How would you feel about reading your mango if translation only poping up on mouse over?
Cutting out cleaning and typesetting could save valuable time and lots more stuff could get translated.
Moon capable readers could also critique or improve on the translation.
Other urls found in this thread:
github.com
danbooru.donmai.us
twitter.com
Would be nice, but afaik only the boorus allow this.
Didn't we have that thread some times ago?
It's not like translators often clean and typeset manga. They usually translate as much as they want/can, so the overall amount of translated stuff wouldn't increase. Personally, I find this kind of typesetting a bitch to read, and it looks like ass.
I had one project stalled and died because cleaner / typesetter was a perfectionist and got burnt out.
Then I translated this one other chapter at a group's request to replaced a low effort release on batoto and it never got typeset.
Personally, I think I'd translate a lot more shit if I could do it all myself.
Not very convenient but it would be great for some passive moon study
Is it possible to install Danbooru on shared hosting? Or is VPS required?
github.com
I'd actually like that a lot. You can see exactly what font the author used, giving a better understand of the scene, and you can actually see the words the characters are saying, helping imagine the scene (if you imagine the characters speaking in japanese). I know this will sound dumb, but it feels like subs for manga.
Although it's a bit inconvenient when the text covers part of the action/panel.
as someone who primarily consumes large amounts of content in a single sitting, an overlay that requires mouse movement and otherwise clutters up the image seems like a pain in the ass to use
on the other hand I could totally see your idea being a speed-translation scene godsend
blast out the translation, then have someone copy/paste into an overlay engine, bonus points for cleaning/leveling the source scans a bit
Only good for one page things. Entire chapters and volumes are a no go.
At first I thought this idea was fine, but then I had some GTO flashbacks and I'm not really sure any more.
It would be an improvement on that shitty censorship people were talking about in that Tenkuu Shinpan thread yesterday, though. And like people said, probably a good workaround if the nips ever decide to really crackdown on scanlations.
It's an interesting idea. I've done typesetting and image editing for a long time. It takes a lot of a time and effort, especially to make something that properly conveys the original style intent of the manga. There are a lot of technical difficulties that are present when inserting English text and sfx into Japanese formatted pages.
I think a better method of doing this would get rid of the yellow box overlays and instead of using a mouse, you could just press left/right arrow keys to make the translation pop up in the order that dialog and sfx are read. To change pages, you would use page up/page down.
Of course, this method isn't perfect and could use improvement. But I think there would be something to gain by using this method because digital releases could come out much faster. For official releases, this would cut out a lot of the cost associated with publishing digital manga and allow companies to release many more titles at a faster rate.
I'm getting flashbacks to the GTO scanlations.
>It takes a lot of a time and effort
People really under-estimate this. Especially true when redrawing is involved.
Yellow box is obviously distracting and not ideal. Keyboard control would definitely be nice.
Perhaps someone could fork danbooru and make it manga reading friendly?
This is where the page in OP is from btw:
danbooru.donmai.us
>press left/right arrow keys
Using the mouse scroll would be better. Pressing two keys is too much effort.
mouse wheel control would probably be impossible to code for websites though.
Best case scenario: boorus for manga
Worst case scenario: GTO
I think it's worth trying. If someone could fork a booru, we could put up some sample pages with an already-done chapter to see how it feels.
Ideally, it would support several methods i.e. mouse, keyboard, scroll wheel etc.
But like said, using the scroll wheel runs into various problems with web pages.
If it made the two things I'm interested in getting TL'd, then I'd be fine with it. Anything to make that process easier is welcomed. But I feel like it would open doors to translators conflicting with one another over the accuracy of translations.
It would help the translation threads here but aside from that it wouldn't be very good as a permanent thing. If Sup Forums implemented it it would be very good. Other boards have things special board specific like the programming thing in Sup Forums in the text as example.
Text overlay for digital manga
Full lettering for physical releases
I never understood cleaning. It's a higher form of autism that is beyond me.
>on that shitty censorship people were talking about in that Tenkuu Shinpan thread yesterday
what was this?
