Parallel colaborative fansubbing proposal

[Starks]

At the same time, 1000 computers in a distributed computing grid doing an encode would be sweet though... If that's even possible.

<_<

[martino]

http://forum.doom9.org/showthread.php?t=117889

[Jiyuu]

Originally Posted by Quarkboy

The fansubbing community is not unified and has too many enemies to allow open editing ala wikipedia.

Imagine for a second if a show was subbed this way: If some anti-fansub people heard about the site, they'd immediately start vandalizing the script. Or a rival fansub group that wanted people to watch their release instead.

Heck, the wiki for fansub groups that lists which shows which group is planning on doing gets vandalized enough as it is, I can only imagine the annoyance if there was an actual script editing site.

And that doesn't even approach the question of differences of opinion. Someone doesn't like the sound of a line, changes it, than the original person changes it back, etc, etc...

In the end, the only way it could possibly work is to appoint moderators for each script that would have to approve EVERY change before it was applied and have a seperate forum for dispute resolution, etc... And because of that I don't think there would be any advantage from what most groups do now.

Also, it goes without saying that this would produce very poor translations. 1000 translators doing 1 sentence each does not make a good translation.

what you are saying are things that were, and still are, said about wikipedia in general.
there are plenty of ways to fix this, among other things are only allowing registered users to edit and yes, moderators that help sort, and lock if needed, problematic items.
the point is it's much easier to restore vandalized things then actually doing that (especially so if its for e-mail verified users only).

about something more specific like

a rival fansub group that wanted people to watch their release instead.

well "rival"(what rival? its not even a "group" just random people contributing to a script, one that can be used by other fansub groups to make actual fansubs if they wish to do so) groups can do these kind of things without a wiki. much more destructive things are far easier then vandalizing a script. just to name a few you can overload the server with a large enough group of people,
you can poison the bittorrent swarm, you can even report a group to some company and possibly cause them to get a C&D from that company.
really, doing evil has so many other ways to be done, whats the logic of being that afraid of something like that, especially when you can restore stuff to the way they were before with the click of a button?

And that doesn't even approach the question of differences of opinion. Someone doesn't like the sound of a line, changes it, than the original person changes it back, etc, etc...

again, this is something that happens in normal wikis and get solved.
in this case you can enable forking the script if its really necessary..

and

Also, it goes without saying that this would produce very poor translations. 1000 translators doing 1 sentence each does not make a good translation.

this again is the same thing being said about wikipedia all the time, yet we all use it and generally trust it..

and again, technically if a translated line has its kanji/romanji/hiregana/whatever attached to it so anyone can see it should be easy to notice errors.
again the idea is that many people contribute as much as they feel like.
maybe I'm mistaken and every single person that will actually be able to contribute will be a self centered egoist who wouldn't feel like helping out if it doesn't grant him a huge credit or something, but i believe there are some people that will be willing to help not to get credited but to have other people enjoy that same anime they enjoy.

and even if there won't be any "professional-level" fansubber willing to help, the idea is that this way if 20 people will help a little to something no one wants to do it will at least be somewhat helpful, as opposed to not having anything..

don't get it wrong, I'm not one of those people that try to tell people they should do things for free.
I'm just saying it seems like a way to get something you can't have today.
there are plenty of people that while they don't know Japanese well enough to actually join any group they picked up enough to translate a couple of lines from some episode.
with a group of people like this getting a basic script up, it will first of all help all of help all the other people in that level, as while they understood some of the lines they didn't get other lines.. that way they can share that knowledge and get some back instead of being stuck with understanding half an episode.



also another benefit i think can come from something like that is that if groups see lots of activity on something they might get the motivation to actually do it.
instead of people coming to them and bugging them to do it.

Originally Posted by martino

http://forum.doom9.org/showthread.php?t=117889

wow thanks.
this could actually help me
at work i have access to like 100 computers that are p4 3ghz to C2D 6300..
ofter 6PM or so i can use all that computing power to encode.. that just sounds awesome

[DryFire]

Originally Posted by Jiyuu

this again is the same thing being said about wikipedia all the time, yet we all use it and generally trust it.

Wikipedia has set certain standards (their manual of style and other things) that fansubs generally lack.

Translation and editing is a bit different than the (ideally) factual style of encyclopedia posts. You would need to have translators and editors agree on what common terms should be localized or transliterated as and how much localization.

I suppose it could all be solved by differing levels of localization and other things, but how is that considerably different than current fansubs? Contributors will often find themselves falling into different camps, much like current groups.

