TetriSubs: Ecstasy of Order translation and subtitling project

Thread in 'Discussion' started by Kitaru, 27 Jan 2014.

1. Kitaru

EDIT 2018/01/20: Time for an update! I finally adjusted the timings enough to create proper SRT files for English and Portuguese!

One unresolved issue I'm still mulling over: the translations for on-screen text in the Portuguese draft would still cause the same text reflow issues in an SRT export that we were experiencing with the overlapping spoken lines. For now I've exported the Portuguese SRT with only spoken lines shown.

English: SRT | .ass project file
Portuguese: SRT (No Text) | .ass project file

I'm reconnecting with some of the original project members at the moment to see which languages we can resume work on, but if you're interested in contributing, please let me know. You can open the .ass project files linked above in Aegisub, or I can provide a Google Sheet or Excel file to work off of.

(I have some messy Python scripts I'm using at the moment to help get the project files re-concatenated to output an SRT file (since it seems that otherwise Aegisub sometimes outputs some of the lines displaying at the same time out of order for some reason and I couldn't find a place to fix that for some reason); I'd release the scripts too, but they're seriously messy and the workflow is not intuitive. )

I also fixed the broken Dropbox links to the original draft files in the post below.

--

Another post that has been on the backburner for a while...

We've been working in Google Docs and IRC for a bit, but I wanted to put forth a public thread to help with organization since things have slowed down again.

Here are some of the sample drafts of what we have so far:
https://www.dropbox.com/s/ng1nnu2n7b6f45q/EoO_English.ass?dl=0 (Thanks to clincher and steadshot for support in the initial transcription and timing!)
https://www.dropbox.com/s/h3tjcxfiysmu69d/EoO_Portuguese.ass?dl=0 (Thanks to Natalia who volunteered to translate!)

We also have a Japanese translation in the works thanks to EMA and SQR. No draft of that available as of yet.

Work on the project thus far has been a bit ad hoc. So, I'd like to nail down the English draft to a final state and sync up any existing translation efforts to match the line breaks and timing for that before moving on, if possible. I know there are still some rough/busy sections that I'd like to get to flow more smoothly. If anyone with the movie can help watch the movie through with the draft I've linked and give feedback, it would be much appreciated!

Last edited: 21 Jan 2018
2. Kitaru

@Steadshot: I know you have some comments floating around -- could you link them here or PM them to me when you get a chance? Thanks!

Sure, these are my notes for roughly the first 30 minutes.

4. clincher

I just watched half an hour and tried to note some stuff.

Code:
30:41 - and then eventually*
30:42 - back playing Tetris*
33:16 - while I was playing Tetris*
36:22 - it's not
38:53 - [pretty sure he says 'play']
40:22 - [some mumbling not subtitled]
43:52 - 'of a' or 'of the'
58:25 - be able to get*

Those are my notes for the second half hour. I'm not sure how demanding we have to be about the faithfulness of the subtitles, though. Sometimes there's details that don't really make a difference.

If they need to be completely faithful, then I'll have to go back to this half an hour and note down everything I left out.

EDIT: Kitaru, any idea if someone is working on the Spanish subtitles?

Last edited: 29 Jan 2014
5. K

oh, the english subtitile timing is finally done. i'll watch the movie and work on french translation

6. Kitaru

@steadshot, clinch: Thanks, I can probably do some editing this weekend.

@clinch: As far as I know, Spanish is open.

@K: It has been finished for a while, but IRC didn't give it proper visibility. But yeah, it's basically just down to refinement now as far as English goes. Feel free to start working on it. The line segments for the subtitle files might shift around a bit as things are fixed, but it should be mostly stable.

7. clincher

Just watched another 15 minutes and noticed a few more things:
Code:

60:58 - 'at my highschool'?
61:55 - wanna see it* [don't know if we're being that faithful to the speech]
63:19 - two buttons yet*
68:08 - ['really, really' or without the comma?]
71:14 - is that, well, Robin*
71:44 - will be attempts*
74:12 - some sick(?) Tetris*



8. Kitaru

I still need to reorganize my work environment for this on my new computer, but I don't want to forget about this project.