Cleaning also includes the most basic shit like leveling and aligning pages. Even OP's idea would require cleaning; what it's removing is typesetting and redrawing.
Check the archives, it's not as if that series is a frequent conversation topic.
>If it made the two things I'm interested in getting TL'd, then I'd be fine with it. Anything to make that process easier is welcomed. But I feel like it would open doors to translators conflicting with one another over the accuracy of translations.
You make it sound like that's a bad thing.
There's nothing a vote system can't resolve.
Work flow:
A translates a bubble (v1)
B comes over and edits it (v2) (it goes into effect immediately). A gets a notification that it's been edited and a view history icon is added to that bubble.
A scoring / voting system will determine which version is displayed. If scores are the same, newest version is displayed. v1 starts with score of 0; v2 starts with score of 1.
From here, a few things can happen:
If A agrees with B's edit, do nothing. B's v2 will show as default. alternatively, A can also delete v1.
If A doesn't agree with B's edit, A can le upvote A's v1 once and make it have a score of 1. Or A can improve on B's v2 and create v3 which will have a score of 1 as well.
If B does not agree with A, B can create v4 with a score of 0 (as each user can only assign 1 point to one version) which will not be displayed as default. This is also designed to favor the original translator.
At any point, a user could start a translation check request which 1) alerts moderators, 2) gets added to the TLC feed on the site's main page, and 3) shows a disputed icon on the bubble.
User C, D, E, F can then come along and vote for the version they like most to resolve the dispute, or create their own version.
CROWD SOURCED TRANSLATIONS.
a more useful idea to promote would be a layered image standard with support for metadata-powered information overlays baked right in
I imagine something along the lines of subtitle files where the origin, size, endpoint and shape of a text area is specified and given an identifier, then content is associated with the identifier to be displayed as intended when the image is viewed
but maybe I'm just talking out my ass
>But I feel like it would open doors to translators conflicting with one another over the accuracy of translations.
This. It would be a shitstorm of epeen translators. Translators can literally argue forever about the translation of almost any line.
It actually would be a horrible thing. There is no way to vet the quality of the translators and just because you make a vote doesn't mean the majority is right or ever qualified to have an opinion about the lines in question. It would be a total meltdown and just lead to endless flaming.
Design by committee is fucking shit and don't you ever forget it.
Yeah, you have no idea what you're talking about.
fuck you, I want my images to have embedded databases
How about a drop-down menu that would give you the translations from different groups? Not saying this would stop scanlator epeen drama, because nothing can stop those faggots, but it wouldn't be too different from how batoto does it with different language versions.
>new image standard
If it doesn't run on jim doe's chrome browser without extra plugins, it won't attract widespread acceptance.
You have to realize tat 80% of translation you see is shit. autocratic decision making is not the answer. Respectful collaboration is the path to higher quality translation.
I personally don't see this becoming a shitstorm. When anons do live translation threads and they get corrected, they're pretty chilled about it.
If need be, users who produced 10,000 quality submissions could gain contributor status and their opinion would be worth like 10 points when applied to something they did not translate. Perhaps.
Each use can also exercise their rights to choose by blacklisting submission by specific users.
>When anons do live translation threads and they get corrected, they're pretty chilled about it.
user translations are really different from most named translators, though.
Like it makes perfect sense to us to not care in the slightest if someone changes our translation, because we're working anonymously and not getting any credit. But the majority of scanlators that work in groups under names? They're always arguing about "sniping," or getting pissy if you make edits to a chapter, and generally acting like they have exclusive rights to whatever manga they're working on.
If you intended this system only for user translators and the handful of decent translators, it would probably be fine. But a lot of the regular scan groups would never join in if they can't slap their names on everything.
>You have to realize tat 80% of translation you see is shit
Don't throw out nonsense statistics.
>Respectful collaboration is the path to higher quality translation.
Which is impossible on an open website. You have to hire staff to do what you're proposing.
>I personally don't see this becoming a shitstorm.
Because you're naive and inexperienced.
>When anons do live translation threads
That's a completely different scenario where one individual has the final say.
