companies ain't learning anything. lol 'manga has "continuously suffered" damages from online piracy and illegal English translations'

Dex-chan lover
Joined
Jan 9, 2023
Messages
2,239
Going to copy and elaborate on a question I posted in the Ancient Magus' Bride Ch. 96 thread, because I think a conversation about this (likely to expand) trend is worthy of having out in the open. It's worth noting beforehand that I've had a bit of time since I first asked the question to do more reading up on Mantra and their claim the engine is more than just supercharged MTL. The engine apparently actually breaks down the source image and identifies nontextual factors (gender of speakers is specifically mentioned) to better guide the engine in selecting the best output. In my mind this targets the current greatest advantage of a human translator--the possession of in situ and cultural knowledge that allows them to choose the better word/sentence in context. I doubt Mantra are quite there yet...but I find it likely they will be before we know it.

Back to the actual question:

I'd be interested in seeing an objective pro/con analysis of the Mantra-assisted translation. The claim on the Mantra press release with their partnership with Shueisha is that the error rate of translation in their model is ~1.6%. No source for that number, best I could see in the PR, but I would not expect much from a public-facing corpo press release. The real question is, does the alleged accuracy hold up on inspection?
-----
I've seen some comments that the translation suffers from odd phrasing or poor word choice. Fair criticism, although I see that as equally a fault of the editor/proofreader as of the AI. I should also note that Mantra itself claims that its LLM was trained on fan translations (hopefully under an appropriately structured NDA...but not hopeful on that end because piracy goodbad :fml:) so any deficiencies are all y'alls fault :lul:. No useful comments on what the "error rate" could possibly mean, so would appreciate actual insight into that figure beyond "it doesn't make sense, so it's wrong!".

Pixiv/Fanbox apparently has an autotranslation tool powered by Mantra? I don't have any experience with those sites, so...any stories about how well it works?

Also, a 2020 paper on the Mantra model, which I've only briefly scanned for the juicy gossip, indicates that the strength of their AI LLM over traditional MTL is that it is "context-aware" to better guide the model in selecting the correct words to ultimately use, by adding non-textual factors derived from OCR of the source image into the LLM. Not misgendering (wince) characters a la Ichigo Reader sounds pretty swell. I have no background in the statistical models they use to evaluate accuracy, so...I'll have to assume that the BLEU model is a widely accepted and appropriate one. Doesn't quite explain where that 1.6% figure comes from, alas. Expert input greatly appreciated.

-----
Back to the original point: if that alleged level of accuracy does hold up, even within a ballpark, that means that:
1) instead of paying a translator hundreds of dollars (or equivalent in JPY) for a translation that may take hours, they can pay pennies for one done with 98%+ accuracy in just seconds. Output/JPY jumps by several orders of magnitude (that's factors of ten/hundreds/thousands, for you who don't math). Or another way...imagine translating the entirety of Detective Conan at the same cost as a single human-translated chapter.
2) potentially more time spent on ED/PR once the raw AI translation is complete, but only a 1.6% (-ish) increase in the big scheme of things, so the savings from (1) should still vastly outweigh any increased ED/PR costs. Depending on how editing is currently structured and budgeted (and there should be ED/PR even if a human did the original TL) this may not increase costs at all. So...massive savings from (1) - negligible increased costs from (2) = some savings per chapter/title
3) Savings per title = more titles(?) (yeah...not gonna touch that discussion. Evil Corporation (R) has little relevance to whether AI translation is a good or bad thing)

Thus far all I've heard is complaints no better than an allergy to AI/MTL without any, you know, facts to back up that feeling. I'm no particular fan of AI, and I do worry about the loss of jobs coming from them (my own job category will likely disappear in 5-10 years at the current rate) but I hate it when a community decides things are bad based entirely on hearsay and ego. I want to see the bad, along with the good, and let it be judged accordingly.

TL;DR if there are objective analyses breakind down the actual performance of Mantra (or any competitor, honestly) for better or worse, go ahead and spill the tea please. Sources much appreciated. Practical experience in the translation and publication fields would be even better.
 
