Does Nozawa have an understudy?

[Cure Dragon 255]

Ok,but I still agree with them that it wont be as easy as Yuli says it will be and that it probably be really expensive. And even if it wasnt there are laws that forbid this. Robot Jones anyone?

[Dr. Casey]

Ok,but I still agree with them that it wont be as easy as Yuli says it will be and that it probably be really expensive.

Given that the concept is probably brand new to you, I would give more consideration to the opinion of someone who's been keeping up with the subject and reading up on it for years. The technology will be available and affordable by the end of the next decade, though that doesn't mean it will put seiyuus are put out of work; these are beloved personalities who fans enjoy seeing and hearing.

[Cure Dragon 255]

I guess you are right. I really dont know the extent of the evolution of tecnology. Sorry for being so rude.

[Dr. Casey]

I guess you are right. I really dont know the extent of the evolution of tecnology. Sorry for being so rude.

Don't feel bad, I've kind of been a dismissive ass and you haven't been rude at all.

But as for an understudy, I think it's entirely likely that Nozawa has one; she owned a talent agency from 2006 to 2012.

[Michsi]

I think there is a side to this debate that people are omitting, namely that VA's are pretty much considered celebrities, especially in Japan, and that is a Business in and of itself. They're pretty much treated as idols and their name alone can bring value to an anime that has them, and they are even used to promote the shows that they are in. They sell singles, merchandise and appear on talk shows, just like any other type performance artist.

So whatever financial benefit may come from replacing them with AIs would probably not be worth losing what VA's brings to the industry in other ways.

[Robo4900]

I think there is a side to this debate that people are omitting, namely that VA's are pretty much considered celebrities, especially in Japan, and that is a Business in and of itself. They're pretty much treated as idols and their name alone can bring value to an anime that has them, and they are even used to promote the shows that they are in. They sell singles, merchandise and appear on talk shows, just like any other type performance artist.

So whatever financial benefit may come from replacing them with AIs would probably not be worth losing what VA's brings to the industry in other ways.

Indeed. That is one thing I touched on that people seem to have largely ignored.
Seeing things like "Starring Ian McKellen, Patrick Stewart, Benedict Cumberbatch, and Mr. T" draw big numbers in major movies, and a similar thing applies for voicework. Everyone loves Tara Strong, Rob Paulsen, etc., so cutting VAs out of the equation would be shooting yourself in the foot in a lot of ways.
Plus, once again, I don't think unions would allow it at all.

That's just it, though. The fact Rogue One's synthetic material looked so bad and yet we're already getting to a point where we're freaking out these things is more or less a sign of how fast things are progressing.

It isn't. Not even close.

CG Tarkin wasn't just a bunch of computers assembling something, and it wasn't a replacement of an actor at all. An actor performed the entire role in full performance capture gear, and that information was then loaded into a computer, where a manually-created Tarkin model had its movements mapped to the performance of the other actor. But even then, it's not quite right, because piles and piles of manual adjustments had to be made to get what we saw on-screen to look any good.
Finally, the voice, the ultimate crux of this whole discussion; that wasn't generated by a computer, not even close, the actor doing the performance had to spend a long time perfecting the voice for the role. CG Tarkin wasn't a computer replacing the job of an actor, CG Tarkin required the work of a talented actor with the ability to imitate Cushing's mannerisms and voice, and several huge teams of digital artists. CG Tarkin actively works against the point you're trying to make here.

[Michsi]

I think there is a side to this debate that people are omitting, namely that VA's are pretty much considered celebrities, especially in Japan, and that is a Business in and of itself. They're pretty much treated as idols and their name alone can bring value to an anime that has them, and they are even used to promote the shows that they are in. They sell singles, merchandise and appear on talk shows, just like any other type performance artist.

So whatever financial benefit may come from replacing them with AIs would probably not be worth losing what VA's brings to the industry in other ways.

Indeed. That is one thing I touched on that people seem to have largely ignored.
Seeing things like "Starring Ian McKellen, Patrick Stewart, Benedict Cumberbatch, and Mr. T" draw big numbers in major movies, and a similar thing applies for voicework. Everyone loves Tara Strong, Rob Paulsen, etc., so cutting VAs out of the equation would be shooting yourself in the foot in a lot of ways.
Plus, once again, I don't think unions would allow it at all.
.

The recent focus on physical appearance definitely shows that they are trying to promote many of them as celebrities and idols, be that for better or worse. And yes, people often forget about unions too when talking about machines replacing workers. Even if the technology is there, it's not that easy.

[Dr. Casey]

Finally, the voice, the ultimate crux of this whole discussion; that wasn't generated by a computer, not even close, the actor doing the performance had to spend a long time perfecting the voice for the role. CG Tarkin wasn't a computer replacing the job of an actor, CG Tarkin required the work of a talented actor with the ability to imitate Cushing's mannerisms and voice, and several huge teams of digital artists. CG Tarkin actively works against the point you're trying to make here.

Rogue One was released a year and a half ago, which for the purposes of this conversation was a long time ago. The first noteworthy leap forward I ever noticed with voice generation was with the Google Duplex video I linked earlier in the thread that was only reported on towards the beginning of this month.