>If need be, users who produced 10,000 quality submissions
Judged by what criteria and by who? That's not logistically feasible.
>Each use can also exercise their rights to choose by blacklisting submission by specific users.
That's an even bigger rivalry fueled shitstorm. Get a grip.
>When anons do live translation threads and they get corrected, they're pretty chilled about it.
difference is those anons aren't the ones depending on the patreon/paypal bucks
>That's an even bigger rivalry fueled shitstorm. Get a grip.
How does user 0f3ea4 adding 1as30's into his personal blacklist lead to any rivalry fueled shitstorm?
Tanoshii.
Actually he's not far off. You could add this information to the end of a jpg (bytes 0xFF, 0xD9 indicate end of image) then use a program or plugin to read that extra data, placing the text as required. If you don't care about the size of the file you could even include multiple languages.
Even if you had to go as far as creating a program to parse a jpg and display it, in order to be able to overlay your words on it that's pretty easy shit that a high-schooler in a programming class could do.
I do like this idea better than just a web version, since I want to keep the stuff I reread.
Works for boorus, not for every other fucking thing.
It's not like you're saving much time in a team, since an ideal scanlation pipeline will have the second page typeset while the third is being translated and cleaned, and the first is being QC'd.
Scanlation request booru would be rad, where translators do spot translations of unclaimed series or shorts.
>since an ideal scanlation pipeline will have the second page typeset while the third is being translated and cleaned, and the first is being QC'd.
You have no idea how to scanlate manga.
>where translators do spot translations of unclaimed series or shorts.
Yes please, I always have random pixiv comics that need translation.
No, it would never work because you cannot typeset an image like you would subtitle a video.
>It's not like you're saving much time in a team, since an ideal scanlation pipeline will have the second page typeset while the third is being translated and cleaned, and the first is being QC'd.
I think that only works for live translations here. Chances of getting a regular group all available at the same time to do this are super low.
It also only works for bare minimum typesetting and for things that have official digital jp releases.
How do I get gud at learning moonspeak
Yeah, tl usually outpaces editing by a fair bit.
Get a teacher. Read lots of Japanese every single day.
imagemaps have been a thing in html since forever, it's not much of a stretch to apply a similar concept to show text
>live translations
That's what I was thinking of, thus "ideal".
Unprofessionally, it's the asynchronous 'each person does it when they're ready'. But those cases aren't very concerned with how long a release takes, so there's already time to clean and typeset.
Yeah because subtitling a video works using MAGIC and can't be done anywhere else.
A simple example that could get widespread appeal would be to do it using css in the browser using a plugin. You telling me you can't store the starting position, angle, font-size, font-family, etc. and then overlay it? Come on man, I've literally already done this before.
You've never typeset anything in your life. Please, go make your imaginary program. I'm waiting.
Someone ask Sup Forums to code a demo.
This thread is specifically about NOT having to typeset.
Maybe not super complex things that require redraws and all sorts of shit, but something like this page? Easy as fuck man.
Hell, there was a guy who did this years ago with Great Teacher Onizuka, it looked like shit back then because the internet was shit but a more sophisticated solution would be possible.
Regardless, people don't like innovation, even when it's good for them. Maybe you don't remember people trying to push mkv and all the crying and whining that happened since people were used to hardsubbed avi?
Yes please. Cleaning and typesetting take way too fucking long.
Not to mention all the manga that has never been released despite being translated because the clean up is impossible.
I have thinking about such service before but could never figure out a way to implement it.
Go ahead and try to do it. You're going to fail because you have no idea about what you're even trying to do. Please. Go typeseting a volume of manga with this method. Go for it. You can keep talking shit all day. Produce some results.
I do do typesetting and cleaning and all that, though not programming like those two. Assuming they're not wrong about the programming end of things, I'm not really seeing what the problem is. Any page I've done could easily be split into 1 leveled image + 1 transparent overlay with all the text and redraws on it. It wouldn't be efficient or anything, but it wouldn't be impossible.
Assuming OP brought this up based on that thread from last night, the original idea was proposed as a way to circumvent potential copyright claims.
