One one hand, the troonslators will be automated away completely.
On the other, the bots might be told to impose largely the same values.
On the third hand, pirates can auto-translate anime themselves without that limitation, and a lot of the troonslators' leeway comes from the fact that they don't need to be explicitly told to enshittify everything they touch.
I had a conference with a French AI voiceover and it was totally intelligible.
Someone just has to bother to listen to the output and hold the hand of the AI to correct uncanny / incorrect parts.
Troon ''localizers'' can go fuck themselves dilate, they are being replaced by automation, and any sympathy I might have will be strictly limited to based content.
The AI will always struggle for the same reason the AI does in every category. People assume its magical, or aren't paid enough to give a shit, so they do zero vetting or editing on it once its done to make sure the parts where it failed to understand the intention are corrected.
It will still need multiple takes to get some scenes right, same as a human, simply because it needs directing on the correct emotion and all that jazz. But as we can already see, companies aren't going to do that so the final product always ends up slop trash.
Now, unofficial works will likely be much improved to the point of superior because passion projects will have guys willing to do that work. Those are the guys that make the AI art right now that is indistinguishable because they've spent hours messing with the prompts and rejecting dozens of lesser versions before considering it done.
They don't just mess with prompts. Proper AI artists will feed the image back through the system afterwards for minor variations, as well. Beyond this, they'll highlight segments of the image (often hands, but anything) and select-edit JUST those elements, akin to a photoshopping session... And some even straight-up just use photoshop to gussy up the image afterwards, too.
"AI Art" is like saying "Paint art" or "pencil art" or "chalk art". It's a medium and a tool that an artist uses.
The issue is when a non-artist uses it, and effectively tries to pass off low-effort crayon drawings as high art.
Yeah, if someone goes the full mile to "perfect" the work its very much a passion project that requires considerable amounts of work, though maybe not exactly "skilled" work the same way non-AI art creation would. Though you can certainly combine those two to make some truly great works as well.
One of the issues with this entire ethical debate is that the guy saying "draw this girl giving birth" in some program is getting treated as equal to someone going that extreme mile, despite at some point it clearly becoming a creative endeavor all on its own.
You won't convince people with hate boners over it otherwise, but at the same time too many companies/creators have also gone way too "all in" on it without the proper controls/vetting to make it seamless and functional on any level beyond cost cutting.
The downvotes in this thread are hilarious, but everything you say is spot on.
AI is the future, and as a toolset for improving pipeline efficiency, it's an absolute God-send.
The amount of work it nullifies (in the right hands) is uncanny. Good artists -- as you note -- will use it to drastically speed up a number of ways in which voice acting, music composition, sound editing, and animation cells are produced, not unlike how fill-frames in 3D animation drastically speeds up the keyframe animation process.
Every time I mention this in animation communities there are liberal faggots everywhere decrying how AI is just "slop" and would never be used for animations. Within a few years anyone who isn't using AI as part of their toolkit within their workflow will be left behind.
I'm just waiting on Japanese animators to realise they can finally get back to making fluid, highly detailed works like in the late 80s and early 90s in cost-effective ways, all thanks to he help of AI filling in frames instead of working those poor animators to the bone for 16 hours a day.
The same ones who downvote, also would have downvoted the concept of typewriter. It removes all "soul" from calligraphy! It conglomerates a hundred different ways of people writing letters into one frankensteinian too-clean typeset! It's robbing the important jobs of calligraphists and copysetters!
And yes, it is, and does. And yet, they'll use a computer keyboard over writing by hand, 99.99999% of the time.
It is nowhere near as bad as a shithole like reddit, but even on this board there are plenty of people who have a stick up their ass about AI. Two things can be true at the same time:
The current (and forthcoming) iterations of AI are revolutionary in their ability to increase productivity in capable hands. Most people shitting on it act like it is a failure because it isn't perfect or because it doesn't fit their preconceived notion of what AI is that they picked up from sci-fi. But its imperfections do not nullify the sheer magnitude of productivity increase one can extract from it if they are capable.
"AI" has become a buzzword and the vast majority of people and companies are not utilizing it effectively. The insane levels of capital expenditure directed toward data centers is unlikely to be worth it in the long run. The AI craze has formed a bubble and the pop and resulting fallout will be interesting to watch.
In theory you should have one person overlooking to make sure it still works, but we all know these companies are too cheap to do that so it will take a catastrophic failure to get them to hire an overseer.