Dex-chan lover
Joined
Jan 8, 2023
Messages
1,636
Going to copy and elaborate on a question I posted in the Ancient Magus' Bride Ch. 96 thread, because I think a conversation about this (likely to expand) trend is worthy of having out in the open. It's worth noting beforehand that I've had a bit of time since I first asked the question to do more reading up on Mantra and their claim the engine is more than just supercharged MTL. The engine apparently actually breaks down the source image and identifies nontextual factors (gender of speakers is specifically mentioned) to better guide the engine in selecting the best output. In my mind this targets the current greatest advantage of a human translator--the possession of in situ and cultural knowledge that allows them to choose the better word/sentence in context. I doubt Mantra are quite there yet...but I find it likely they will be before we know it.

Back to the actual question:

I'd be interested in seeing an objective pro/con analysis of the Mantra-assisted translation. The claim on the Mantra press release with their partnership with Shueisha is that the error rate of translation in their model is ~1.6%. No source for that number, best I could see in the PR, but I would not expect much from a public-facing corpo press release. The real question is, does the alleged accuracy hold up on inspection?
-----
I've seen some comments that the translation suffers from odd phrasing or poor word choice. Fair criticism, although I see that as equally a fault of the editor/proofreader as of the AI. I should also note that Mantra itself claims that its LLM was trained on fan translations (hopefully under an appropriately structured NDA...but not hopeful on that end because piracy goodbad :fml:) so any deficiencies are all y'alls fault :lul:. No useful comments on what the "error rate" could possibly mean, so would appreciate actual insight into that figure beyond "it doesn't make sense, so it's wrong!".

Pixiv/Fanbox apparently has an autotranslation tool powered by Mantra? I don't have any experience with those sites, so...any stories about how well it works?

Also, a 2020 paper on the Mantra model, which I've only briefly scanned for the juicy gossip, indicates that the strength of their AI LLM over traditional MTL is that it is "context-aware" to better guide the model in selecting the correct words to ultimately use, by adding non-textual factors derived from OCR of the source image into the LLM. Not misgendering (wince) characters a la Ichigo Reader sounds pretty swell. I have no background in the statistical models they use to evaluate accuracy, so...I'll have to assume that the BLEU model is a widely accepted and appropriate one. Doesn't quite explain where that 1.6% figure comes from, alas. Expert input greatly appreciated.

-----
Back to the original point: if that alleged level of accuracy does hold up, even within a ballpark, that means that:
1) instead of paying a translator hundreds of dollars (or equivalent in JPY) for a translation that may take hours, they can pay pennies for one done with 98%+ accuracy in just seconds. Output/JPY jumps by several orders of magnitude (that's factors of ten/hundreds/thousands, for you who don't math). Or another way...imagine translating the entirety of Detective Conan at the same cost as a single human-translated chapter.
2) potentially more time spent on ED/PR once the raw AI translation is complete, but only a 1.6% (-ish) increase in the big scheme of things, so the savings from (1) should still vastly outweigh any increased ED/PR costs. Depending on how editing is currently structured and budgeted (and there should be ED/PR even if a human did the original TL) this may not increase costs at all. So...massive savings from (1) - negligible increased costs from (2) = some savings per chapter/title
3) Savings per title = more titles(?) (yeah...not gonna touch that discussion. Evil Corporation (R) has little relevance to whether AI translation is a good or bad thing)

Thus far all I've heard is complaints no better than an allergy to AI/MTL without any, you know, facts to back up that feeling. I'm no particular fan of AI, and I do worry about the loss of jobs coming from them (my own job category will likely disappear in 5-10 years at the current rate) but I hate it when a community decides things are bad based entirely on hearsay and ego. I want to see the bad, along with the good, and let it be judged accordingly.

TL;DR if there are objective analyses breakind down the actual performance of Mantra (or any competitor, honestly) for better or worse, go ahead and spill the tea please. Sources much appreciated. Practical experience in the translation and publication fields would be even better.
https://www.forbes.com/sites/jackke...-job-losses-at-ibm-and-chegg/?sh=2ae21c8975a4
 
Dex-chan lover
Joined
Jan 8, 2023
Messages
1,636
Going to copy and elaborate on a question I posted in the Ancient Magus' Bride Ch. 96 thread, because I think a conversation about this (likely to expand) trend is worthy of having out in the open. It's worth noting beforehand that I've had a bit of time since I first asked the question to do more reading up on Mantra and their claim the engine is more than just supercharged MTL. The engine apparently actually breaks down the source image and identifies nontextual factors (gender of speakers is specifically mentioned) to better guide the engine in selecting the best output. In my mind this targets the current greatest advantage of a human translator--the possession of in situ and cultural knowledge that allows them to choose the better word/sentence in context. I doubt Mantra are quite there yet...but I find it likely they will be before we know it.