[cyth]

@Jiyuu: You could set a simplified fansub standard based on what most groups do. Those who are able to swallow their pride for their personal standards will be able to participate. If you introduce this project to newbie fansubbers, most of whom are still blank pages regarding standards, you won't have much difficulties building a nice little army of speedsubbers.

Seriously, if you want this to work, you'll eventually need to set up some standards. Why not swallow that bullet and be done with it? I think you have other issues to worry about though. Centralization of illegal practices can introduce legal issues quicker, for example.

[getfresh]

Using SVN would work fairly well also since you can compare versions. You could even have a system where people can rate the different versions submitted for each line. This way when the final selection of line versions to use on the release script are made they can be partially based off of the rating system.

[Jiyuu]

Originally Posted by DryFire

Wikipedia has set certain standards (their manual of style and other things) that fansubs generally lack.

Translation and editing is a bit different than the (ideally) factual style of encyclopedia posts. You would need to have translators and editors agree on what common terms should be localized or transliterated as and how much localization.

I suppose it could all be solved by differing levels of localization and other things, but how is that considerably different than current fansubs? Contributors will often find themselves falling into different camps, much like current groups.

that is why I said one of the technical things needed might just be the ability to fork the text..
that way it'll also make it simpler for people to translate stuff to other languages..

having said that, i do think rules for such a thing will be a must, there would have to be a definition of a couple levels of localization so you'll be able to have a japanese track, possibly a romanji(basically if someone wants to help but doesn't actually know what kanjis are correct he'll be able to listen and write what he hears, that way when someones writing the kanji and is not sure of what he heard he'll be able to compare with the romanji or something like that), and like up to say 3 levels of localization(it'll obviously have to be discussed and decided about how many localization levels will be, and proper rules for each level).
that way when someone wants to help and he likes to use san, sama, chan etc. he'll know he is considered level2 and will type it there(also incase he wouldn't know, its still alot better then there not being anything, so even if he posts it to the wrong level it'll still help)
and once a translation is available in any type of style it already is far simpler to change it to another level, even without too much knowledge of Japanese.

but how is that considerably different than current fansubs? Contributors will often find themselves falling into different camps, much like current groups.

basically the idea is not to target just the same people who already are in fansub groups, what i want is to "use" the power of people who are not good enough to be in fansub groups but still have some knowledge, basically fansubs take people that know Japanese in a pretty high level, between that level and not knowing Japanese at all i think there are plenty of people that know some Japanese but not all(you can see plenty of bloggers that while they don't completely know Japanese are still able to blog about episodes from raws, they usually don't get all the text but they do get some of it. and that can help people who know nothing at all) so while i do think that to produce something of the level of actual fansubs it'll still require people in the level of fansubbers, but they'll have a lot less work as most of the simple stuff will already be done, and all it'll take is fixing some of the mistakes, as opposed to actually translating everything.
and that even if it'll cause people to "fall into different camps" each of the camps will be working and helping(not necessarily actively, it could just be the fact that they already translated it in some way..) as opposed to different groups doing all the work again and again..

I think you have other issues to worry about though. Centralization of illegal practices can introduce legal issues quicker, for example.

that is my main concern, the thing is i'm not sure about the legality of the translation.
i mean, as long as you are not uploading the actually video to the translation site it should keep most of the heat away.
it will make no sense for commercial groups (I'm talking Japanese groups btw, i haven't though about the licensed anime issue) to go after a site that enables people to download translations without the actual content, while i think that technically speaking the law does not allow translation of content without the approval of the author but i think this is alot less of an issue then actual fansubs that are released with the original content.
so its a far less shady tone of gray then normal fansubs.
and i think that if and when the industry actually notice something like that, it could be possible to discuss something like this with then, cause while the possibility of fansubs hurting actual sales is pretty obvious
it is not as sure about simple translations.

Using SVN would work fairly well also since you can compare versions. You could even have a system where people can rate the different versions submitted for each line. This way when the final selection of line versions to use on the release script are made they can be partially based off of the rating system.

a rating system is optional, but i think this will only be necessary when there be a large amount of versions for each file..



i personally think about this as open sources fansubs, the industry makes closed source proprietary subtitles,
most of the people make it in a way that while they don't charge money or anything still make the fansubs closed source (in the programming world they are the minority i think, such as haali)
and I'm suggesting a fully open sources model,
so anyone will be able to help.

there are lots of technical issues such as how to unify the timing, and as mentioned the different levels of localization and i'm sure there'll be plenty of other issues.
i think the main issue that will decide if something like this will succeed is how many people will be willing to contribute.
cause if you have people that actually want to help, you can always solve the other issues in some way..
but even if you have a perfect system, if people don't feel like helping then whats the point..