Probably should have a few really, both in terms of effeciency and extra sets of eyes, relative to the size of the project, but what keeps happening is one guy does literally everything using a prompt/program left over from another project.
It is continually amusing that the same problem you might see on a lazy MTL scanlator working by himself on a niche series repeats with billion dollar corps working on major properties.
To be fair, kanji can be a bitch if your eyesight isn't the best, especially with some of the 12+ stroke characters. When the difference between two kanji is a single line inside a triple compound character, you're basically doing a mix of figuring it out from context and matching the character's general shape.
But aside from that, it's not that bad. Just takes time and practice, like any skill.
No pity from me after how they treated Vic. That man got crucified and his life's work ruined for doing nothing wrong.
Hope the likes of Monica Rial, Jamie Marchi, Ron Toye, Kara Edwards, Michele Specht and even though she wasn't involved, Grey DeLisle suffer bigly from this.
Us: "AI will make the weaker actors have to raise their game or get replaced with AI. Sink or swim, bro, sink or swim!"
Reality: "Actors cost money, bro. Just use AI. Do you really think a trillion-dollar company has $200 laying around to pay an actor for an hour of work? That's why you're poor, brokie."
Yeah, I never could manage to watch them as they always sounded so awful while the Japanese felt much more natural. I assumed it was just for people who couldn’t be bothered to read.
Yeah, although I will admit the handful of good dubs I have watched I have noticed I can really take the time to enjoy the background details more. If dubs were half as good as subs I would be willing to admit they're worthwhile.
Given we're not beholden to 22-minute timesegments anymore, everything is online, they definitely can. Most of the stilted and bad voice acting is due to lefty loons, yes, but mechanically it's due to needing to match lipflaps, scene timings, and flows, from two VERY different languages. "Gomenasai", 4 syllables. "Sorry", 2 syllables. "I am sorry" -> Same meaning and syllables, but adding the words changes the tone and flow, a casual person wouldn't say it like that.
But in the age of streaming and on-demand, where every episode of Hazbin Hotel, in example, is a different length because it's streaming-only so why not, they could totally, at least for "profitable" or "flagship" anime, adjust talking scenes, and if the episode winds up being a minute longer or shorter... Oh well! Let the voice actors act their hearts out, and then edit the images to match, not the other way around, and the voice acting quality will go way up.
Obviously, this "obvious solution" has problems, such as Shonen talking-during-fight-scenes, which would be more effort to edit, but some middle ground adjustments, being free from needing to be exactly 22 minutes and 30 seconds, makes sense in the modern streaming format.
Which is probably why they went this way with it. Its usually garbage that nobody watches, so they went with the lowest investment to fulfill their contract and just put it on the site knowing basically nobody would buy it anyway.
Namely that Macnamaran leadership has taken hold of most companies. They're obsessed with quantitative analytics and have abandoned qualitative.
So they'll jump at the chance for fifty percent savings on twenty percent quality. Because it looks better on the spreadsheet for one, and for two the average prole can barely tell the difference anyway.
my favorite is Frank Sinatra singing "Cruel Angel's Thesis", also from two years ago, since it's proper for the conversion: English voice sampling, Japanese output. Proper AI dubbing.
There are some errors, and native Japanese-born do note that "he has a strong accent", but that's the limit of the complaint, and it's more just observation. Proper AI dubbing is possible, the issues are as they've always been: Lazy/hurried work, combined with needing to "localize". It doesn't matter how good your AI dub is, if your dragon maid is talking about the patriarchy's gaze instead of her shopping trip.
Is it added option where they previously is no available English dub? In that case I'm fine with it but I agree Ai voice still needs lots of work it lacks depth and proper emotions
One one hand, the troonslators will be automated away completely.
On the other, the bots might be told to impose largely the same values.
On the third hand, pirates can auto-translate anime themselves without that limitation, and a lot of the troonslators' leeway comes from the fact that they don't need to be explicitly told to enshittify everything they touch.
I had a conference with a French AI voiceover and it was totally intelligible.
Someone just has to bother to listen to the output and hold the hand of the AI to correct uncanny / incorrect parts.
Troon ''localizers'' can go
fuck themselvesdilate, they are being replaced by automation, and any sympathy I might have will be strictly limited to based content.The AI will always struggle for the same reason the AI does in every category. People assume its magical, or aren't paid enough to give a shit, so they do zero vetting or editing on it once its done to make sure the parts where it failed to understand the intention are corrected.
It will still need multiple takes to get some scenes right, same as a human, simply because it needs directing on the correct emotion and all that jazz. But as we can already see, companies aren't going to do that so the final product always ends up slop trash.