Back to the actual question:

I'd be interested in seeing an objective pro/con analysis of the Mantra-assisted translation. The claim on the Mantra press release with their partnership with Shueisha is that the error rate of translation in their model is ~1.6%. No source for that number, best I could see in the PR, but I would not expect much from a public-facing corpo press release. The real question is, does the alleged accuracy hold up on inspection?
-----
I've seen some comments that the translation suffers from odd phrasing or poor word choice. Fair criticism, although I see that as equally a fault of the editor/proofreader as of the AI. I should also note that Mantra itself claims that its LLM was trained on fan translations (hopefully under an appropriately structured NDA...but not hopeful on that end because piracy goodbad :fml:) so any deficiencies are all y'alls fault :lul:. No useful comments on what the "error rate" could possibly mean, so would appreciate actual insight into that figure beyond "it doesn't make sense, so it's wrong!".

Pixiv/Fanbox apparently has an autotranslation tool powered by Mantra? I don't have any experience with those sites, so...any stories about how well it works?

Also, a 2020 paper on the Mantra model, which I've only briefly scanned for the juicy gossip, indicates that the strength of their AI LLM over traditional MTL is that it is "context-aware" to better guide the model in selecting the correct words to ultimately use, by adding non-textual factors derived from OCR of the source image into the LLM. Not misgendering (wince) characters a la Ichigo Reader sounds pretty swell. I have no background in the statistical models they use to evaluate accuracy, so...I'll have to assume that the BLEU model is a widely accepted and appropriate one. Doesn't quite explain where that 1.6% figure comes from, alas. Expert input greatly appreciated.

-----
Back to the original point: if that alleged level of accuracy does hold up, even within a ballpark, that means that:
1) instead of paying a translator hundreds of dollars (or equivalent in JPY) for a translation that may take hours, they can pay pennies for one done with 98%+ accuracy in just seconds. Output/JPY jumps by several orders of magnitude (that's factors of ten/hundreds/thousands, for you who don't math). Or another way...imagine translating the entirety of Detective Conan at the same cost as a single human-translated chapter.
2) potentially more time spent on ED/PR once the raw AI translation is complete, but only a 1.6% (-ish) increase in the big scheme of things, so the savings from (1) should still vastly outweigh any increased ED/PR costs. Depending on how editing is currently structured and budgeted (and there should be ED/PR even if a human did the original TL) this may not increase costs at all. So...massive savings from (1) - negligible increased costs from (2) = some savings per chapter/title
3) Savings per title = more titles(?) (yeah...not gonna touch that discussion. Evil Corporation (R) has little relevance to whether AI translation is a good or bad thing)

Thus far all I've heard is complaints no better than an allergy to AI/MTL without any, you know, facts to back up that feeling. I'm no particular fan of AI, and I do worry about the loss of jobs coming from them (my own job category will likely disappear in 5-10 years at the current rate) but I hate it when a community decides things are bad based entirely on hearsay and ego. I want to see the bad, along with the good, and let it be judged accordingly.

TL;DR if there are objective analyses breakind down the actual performance of Mantra (or any competitor, honestly) for better or worse, go ahead and spill the tea please. Sources much appreciated. Practical experience in the translation and publication fields would be even better.
there's no way i'll accept this statement "manga has "continuously suffered" damages from online piracy and illegal English translations"
 