Rather than creating a totally unique standard, I've thought for a while that it would be neat to have a type of file that does essentially what you're talking about with a stripped-down form of HTML packaged together with multimedia data.
Like here's a page, then
Sounds like a nightmare for archiving.
here's what their program would put on top of it, I assume?
Would be crazy inefficient for series with detailed redraws, but not impossible like you're saying.
Why do you believe this to be so hard? What is your secret that you're not revealing? Do you think picking the same fonts that every group uses and sticking it in a white box is some sort of skill?
>circumvent potential copyright claims
Would it? If the work isn't copyrighted in a region, distributing the original isn't illegal. If the work becomes copyrighted, the rights-owner could copy the fan text for the copyrighted release and bar it that way (would it be prior art? does it even matter?). I'd imagine copyright courts have some prior ruling on this, given how localizing foreign works isn't new.
and let some fuckhead like you enjoy the fruits of it free and clear? hahaha, that's cute
I'm actually working with a largeish domestic comic publisher on something similar, may throw some time at the concept as a side project sometime
Easy, take whatever it is you do in photoshop, and do it in a similar program that either saves it as another layer OR saves it as raw data and use some other program to render it.
Hell, with all this AI creating coloured photos from black and white it should be possible if one was so inclined to create something that does a lot of the cleaning work as well.
OP here. I have no idea what "the thread last night" or "potential copyright claims" refer to. I was mostly translating shit in the kemono friends thread.
I just want collaborative translation and more efficient scanslation by cutting out redrawing / typesetting.
Redraws are NOT NECESSARY when the English text is just hovering above the image when you want to read it and is hidden normally.
Typesetting does not need to be fancy when you can see the original japanese font used. Just need to set coordinates.
I don't know, I have no idea how moonland copyrights work. But maybe the fact that you're not distributing a scanlation, just the pieces of it, could cover your ass legally?
But since manga publishers almost always let scan groups quietly do their thing, it's not even necessary at this point. The conversation sprang up because Helvetica was dropping Grand Blue after it got licensed, but it's not like they got a C&D.
Oh that's funny. Someone brought up almost the exact same idea there so I assumed you were thinking on it some more.
>Typesetting does not need to be fancy when you can see the original japanese font used. Just need to set coordinates.
If you're doing all that, setting a font is not a big deal. You could still make it look nice. If it looks like shit nobody is going to want to read that over traditional scanlations.
I would. I don't care how clean it looks, just give me the damn translations.
>The conversation sprang up because Helvetica was dropping Grand Blue after it got licensed
I just started reading Grand Blue, you have ruined my week.
Luckily it's popular and likely to get picked up by someone else. Pray to god it's not a shit tier group.
this desu
anything that lowers the bar to participation means more shit gets translated faster - and if the translation quality is shit but there's enough interest due to, you know, content actually being available, that could cause "real" groups to pick up projects they never would have known or cared about before
the circle of life and all
>Redraws are NOT NECESSARY when the English text is just hovering above the image when you want to read it and is hidden normally.
For your idea with a web-based booru-like system, I agree. It'd be great for speed scans or shorts.
When it comes to what's saved on my HD, I want it nice and clean and permanently typeset.
How Elevens do copyright doesn't matter unless they partner with other countries for umbrella copyright filing, which hasn't and likely will never happen. If someone doesn't have exclusive copy-right for the work in a country, there's no copy-right in that country to infringe upon.
>Helvetica was dropping Grand Blue after it got licensed, but it's not like they got a C&D
Sounds like a moral "support the creator" decision, though it's not like a C&D can be filed for actions prior to the copyright being granted.
Let's take bets on which shit group will pick it up.
Jaimini's box?
Mangastream?
dropout?
One Time Scans?
Norway Scans?
Mangabox.
Or that one that ragequit after putting their stupid group chatter on top of actual panels.
>putting their stupid group chatter on top of actual panels
>one
We can dream.
Death Toll, took me a minute to remember. This is specifically the group that first tried to escalate by making the text more obnoxious, then gave up and dropped the series after someone made versions with their TN shit removed.