Now, unofficial works will likely be much improved to the point of superior because passion projects will have guys willing to do that work. Those are the guys that make the AI art right now that is indistinguishable because they've spent hours messing with the prompts and rejecting dozens of lesser versions before considering it done.
They don't just mess with prompts. Proper AI artists will feed the image back through the system afterwards for minor variations, as well. Beyond this, they'll highlight segments of the image (often hands, but anything) and select-edit JUST those elements, akin to a photoshopping session... And some even straight-up just use photoshop to gussy up the image afterwards, too.
"AI Art" is like saying "Paint art" or "pencil art" or "chalk art". It's a medium and a tool that an artist uses.
The issue is when a non-artist uses it, and effectively tries to pass off low-effort crayon drawings as high art.
Yeah, if someone goes the full mile to "perfect" the work its very much a passion project that requires considerable amounts of work, though maybe not exactly "skilled" work the same way non-AI art creation would. Though you can certainly combine those two to make some truly great works as well.
One of the issues with this entire ethical debate is that the guy saying "draw this girl giving birth" in some program is getting treated as equal to someone going that extreme mile, despite at some point it clearly becoming a creative endeavor all on its own.
You won't convince people with hate boners over it otherwise, but at the same time too many companies/creators have also gone way too "all in" on it without the proper controls/vetting to make it seamless and functional on any level beyond cost cutting.
The downvotes in this thread are hilarious, but everything you say is spot on.
AI is the future, and as a toolset for improving pipeline efficiency, it's an absolute God-send.
The amount of work it nullifies (in the right hands) is uncanny. Good artists -- as you note -- will use it to drastically speed up a number of ways in which voice acting, music composition, sound editing, and animation cells are produced, not unlike how fill-frames in 3D animation drastically speeds up the keyframe animation process.
Every time I mention this in animation communities there are liberal faggots everywhere decrying how AI is just "slop" and would never be used for animations. Within a few years anyone who isn't using AI as part of their toolkit within their workflow will be left behind.
I'm just waiting on Japanese animators to realise they can finally get back to making fluid, highly detailed works like in the late 80s and early 90s in cost-effective ways, all thanks to he help of AI filling in frames instead of working those poor animators to the bone for 16 hours a day.
The same ones who downvote, also would have downvoted the concept of typewriter. It removes all "soul" from calligraphy! It conglomerates a hundred different ways of people writing letters into one frankensteinian too-clean typeset! It's robbing the important jobs of calligraphists and copysetters!
And yes, it is, and does. And yet, they'll use a computer keyboard over writing by hand, 99.99999% of the time.
It is nowhere near as bad as a shithole like reddit, but even on this board there are plenty of people who have a stick up their ass about AI. Two things can be true at the same time:
The current (and forthcoming) iterations of AI are revolutionary in their ability to increase productivity in capable hands. Most people shitting on it act like it is a failure because it isn't perfect or because it doesn't fit their preconceived notion of what AI is that they picked up from sci-fi. But its imperfections do not nullify the sheer magnitude of productivity increase one can extract from it if they are capable.
"AI" has become a buzzword and the vast majority of people and companies are not utilizing it effectively. The insane levels of capital expenditure directed toward data centers is unlikely to be worth it in the long run. The AI craze has formed a bubble and the pop and resulting fallout will be interesting to watch.
In theory you should have one person overlooking to make sure it still works, but we all know these companies are too cheap to do that so it will take a catastrophic failure to get them to hire an overseer.
Probably should have a few really, both in terms of effeciency and extra sets of eyes, relative to the size of the project, but what keeps happening is one guy does literally everything using a prompt/program left over from another project.
It is continually amusing that the same problem you might see on a lazy MTL scanlator working by himself on a niche series repeats with billion dollar corps working on major properties.
And they said cold fusion was 50 years away for the last 70 years.
It will never learn to emote on the level of Jessica Calvello or Spike Spencer.
Even if they get dubbing on par with humans - they'll still suck.
Buy raws from japan. Fund studios that don't bend the knee.
Learn Japanese.
Problems solved, worries eliminated.
You will never learn japanese.
はい, 俺は大ばかだよ.
To be fair, kanji can be a bitch if your eyesight isn't the best, especially with some of the 12+ stroke characters. When the difference between two kanji is a single line inside a triple compound character, you're basically doing a mix of figuring it out from context and matching the character's general shape.