Aggregator gang
Joined
Aug 4, 2019
Messages
59
I'd be interested in seeing an objective pro/con analysis of the Mantra-assisted translation. The claim on the Mantra press release with their partnership with Shueisha is that the error rate of translation in their model is ~1.6%. No source for that number, best I could see in the PR, but I would not expect much from a public-facing corpo press release. The real question is, does the alleged accuracy hold up on inspection?
Here's an objective analysis from someone who's read the shit they drop for $2/mo in their app: It's garbage lmao. It's not necessarily a matter of "accuracy," (it's obviously hard to tell if a translation is accurate without knowing and reading the source language) but rather it has zero ability to "voice" characters, so every single bit of dialog ends up sounding like an out of character summary of what they would be saying rather than feeling like an actual distinct character talking. The end product is essentially completely sucked dry of flavor except from what scraps you can glean from the art alone. Anything that's subtext will be lost. More than half of any given character's personality will be lost. Most jokes will fall flat, you might not even notice they were trying to be funny. It'll pass the time, you'll technically know the plot, but at that point you might as well just read a fandom wiki summary of the manga instead.
It's like the manga version of grinding up meat to make a slurry. You might enjoy the cheap hamburger on some level, but it'll never be anywhere near the steak it was before they dumped it in the grinder.
 
Last edited:
Contributor
Joined
Feb 4, 2018
Messages
7,924
Right.

I bought manga since a long time (20 years i guess, started with Kenshin, GTO, and Love hina), but i bought many more BECAUSE of the fantrad like everything we get here on dex.

I will not buy a manga randomly without reading at least some chapters first, or even before the end bcz a lot of manga, especially rom-coms, easily turn from ok-ish to pure trash.
 
Dex-chan lover
Joined
Sep 26, 2018
Messages
1,956
It's worth noting beforehand that I've had a bit of time since I first asked the question to do more reading up on Mantra and their claim the engine is more than just supercharged MTL.
Sora the Troll (a JP TLer, voice actor, our aho) showcased how to use ChatGPT4 to have good JP -> EN translations. I wouldn't put it past Mantra for not having actual human personnel to do Editing and PR when the AI itself can do it.
 
Dex-chan lover
Joined
Jan 9, 2023
Messages
2,239
Sora the Troll (a JP TLer, voice actor, our aho) showcased how to use ChatGPT4 to have good JP -> EN translations. I wouldn't put it past Mantra for not having actual human personnel to do Editing and PR when the AI itself can do it.
Editing and PR are still be handled at the publisher's level, I think, rather than Mantra's. It's a fair point, though, that with current technology there is a lot of financial pressure to cut the expensive humans out of the process streamline operations.

Here's an objective analysis from someone who's read the shit they drop for $2/mo in their app
Subjective hyperbole aside, do you have some concrete examples to back up your conclusions? I have seen comments in the Magus thread noting a decrease in the effective reading grade level of the AI output (vs a fan TL), which is the kind of analysis I would like to see. Not trying to knock your position here, just genuinely curious to see details beyond broad generalizations.
 
Aggregator gang
Joined
Aug 4, 2019
Messages
59
Editing and PR are still be handled at the publisher's level, I think, rather than Mantra's. It's a fair point, though, that with current technology there is a lot of financial pressure to cut the expensive humans out of the process streamline operations.


Subjective hyperbole aside, do you have some concrete examples to back up your conclusions? I have seen comments in the Magus thread noting a decrease in the effective reading grade level of the AI output (vs a fan TL), which is the kind of analysis I would like to see. Not trying to knock your position here, just genuinely curious to see details beyond broad generalizations.
I first noticed when going back to reread Golden Kamuy, a favorite series of mine that I first read with the excellent fan translation by Everyday Heroes here on mangadex:
https://mangadex.org/manga/8847f905-550d-4fe6-bcda-ac2b896789c7/golden-kamuy
For the reread I instead picked up the official VIZ translation of the manga:
https://www.viz.com/shonenjump/chapters/golden-kamuy
Compare these two side by side, and you'll notice a massive quality difference when it comes to dialog in particular. The Viz translation is absolutely garbage at making the character voices distinct from each other, and has quite a few lines that make no sense and look like they were translated in a vacuum without considering the context of the scene.
I cannot say for sure that Vis used machine translation tools to work on this, as they don't admit it, but it's got all the hallmarks of that kind of work. If you look around, you'll find a ton of licensed series that release chapter by chapter and try to come out asap do.