[Tofusensei]

Originally Posted by Jiyuu

so while i do think that to produce something of the level of actual fansubs it'll still require people in the level of fansubbers, but they'll have a lot less work as most of the simple stuff will already be done, and all it'll take is fixing some of the mistakes, as opposed to actually translating everything.

I can tell you're an idealist, which is great. However, I think you're vastly underestimating the amount of time/work "fixing some of the mistakes" actually takes for an experienced translator versus just translating everything from scratch.

I do think, however, as opposed to focusing on language skills, focusing on technical skills is a better approach.

Look at the number of people who watch hundreds of fansubs on streaming sites like Youtube yet cannot figure out how to download fansubs directly (even though the URL is usually right in the video!). If you can make a web 2.0 platform that will make fansubbing easier for those who are afraid to use irc/aegisub/ftp/etc. you can create a real impact. Sort of like NicoNico (or now Youtube) allows you to overlay subtitles. Dumb it down and capture the power of the aforementioned folks.

-Tofu

[Jiyuu]

in most animes about 60-70 percent of the text are pretty simple sentences that most people who watch anime for enough time will probably be able to translate..
at worst a good translator will merely need to fix the entire translation of the rest of the sentences, so it should still save plenty of work.
and incase such a translator won't come you'll still have some translation.. which is better then no translation.

i know that sometimes fixing something that is broken takes almost the same, or even more, time then to do it from scratch yourself.
but if we'll always wait for the best programmer/translator/editor/timer/whatever we'll never get anywhere..

moreover as i said i am also talking about a system similar to what you described, which ontop of that will support an open editing and wiki like properties that will make it possible to join a lot of medium level to get something more then their individual knowledge..

[martino]

Originally Posted by Jiyuu

in most animes about 60-70 percent of the text are pretty simple sentences that most people who watch anime for enough time will probably be able to translate..

I'm surprised that somebody has actually said this...

If that were the case then most groups wouldn't have a problem with translators, and you'd probably have a hell lot of crap out there, translation-wise. Enjoy the rest of your day.

[Jiyuu]

but its a fact there are plenty of people that understand a large portion of the episode #_# without being fansub-translator level.. X_X

and if you already mentioned the "fanasubbers wouldn't have an issue" well what can you say about all those speed subs that are like, 70-80-90% correct?
i got to see plenty of those, people always complain about there being fansubs with lots of mistakes and such.
so while the people who translated an anime with about 70-80% accuracy might not be that smart of an idea as single-translators for fansubs people like that will be able to help a lot in a system like the one i suggested..

[Starks]

Why does everyone flock to SVN for version management? <_<

Give a modern system, like Mercurial, a try.

[Skyw4lker]

Originally Posted by Jiyuu

but its a fact there are plenty of people that understand a large portion of the episode #_# without being fansub-translator level.. X_X

Guesslating != Translating.

Besides, there's a huge difference between simply understanding what's being said and carrying that over into another language.

[Jiyuu]

is it that hard to accept the idea that there are people that can understand a phrase like
doyate kanjiruka oshiete yo
but wouldn't know what
ii kusuri! ie doku ka?
meant..

you know, its not all or nothing you can understand some of the episode without understanding the rest..
its not that either you don't even know what "baka" means, or you are capable of reading ancient Japanese literature..
theres plenty of a middle stage.

[dj_tjerk]

There's only one way people will actually help when you set up something like that; keep stats of how many lines one has translated/timed/edited.. stats == epenis...

[Jiyuu]

XD
thats one way to give people motivation

and people will get ranks according to those stats..
after all having a special title by your name=even more epenis

and over that we can make it so people can buy items with the points from editing,
and the types of edits you make give you skills.
and then we get an MMOFSRPG


oh and about

Look at the number of people who watch hundreds of fansubs on streaming sites like Youtube yet cannot figure out how to download fansubs directly (even though the URL is usually right in the video!). If you can make a web 2.0 platform that will make fansubbing easier for those who are afraid to use irc/aegisub/ftp/etc. you can create a real impact. Sort of like NicoNico (or now Youtube) allows you to overlay subtitles. Dumb it down and capture the power of the aforementioned folks.

i already linked to it but http://www.stormberry.tv/ has just that