I recognize that this all sounds like a bunch of delusional crazy talk to people hearing about this stuff for the first time, so I won't waste time arguing about the technical feasibility of any of this (especially since the more societal-based facets that you and Michsi bring up, which I mentioned myself, are both valid points). This is roughly the equivalent of someone who had never heard about the internet being told in 1990 by someone who'd been watching it slowly develop during the 1980s that it would be a service used by a couple hundred million people by the decade's end. "Uh... this whole internet thing sounds weird. Why would something like that ever become popular?" The technology had been maturing in the background and plenty of successes had already been achieved, but because Average Non-Nerdy Guy 1990 wasn't aware of them, to him none of that mattered and the pro-internet guy was just saying stupid, bizarre things. This thread is a good example of why I don't bother talking about technology to most people that aren't into it - if something sounds ridiculous or outlandish on its face (and I realize that this does), people generally just tune you out. Not speaking to Robo personally so much since he is at least trying to engage, just commenting on the tenor of the thread in general.

Honestly, though, even if this technology didn't make a single dent on the corporate world, the real magic is the extra power that it will grant to regular, everyday people creating amateur projects. Like I have a long, 2.5 million word story that I think would be fun to convert to Visual Novel format during the 2020s. I could create character sprites by myself using the in-program editor, and backgrounds would be no problem (I like using real life photographs as backdrops to increase the sense of immersion), but voice acting would be way too ambitious of a project. Or at least, it would be right now - by the time I get around to the voice acting section around the middle 2020s, resources will be available that can help with that (and no, they won't be robotic, flat-lined voices, before someone gives some tired remark along those lines - voice synthesis in the lab is already very human-like and authentic, and by the middle 2020s should be out there in the wild for anyone to use).

That's the real beauty of the technology, I think - democratizing the voice acting process so that it can be available to everyone. Businesses and corporations can do things the way they always have (or not, who knows, that will be up for everyone to decide when the time comes), whereas small groups and teams of people can - if they so wish or if it's necessary - give good quality voice acting to their projects without worrying about people flaking out and quitting halfway through, and people like myself can single-handedly create voiced dialogue without fretting and worrying about the logistics of "why would anyone help out a complete nobody like myself voice this story?"

[Danfun64]

Voice acting as a career has maybe a decade and a half left of life in it, if that. And considering that many animated shows don't have high budgets to begin with and Japan is already thrifty as it is (IIRC the costs for an episode of an anime is a third of that of an American cartoon?), VAs for them are probably not going to survive the end of the 2020s.
...
TL;DR: Machine learning can replicate human voices. Not only that, but it can replicate any voice within a minute and it can do so to such a disturbingly realistic level that some people now have trouble telling the difference between a human voice and a synthesized one. This is all happening at an accelerated rate— we could not have predicted such extreme progress even two years ago, and it's still progressing forward faster than we may be comfortable with knowing. If you told me in 2016 that voice actors may be out of a job en masse by 2028 thanks to AI, even I would have said that was foolishly optimistic because surely computers can't hit the right subtle intonations, cadences, timbres, etc. Now I know better and, considering how few people are also following this, I dread the day it slams into our reality because we are going to have a societal-wide nervous breakdown.

Here's a video someone made on a related topic that might be relevant. https://www.youtube.com/watch?v=XNgWj5LURDs

[majinwarman]

Voice acting as a career has maybe a decade and a half left of life in it, if that. And considering that many animated shows don't have high budgets to begin with and Japan is already thrifty as it is (IIRC the costs for an episode of an anime is a third of that of an American cartoon?), VAs for them are probably not going to survive the end of the 2020s.
...
TL;DR: Machine learning can replicate human voices. Not only that, but it can replicate any voice within a minute and it can do so to such a disturbingly realistic level that some people now have trouble telling the difference between a human voice and a synthesized one. This is all happening at an accelerated rate— we could not have predicted such extreme progress even two years ago, and it's still progressing forward faster than we may be comfortable with knowing. If you told me in 2016 that voice actors may be out of a job en masse by 2028 thanks to AI, even I would have said that was foolishly optimistic because surely computers can't hit the right subtle intonations, cadences, timbres, etc. Now I know better and, considering how few people are also following this, I dread the day it slams into our reality because we are going to have a societal-wide nervous breakdown.

Here's a video someone made on a related topic that might be relevant. https://www.youtube.com/watch?v=XNgWj5LURDs

That video was great.

[Mister_Popo]

I wonder how long we can expect from Nozawa to go along. She will be in the new Heroes anime as well in the movie. But what after that? If Super would get a second 'half' that's just as long as the first half, and run for another 3 years, would be it realistic to expect this from a 85-year-old-lady? I could appreciate it if she'd still do it, but i'd accept she wouldn't do another 150 episodes. Maybe i'd be better to start with a 'new voice' when the series return, in stead of she has to be replaced due to health issues along the way. She deserves to be retired at a certain point as well, no matter how much we would like she'd keep going.

I hope she does have an understudy, and the actor or actress who replaces her, is able to mimic her voice as accurate as possible. I don't think they want to take chances with such an iconic role by replacing her with an actor that sounds completely different.

[8000 Saiyan]

I wouldn't be surprised if she had one.

[DragonBallFoodie]

I was thinking she has an experienced voice actor to take over, but an understudy sounds likely too.

Yasuo Yamada, Lupin III's voice actor, had Kanichi Kurita as an understudy and replacement (which happened when Yamada passed away in 1995). Kurita hasn't done any major non-Lupin roles.

[Mister_Popo]

In Japan Bulma's new voice was revealed in a commercial for cleaning products. She does sound very simular to old Bulma, i barely hear any difference, which indicates she probably had an understudy as well. Wouldn't be surprised they showed Goku's new voice before an actual episode aired, in order to test the waters.