Mangadex gets hit by DMCA - more than 25% of chapters gone | staff FAQ update

[hahanaho]

Welp, guess that explains why I couldn't see any logic in what got purged and not from my trusty tablet reader. And if that many series affected, odds are there's plenty to come as the requests gets processed. Sad times, but I guess all nice and convenient things must come to an end.

On an unrelated note, I wonder if I could muster up enough brain cells able and willing to learn Japanese.... 🤔 Probably not.

I mean then the question will be how to source the Japanese raws of stuff you want to read. Not that simple.

 

[ThetaSigma]

At the very least, this forced me to get a good look at my list for the first time in a long while, probably since the launch of v5 I guess. Most of the stuff I'm actively reading has been spared (although no idea if the scanlators are gonna continue here or not), but turns out that some manga that I had put on the back burner because new chapters weren't coming out had just been dropped in favor of an official translation. Welp, guess I had become a bit too complacent.

Just wish French scanlation groups were a bit more present on mangaupdates, because I don't really have it in me to check their twitter feeds every day or join their discord groups.

I personnaly use the website Nautiljon. It has a planning where you can see the future manga available for each month. You can also see if a manga is available in France, was, or is not, in the manga's page on the website. Imperfect, but it helps a lots.

I mostly use Nautiljon to check the licensing status of stuff I'm interested in (pretty damn useful for anime these days, since our overlords have stopped communicating about expired and soon-to-be-expiring licenses) and while I've been using the planning page on and off, them not having an option to make their listing more compact is kind of a turn-off for me, 300 rows with the covers is a lot to go through sometimes. Oh and manga-news is pretty neat too, you can export the monthly schedule as csv, and also I think they're the only ones who do a pretty comprehensive weekly roundup of the releases.

 

[bondejeger]

I mean then the question will be how to source the Japanese raws of stuff you want to read. Not that simple.

You buy it. How hard can it be.

 

[JagdPanther]

Did not expect to be spending my Saturday morning researching which series I'm reading got hit and how far behind the official English is, if there even is an official English.

My God, some of the series I'm reading the official English is effectively more than 3 years behind.

 

[JVick]

Did not expect to be spending my Saturday morning researching which series I'm reading got hit and how far behind the official English is, if there even is an official English.

My God, some of the series I'm reading the official English is effectively more than 3 years behind.

For me, it's discovering how the 'legit' website do it. Seeing mangaup cut in 4 each chapter of Kusuriya no hitorigoto is wild

 

[JagdPanther]

For me, it's discovering how the 'legit' website do it. Seeing mangaup cut in 4 each chapter of Kusuriya no hitorigoto is wild

I didn't get what you meant until I just looked. That's... just... WHY?

I'm systematically going through my entire Reading (and Completed) list in my MD Library right now. I've found at least 7 series so far that don't have an Official English link (nor one I can find with a cursory Google search) that got nuked. That's pretty dang annoying.

 

[bondejeger]

I didn't get what you meant until I just looked. That's... just... WHY?

I'm systematically going through my entire Reading (and Completed) list in my MD Library right now. I've found at least 7 series so far that don't have an Official English link (nor one I can find with a cursory Google search) that got nuked. That's pretty dang annoying.

Most likely AI/MTL assisted companies like Orange/Emaqi are going to pump out official translations of the thousands of manga they’ve licensed very soon.
Honestly I think this might be the death of scanlation as we know it.

 

[JagdPanther]

Most likely AI/MTL assisted companies like Orange/Emaqi are going to pump out official translations of the thousands of manga they’ve licensed very soon.
Honestly I think this might be the death of scanlation as we know it.

Ewww.

One of the things I've been thinking about as I've been going through my list is that, on the one hand, some official English translations definitely are superior... but on the other hand, I like what some scanlation teams do to try to provide extra context (TL notes), keep a translation faithful (or go the extra mile to make it make sense in English), and a lot of other stuff I don't always see the official translations do. I feel like AI/MTL will be none of that.

 

[bondejeger]

Ewww.

One of the things I've been thinking about as I've been going through my list is that, on the one hand, some official English translations definitely are superior... but on the other hand, I like what some scanlation teams do to try to provide extra context (TL notes), keep a translation faithful (or go the extra mile to make it make sense in English), and a lot of other stuff I don't always see the official translations do. I feel like AI/MTL will be none of that.

Depends, MTL has come a long way in just a few years. Most of the time you get a fairly decent liberal translation. As long as you have a proofreader proficient in Japanese that can correct a few mistakes and add some trivia like that, you probably won’t notice much if a difference.

 

[Angry_Panda]

Most likely AI/MTL assisted companies like Orange/Emaqi are going to pump out official translations of the thousands of manga they’ve licensed very soon.
Honestly I think this might be the death of scanlation as we know it.

Formulating the evil business plan.

1. License works that have been scanlated.
2. Issue DMCAs to have those scanlations removed.
3. Prepare 'official releases' by re-typesetting the translations from the scanlations.
4. Release.
5. Profit.

Groups can't really complain about this, since their distribution of the scanlations is a violation of copyright, hence attempting to challenge the use of their translations is an admission of that violation and would open them up to civil action.

...who's in?

 

[bondejeger]

Formulating the evil business plan.

1. License works that have been scanlated.
2. Issue DMCAs to have those scanlations removed.
3. Prepare 'official releases' by re-typesetting the translations from the scanlations.
4. Release.
5. Profit.

Groups can't really complain about this, since their distribution of the scanlations is a violation of copyright, hence attempting to challenge the use of their translations is an admission of that violation and would open them up to civil action.

...who's in?

