanty apostrophes may be doable with a prime character (U+2032).

Harmy said:

Here are the new and hopefully final credits:

http://uloz.to/xkG3V6u2/credits-new-mp4

Did you check the apostrophes against the 35mm scan?  Were they clear enough to make any verdict about slantiness/curliness/chunkiness?  They still look a bit anachronistic in their current state.

Oh, absolutely.  But that still supports the idea of the new beeps as part of an overall intentional personality transplant, not as a mere technical detail.

Wasn't there a thread about Vader's floating glass of water in the original drafts for Star Wars, which made it into a comic adaptation?  Basically the guy can't even be bothered to hold his own cup--it's the very definition of frivolous use of the Force.

We'll never know exactly why that wasn't filmed, but I like to think part of it was that in the early days, Lucas accepted criticism and when people told him something was idiotic and unnecessary, he didn't do it.  Now we have three complete idiotic and unnecessary films, so yeah, frivolous use of the Force comes right along with that.

FWIW, I noticed this while doing my SDH subtitling.  Since I tried to match each sound to an emotional adverb (yeah, it's terribly subjective, but also a terribly important part of his "speech" for deaf viewers), ROTJ R2 seemed to have a bit of a different personality.  The R2 in SW and ESB was a rude wise-ass with good intentions.  R2 in ROTJ has a strong pollyanna vibe.  For ROTJ R2, it seemed everything was always going to turn out just fine, you'll see.  SW/ESB R2 had a lot more negative things to say.

Regarding the text of the credits themselves, could you or someone with access to the 35mm scans see if it's possible to tell if apostrophes used straight-down marks (as in the preview version) or not?  By 1980, it was increasingly likely for straight-down apostrophes to appear in commercial typefaces, but I think either curly marks or angled straight marks would still have been more common at the time.

They're in lots of places: Lando's aide, Luke's gunner, Wedge's gunner, JOANNE D'ANTONIO, KEVIN O'CONNELL.  May have missed one or two.  Especially in the larger sizes, they look a bit out of place.

I haven't found SFX or music to be noticeable until quite a bit more than 2 frames offset, YMMV.  I'd be interested in that German audio when you have it ready, BTW.

Oh, and I guess I should add that in my dub collection, most were synced to or derived from NTSC video, so trimming video would actually improve their sync too.

The only ones in my dub collection based on either PAL or some unknown frame reference are: Czech, Hungarian, German, Italian, and Russian non-voiceover.  Pretty manageable if we want to check/fix them, but I honestly don't think a fix is needed.  And we'd have to go back to original sources because all I have is lossy versions, and we'd lose more quality from lossy recompression than we'd be gaining in sync.

FWIW, in my opinion, a two-frame difference only makes a difference for English audio tracks.  Differences that small are only really noticeable with lip sync (and barely noticeable at that).  Since dubs don't really match lip sync 100% anyway, I'd say two frames is within the range of acceptable error for dubs, and two frames delayed is also more natural than two frames advanced.

We also need to remember to check the 16mm mono mix while we're checking English audio tracks.

Harmy--I thought your video was already based on the NTSC frames.  It uses PAL frames?  Yeah, trimming video seems the best choice if it's PAL frames.

I personally don't care if you call it 8.0, but there's always such newbie confusion about version numbering, especially when the version numbers don't match between films.

If 2.5 is the last of the 2.x Star Wars series, then I'd say aim for 2.5 being the end of the 2.x Empire series.  Whatever that ends up meaning to you ;)

Can't tell from the compression if it's any better than the current upscaled GOUT. I'm assuming it is, but who knows? I assume a little-to-no-compression still could be a quick way to determine the relative quality versus upscaled GOUT.

n00b said:

Is there really an occasion where someone needs subtitles and the overlay for rounded corners at the same time? TESB doesn't have any burnt-in subtitles, every existing GOUT-synched language track can be added to the file and everybody speaks English anyway.

A little offtopic, but I'll hopefully be able to take care of it in one post.  I don't know about this particular framing, but similar graphical frame overlays have been done for open-matte DVDs where the user could optionally apply black bars to achieve the original aspect ratio.  That scenario also presented issues for people who needed subtitles.

The best way to add subtitle tracks is via graphical files, not SRT files.  You are correct that the SRT files could be added without interfering with the rounded corners overlay.  There are lots of reasons this is a less than ideal way to manage subtitles, though--no black borders for SDH subtitles, no subtitle shift to avoid onscreen text, lots of client subtitle rendering bugs out there, etc, etc.  I imagine some subtitle files are completely unusable for most people in SRT format, and I really don't recommend using the SRT format unless you have no other choice.  And, although I'm working to address this, some languages, including those spoken by nearly half the population of Earth, are not yet available in SRT format, so graphical is the only way to go there anyway.

Lastly, not everyone speaks English.  Really.  I get PM'ed all day from people who have friends and family who can't watch English-only preservations without either a dub or subtitles.  Often, English-speakers need English subs--SDH subs, for example, and in South Asia, it's common to show English-language movies in English with English subtitles for people who have trouble with the various English accents.

AntcuFaalb said:

FWIW: It should still be possible to implement the rounded corners as a graphical subtitle track, as Laserschwert said.

Yeah, I could throw one of these together pretty trivially, I believe, if there's interest.  The problem, of course, is that you couldn't use this overlay and also use graphical subtitles at the same time.

EDIT: I vote C.