Edit: checking the Viz translation again, the gap isn't as obviously wide from the beginning, although it's still there. The biggest sign in chapter one is that the official translation refers to "those bodies sharing the same cell," which did not actually make sense in context. Those men were in prison together but absolutely not in the same cell, there's dozens of them lol. The fan translation says "those bodies that spent so much time suffering together," which makes far more sense.
After a while it got so bad that I recognized problems despite making the comparison from a years old memory instead of reading the two versions side by side. I distinctly remember a conversation where the official translation keeps using "trivial" to describe a shitty joke and it makes the dialogue incredibly confusing. If you're trying to say a joke is shitty and someone shouldn't make it, why would you call it "trivial"? The fan translation instead calls the joke "petty" and in general the dialog works way better. There's like four different people talking and in the official translation they all sound the same and use the same phrasing, even though they're arguing, but they're all distinct voices in the fan translation.
 
Last edited:
Dex-chan lover
Joined
Oct 9, 2019
Messages
1,988
Dex-chan lover
Joined
Jan 8, 2023
Messages
1,636
I mean I could theoretically use Google Translate to translate manga chapters and then clean up the output so that it sounds grammatically correct in English without knowing an ounce of Japanese and most people would be none the wiser
yup 100%
 
Dex-chan lover
Joined
Mar 17, 2018
Messages
362
its not that they don't know they're being willfully ignorant because you have smucks out there believing them so they can just use the whole piracy killed the industry woe is us sony love to use. Remember its not shitty TL quality, its not content being blatantly rewritten and definitely not content being cut out of international releases by egomaniacs, it's the pirates ruining everything.
 
Dex-chan lover
Joined
Jun 24, 2020
Messages
180
Going to copy and elaborate on a question I posted in the Ancient Magus' Bride Ch. 96 thread, because I think a conversation about this (likely to expand) trend is worthy of having out in the open. It's worth noting beforehand that I've had a bit of time since I first asked the question to do more reading up on Mantra and their claim the engine is more than just supercharged MTL. The engine apparently actually breaks down the source image and identifies nontextual factors (gender of speakers is specifically mentioned) to better guide the engine in selecting the best output. In my mind this targets the current greatest advantage of a human translator--the possession of in situ and cultural knowledge that allows them to choose the better word/sentence in context. I doubt Mantra are quite there yet...but I find it likely they will be before we know it.

Back to the actual question:

I'd be interested in seeing an objective pro/con analysis of the Mantra-assisted translation. The claim on the Mantra press release with their partnership with Shueisha is that the error rate of translation in their model is ~1.6%. No source for that number, best I could see in the PR, but I would not expect much from a public-facing corpo press release. The real question is, does the alleged accuracy hold up on inspection?
-----
I've seen some comments that the translation suffers from odd phrasing or poor word choice. Fair criticism, although I see that as equally a fault of the editor/proofreader as of the AI. I should also note that Mantra itself claims that its LLM was trained on fan translations (hopefully under an appropriately structured NDA...but not hopeful on that end because piracy goodbad :fml:) so any deficiencies are all y'alls fault :lul:. No useful comments on what the "error rate" could possibly mean, so would appreciate actual insight into that figure beyond "it doesn't make sense, so it's wrong!".

Pixiv/Fanbox apparently has an autotranslation tool powered by Mantra? I don't have any experience with those sites, so...any stories about how well it works?

Also, a 2020 paper on the Mantra model, which I've only briefly scanned for the juicy gossip, indicates that the strength of their AI LLM over traditional MTL is that it is "context-aware" to better guide the model in selecting the correct words to ultimately use, by adding non-textual factors derived from OCR of the source image into the LLM. Not misgendering (wince) characters a la Ichigo Reader sounds pretty swell. I have no background in the statistical models they use to evaluate accuracy, so...I'll have to assume that the BLEU model is a widely accepted and appropriate one. Doesn't quite explain where that 1.6% figure comes from, alas. Expert input greatly appreciated.