Mostly agree, but they’ll probably make a translation themselves with AI, and then proofread them with scanlations to save money on translators. I doubt they’d want attention from using translations from illegal sources. That’s like using cracked photoshop professionally.

 

[Angry_Panda]

Mostly agree, but they’ll probably make a translation themselves with AI, and then proofread them with scanlations to save money on translators. I doubt they’d want attention from using translations from illegal sources. That’s like using cracked photoshop professionally.

I think you're right, but I wouldn't be surprised if in occasional instances where the AI fails completely, someone resorts to scraping the scans for a workable script.

 

[bondejeger]

I think you're right, but I wouldn't be surprised if in occasional instances where the AI fails completely, someone resorts to scraping the scans for a workable script.

Yeah, you could like just copy paste the scanlation translation into chatgpt and ask it to rewrite it or something so it’s not recognizable. And put in some zoomer lingo for perfection.

 

[eromaster]

Just move the damn host country. Capitalist pigs are too greedy and understand nothing.

 

[Valle_reading_manga]

Anyone have any information on which company is the one that issued the DMCA?

It's not that single publisher. A lot of them ganged up on MD.
As for Comeso and other stuff mentioned further down the line, they were "just" issuing the DMCAs on behalf of the publishers...

 

[bondejeger]

It's not that single publisher. A lot of them ganged up on MD.
As for Comeso and other stuff mentioned further down the line, they were "just" issuing the DMCAs on behalf on the publishers...

If my suspicion is correct that they’re pushing this for Emaqi, then it’s most likely most of the publishers. I think all the big publishers all have invested a lot of money in Orange and Emaqi, and they’ve like I wrote licensed thousands of series.

 

[volodyuka]

Depends, MTL has come a long way in just a few years. Most of the time you get a fairly decent liberal translation. As long as you have a proofreader proficient in Japanese that can correct a few mistakes and add some trivia like that, you probably won’t notice much if a difference.

I write a blog, in russian, with lotsa letters, yes I am that old. It is highly stylised text, with difficult sentences structure (russian generally does not have a sentence structure like english has, so you can play with a lot while writing stuff, also why russian classical prose in english is still good but never will be even a smidge as good as original is, poetry too, doubly so), well, anyway, I sometime run an MTL on my texts and it falls flat on it's machine arse. even modern oh so powerful AIs is usually beaten by my partially demented mind.

now japanese is kinda hard. add to it that manga does not have any structure, it's bunch of balloons with text and sfxs. Human can make sense out of it. It's not easy and takes time. For AI you would have to train a model what would be able to make sense of all it and output manga text as a coherent continued dialog, with introspective speech. And I'm not quite sure anyone would bother with this harder task, they will just translate balloons contents, which will become in a lot of cases disjointed mess. also with how corpos use AIs RN they would not bother hiring editors and just put out this shit as is.

yeah. MTL in hands of someone knowing the language and editing can be a tool. and MTL in a hands of a tool...

 

[bondejeger]

I write a blog, in russian, with lotsa letters, yes I am that old. It is highly stylised text, with difficult sentences structure (russian generally does not have a sentence structure like english has, so you can play with a lot while writing stuff, also why russian classical prose in english is still good but never will be even a smidge as good as original is, poetry too, doubly so), well, anyway, I sometime run an MTL on my texts and it falls flat on it's machine arse. even modern oh so powerful AIs is usually beaten by my partially demented mind.

now japanese is kinda hard. add to it that manga does not have any structure, it's bunch of balloons with text and sfxs. Human can make sense out of it. It's not easy and takes time. For AI you would have to train a model what would be able to make sense of all it and output manga text as a coherent continued dialog, with introspective speech. And I'm not quite sure anyone would bother with this harder task, they will just translate balloons contents, which will become in a lot of cases disjointed mess. also with how corpos use AIs RN they would not bother hiring editors and just put out this shit as is.

yeah. MTL in hands of someone knowing the language and editing can be a tool. and MTL in a hands of a tool...

OCR is pretty good now picking up regular Japanese text, and google lens even manage to pick up most handwritten text manga ocr can’t. As for sfx they’re mostly written in katakana, and you can learn that in less than a day.

 

[volodyuka]

OCR is pretty good now picking up regular Japanese text, and google lens even manage to pick up most handwritten text manga ocr can’t. As for sfx they’re mostly written in katakana, and you can learn that in less than a day.

mate, it's not the problem to catch text from page, it is a problem to make it coherent. like manga does not have usual structure prose has, with dialog lines being clearly separated, who speaks when easily understood and etc. instead you just get a lot of lines, some of which will only be partial, no clue who talk when and etc. again, in time and with proper resources you can teach model to read manga and make sense of it (most of the time) but why bother? just ocr text, put it into chatGPT, use some automated cleaner software (ask chatGPT to clean text bubbles), put 'translated' text in, ???, PROFIT! (and quality will be shit because why bother with quality when you can DMCA anyone and just push your shit?)

 

[bondejeger]

mate, it's not the problem to catch text from page, it is a problem to make it coherent. like manga does not have usual structure prose has, with dialog lines being clearly separated, who speaks when easily understood and etc. instead you just get a lot of lines, some of which will only be partial, no clue who talk when and etc. again, in time and with proper resources you can teach model to read manga and make sense of it (most of the time) but why bother? just ocr text, put it into chatGPT, use some automated cleaner software (ask chatGPT to clean text bubbles), put 'translated' text in, ???, PROFIT! (and quality will be shit because why bother with quality when you can DMCA anyone and just push your shit?)

I think you underestimate how good chatgpt and gemini have gotten understanding context. If you load a whole chapter ocr script into chatgpt, you’ll get a translation that’s far better than you’ve described. A lot has happened with MTL and AI/LLMs the past couple of years.