Other than these sorts of corrections, are there any issues with the segmentation/flow/readability? That's my main concern. I'll try to watch another pass when I get things set back up.

9. Kitaru

"18:05 - maybe split the two sentences into two seperate subs?" Did you mean ~18:00 to 18:05? (That would be Line 283. If you mean 18:05+, can you suggest how to rebalance the speech fillers in that sentence? ) I'm not sure if I want to goof with it in case it leaves things on screen for too short a stretch of time, but it might work OK.

"23:37 - maybe split that sub into two lines (@ the semicolon)" Making note of this.

"25:04 - commas after the like" This is a legitimate usage, not a filler.

"38:53 - [pretty sure he says 'play']" It's very ambiguous whether or not is voiced. I think it might still be "played."

"40:22 - [some mumbling not subtitled]" I'll most likely leave it as is. There is "oh shit" timed but commented out. I don't think the "right there right there" and other barely audible murmuring is too critical.

"43:52 - 'of a' or 'of the'" Ambiguous. Leaving it for now.

"58:25 - be able to get*" It's noisy but I do think "being" is voiced.

"60:58 - 'at my highschool'?" http://en.wikipedia.org/wiki/High_school

"61:55 - wanna see it* [don't know if we're being that faithful to the speech]" Weewannoo see it! Don't know, leaving it for now.

"74:12 - some sick(?) Tetris*" Possibly. It's noisy.

10. Kitaru

Anyway, I addressed all the other stuff and put it in the English file. Since I haven't changed any segmentation, I can port the minor timing tweaks to Portuguese easily.

Time to revive this sucker.

11. Qlex

Wow... I'm having a 100th additional reason to watch this movie now!

Did you mean ~18:00 to 18:05? - Yes.

25:04 - commas after the like's?

I was referring to this line: "We hung out like in the mid 90's and like played games together" where they're fillers. It switches to that subtitle right at 25:04, so I guess it was a confusing timestamp.

13. 0m1cr0n

Hi guys, i've been doing the spanish subtitles and i'm almost about to finish them. Do you still need them?

14. clincher

Whoa, that's awesome. We do still need them. Would you mind sharing them so we could collaborate?

Thanks!

16. Kitaru

I finally did that one last tiny change in the English sub file. I'm willing to call those "final" for now. So, it's all about rekicking the translation efforts now, haha.

17. Qlex

Not sure I will have enough time, but if Jago and I can get back to the thing then it might work eventually!

18. jmzc

Hi: I've seen that you are made Spanish subtitles for 'Ecstasy of Order' . I'm interested on it . Could you send to me by any way ? Thanks

19. Qlex

Kitaru likes this.
20. Kitaru

Ah, that's a pretty nice way of working with things! If I recall correctly it was possible to multi-line paste into AegisSub and that's how I copied the Portuguese spreadsheet over, but this looks really convenient too.

Also, I'm really sorry this project keeps stalling. I feel like I've stretched myself thin, but even still things have lagged far too long. I should really be able to close some of the small loose ends to get where we need to be.

For one, I need to repurpose or rewrite some of my old scripts/tooling so we can generate a proper SRT subtitle file. While SSA was a great format to edit in, and a great format to view in if your media player supports it, there is a requirement out for a barebones SRT as it is widely (if not nigh universally) supported due to its simplicity. This means doing a bit of scripting to look over the existing timings to flag the ones that overlap so we can normalize them: combine and align the timings of lines that occur during the same space of time. Ideally, we should be able to take any finished SSA subtitle file (or spreadsheet, perhaps) and apply the final set of timings and/or text formatting rules to get an SRT. I have ideas in mind for scripting each of these steps but just haven't been able to bring things roundtrip to completion, so I just wanted to put out some sort of update.

Qlex likes this.