-----
Back to the original point: if that alleged level of accuracy does hold up, even within a ballpark, that means that:
1) instead of paying a translator hundreds of dollars (or equivalent in JPY) for a translation that may take hours, they can pay pennies for one done with 98%+ accuracy in just seconds. Output/JPY jumps by several orders of magnitude (that's factors of ten/hundreds/thousands, for you who don't math). Or another way...imagine translating the entirety of Detective Conan at the same cost as a single human-translated chapter.
2) potentially more time spent on ED/PR once the raw AI translation is complete, but only a 1.6% (-ish) increase in the big scheme of things, so the savings from (1) should still vastly outweigh any increased ED/PR costs. Depending on how editing is currently structured and budgeted (and there should be ED/PR even if a human did the original TL) this may not increase costs at all. So...massive savings from (1) - negligible increased costs from (2) = some savings per chapter/title
3) Savings per title = more titles(?) (yeah...not gonna touch that discussion. Evil Corporation (R) has little relevance to whether AI translation is a good or bad thing)

Thus far all I've heard is complaints no better than an allergy to AI/MTL without any, you know, facts to back up that feeling. I'm no particular fan of AI, and I do worry about the loss of jobs coming from them (my own job category will likely disappear in 5-10 years at the current rate) but I hate it when a community decides things are bad based entirely on hearsay and ego. I want to see the bad, along with the good, and let it be judged accordingly.

TL;DR if there are objective analyses breakind down the actual performance of Mantra (or any competitor, honestly) for better or worse, go ahead and spill the tea please. Sources much appreciated. Practical experience in the translation and publication fields would be even better.

Learn from Steam and Valve. Good platform for private creations and professional. At the same time make the platform user friendly [cheap, free and dope]. It's a fucking media platform it's not that hard ffs

 
Last edited:
Dex-chan lover
Joined
Jun 24, 2020
Messages
180
@your confusion

Before Steam piracy of PC-games were HUGE. So rampart infact that many big company's stopped developing for PC. But because Valve focused on the user experiance instead the sellers over time PC-game piracy was in all but fringe minoritys eradicated. "The easiest way to stop piracy is not by putting antipiracy technology to work. It's by giving those people a service that's better than what they're receiving from the pirates." - Gabe Newell


So if manga companys doesnt want to continuously suffer damage from online piracy and illegal English translation they have to step up their game and provide a better service then the pirates.

 
Dex-chan lover
Joined
Jan 25, 2023
Messages
6,337
More mangaka and even illustrators are suffering from shot damage than from piracy itself. See the increase of hiatus "due to health reasons".

Before Steam piracy of PC-games were HUGE. So rampart infact that many big company's stopped developing for PC. But because Valve focused on the user experiance instead the sellers over time PC-game piracy was in all but fringe minoritys eradicated. "The easiest way to stop piracy is not by putting antipiracy technology to work. It's by giving those people a service that's better than what they're receiving from the pirates." - Gabe Newell


So if manga companys doesnt want to continuously suffer damage from online piracy and illegal English translation they have to step up their game and provide a better service then the pirates.

If you kill the thieves, there would be no more stealing.
 
Dex-chan lover
Joined
Jan 25, 2023
Messages
6,337
manga would be no where outside japan if it wasnt for english translators and sites like this. free exposure is great i dont know what much else to say.
It exposed Japan as a den of pedophiles, preying on 16 yo and less and is seen much repeated drafts of Usagi Drop, Uncle - middle school niece grooming, etc. In addition to their obsession to high school students in mini skirts. Not to mention their masterpiece: Boku no Pico.

The TL's did the world a favor of not buying such ... banes. In other countries, having them is already punishable by being burned alive.
 
Active member
Joined
Aug 6, 2020
Messages
50
manga would be no where outside japan if it wasnt for english translators and sites like this. free exposure is great i dont know what much else to say.
not true.
manga didnt start overseas through volunteered translations.

scanlation took off in the late 2000's when dial-in finally evolved into modern unlimited internet data plans.
and it was powered by things like GTO, yugioh, one piece, naruto and other big early 2000's franchises having a several month delay (at the time) between japanese and overseas publication.

it started because european and american publishers found their potential in the 1980's.
and they began litterally gobbling anything drawn in japan, as they got them cheap and in large quantities.

and that even led to pre-publishing magazines being offered.

only after that we had conservation efforts, trying to save or share mangas who werent published or got translated but fell into obscurity.

on my market, which is one of the broadest in the world, the main problem is that volume cost as soared well above 8 dollarydoos a volume and when i was a teen it was below 6, so... yeah. that's a 20% increase that def doesnt show in print quality.
 

Users who are viewing this thread

Top