But aside from that, it's not that bad. Just takes time and practice, like any skill.
This is the way. The hard way usually is, but few there be who take it.
No pity from me after how they treated Vic. That man got crucified and his life's work ruined for doing nothing wrong.
Hope the likes of Monica Rial, Jamie Marchi, Ron Toye, Kara Edwards, Michele Specht and even though she wasn't involved, Grey DeLisle suffer bigly from this.
Us: "AI will make the weaker actors have to raise their game or get replaced with AI. Sink or swim, bro, sink or swim!"
Reality: "Actors cost money, bro. Just use AI. Do you really think a trillion-dollar company has $200 laying around to pay an actor for an hour of work? That's why you're poor, brokie."
Most English dubs are horrendous anyway. Horrendous is something AI is actually capable of so in this case it's not much of a loss.
Yeah, I never could manage to watch them as they always sounded so awful while the Japanese felt much more natural. I assumed it was just for people who couldn’t be bothered to read.
Yeah, although I will admit the handful of good dubs I have watched I have noticed I can really take the time to enjoy the background details more. If dubs were half as good as subs I would be willing to admit they're worthwhile.
Given we're not beholden to 22-minute timesegments anymore, everything is online, they definitely can. Most of the stilted and bad voice acting is due to lefty loons, yes, but mechanically it's due to needing to match lipflaps, scene timings, and flows, from two VERY different languages. "Gomenasai", 4 syllables. "Sorry", 2 syllables. "I am sorry" -> Same meaning and syllables, but adding the words changes the tone and flow, a casual person wouldn't say it like that.
But in the age of streaming and on-demand, where every episode of Hazbin Hotel, in example, is a different length because it's streaming-only so why not, they could totally, at least for "profitable" or "flagship" anime, adjust talking scenes, and if the episode winds up being a minute longer or shorter... Oh well! Let the voice actors act their hearts out, and then edit the images to match, not the other way around, and the voice acting quality will go way up.
Obviously, this "obvious solution" has problems, such as Shonen talking-during-fight-scenes, which would be more effort to edit, but some middle ground adjustments, being free from needing to be exactly 22 minutes and 30 seconds, makes sense in the modern streaming format.
Which is probably why they went this way with it. Its usually garbage that nobody watches, so they went with the lowest investment to fulfill their contract and just put it on the site knowing basically nobody would buy it anyway.
It's hard to say if this is worse than 4kids
Funny enough it's the same problem as jeets.
Namely that Macnamaran leadership has taken hold of most companies. They're obsessed with quantitative analytics and have abandoned qualitative.
So they'll jump at the chance for fifty percent savings on twenty percent quality. Because it looks better on the spreadsheet for one, and for two the average prole can barely tell the difference anyway.
Guessing the Board of Directors wants to give themselves a bonus with all the money they save, and maybe take an Epstein Island vacation with it.
Apparently they used DEI hires or something to do the conversion. One notch above Mecc Opening Night.
For good AI voice, here is a video from two years ago of a radio era Orsen Welles reading a 4chan copypasta.
https://youtu.be/K-TeECmt0Nc
my favorite is Frank Sinatra singing "Cruel Angel's Thesis", also from two years ago, since it's proper for the conversion: English voice sampling, Japanese output. Proper AI dubbing.
And it sounds just fine. https://www.youtube.com/watch?v=FLmt6dQmMTU
There are some errors, and native Japanese-born do note that "he has a strong accent", but that's the limit of the complaint, and it's more just observation. Proper AI dubbing is possible, the issues are as they've always been: Lazy/hurried work, combined with needing to "localize". It doesn't matter how good your AI dub is, if your dragon maid is talking about the patriarchy's gaze instead of her shopping trip.
I don't actually believe dubfags have the discernment to realise this is bad.
notmyproblem.gif
Didn't realize there was a NGNL movie.
For once, something useful has come from the whining of dub watchers.
It's a dub, it had no chance to be anything but horrible.
The problem seems to be it's either too early to use as not fully developed or they're just applying it in a 'that'll do' methodology.
It'd just be easier to tell the woke dub VAs to kick rocks, get in proper Dub artists as they develop AI to be better in the background.
>dub
Shiggy diggy
Is it added option where they previously is no available English dub? In that case I'm fine with it but I agree Ai voice still needs lots of work it lacks depth and proper emotions
This sounds almost as bad as polish lektor of Tomasz Knapik but without the hilarious contracst.
Still better than danger hairs and trannies
Back in my day, you'd be laughed at if you watched the dub.