[POLL] YS Net using AI to get Shenmue 4 made

Good or bad idea?

  • Good idea

    Votes: 20 55.6%

  • Bad idea

    Votes: 8 22.2%

  • I have no strong feelings one way or the other

    Votes: 8 22.2%
  • Total voters
    36
Dietsoap

Dietsoap

Joined
Aug 18, 2018
There's been a lot of talk lately about the possibility of YS Net using AI to generate development assets. The idea is to lower costs, with the hope of bringing the entire project closer to a palatable financial sum for potential investors. What do you guys think? Good or bad idea?

I tried that new Bing DALL-E based AI image generator last night. I'm not sure what compelled me to add "hotdog" in retrospect, but I think that was my undoing.

1696171393674.png

Was this thread really just an excuse to post these shitty images and see what other Dojo lads create? It's possible. Link if anyone wants to try generating some Shenmue images with this: https://www.bing.com/images/create
 
It's really an evolution of the Magic Maze system in some respects. I wouldn't mind as long as it was used well.
 
If I have to choose... I have a feeling that right now I'm going to say an answer that I might regret later. :)

Right now I really want a new game, but not necessarily with endless compromises.

On the other hand, even if something is not very high-quality at the beginning, there is always a chance that fan created mods will improve the game in some aspects. I'm saying yes now and hope I don't regret it later. :ROFLMAO:
 
I'm not voting, because the answer I want isn't there.

If it's done for minor things such minor NPC's and stuff, I don't see the harm, but anything beyond that then no way.

As for the magic maze; isn't there some kind of programming trick that allows for a random layout (think this was the case for the Yellow Head Building?)
 
Generative AI has come a long way, but I still feel like it's at a point where its inclusion in an actual production is tacky and overall detrimental, especially in the case of text to speech.

Even using it for background NPCs would be a step too far for me personally. It just isn't there yet and I would find it more immersion breaking than just a text box with no VO.

There's also the moral issue of if it's really OK to use this kind of tech in commercial settings. I've got thoughts on both sides of the argument but I haven't worked out exactly where I fall on that just yet.
 
For the life of me, I cannot get anyone but Ryo to appear when I prompt something about Shenmue. Half the time it just has Ryo hanging out with a clone (Roy of S3 fame?).

1696268848817.png

Edit: Needed to share these
1696275906316.png
1696275863383.png
 
Last edited:
I had a long conversation with a couple dialogue editors about this topic the other day. Before then I would’ve said yes, but after our chat I’m very skeptical.

In short, their take was that AI dialogue is not even close and could even increase costs due to the sheer amount of curation and input required to make it work properly on the dev side.

But who knows? I’m all for it if YS Net could figure it out :)
 
I wonder if YSNet could make 80% of the generic NPC's have AI-generated dialogue, giving them more time to focus on more important NPC characters.

Think of this in the context of Shenmue 2, where you had more interesting NPC's like the darts players, Guixhang, Joy, Izumi, etc and the more generic npc's that merely existed to make the world feel more populated.

The generic NPC's that simply exist to tell you things like 'I'm sorry, I'm not familiar with this area', could be outsourced to AI. This could be beneficial, because more time could be allocated to 'key' NPC's.
 
OnryoSiva said:
I had a long conversation with a couple dialogue editors about this topic the other day. Before then I would’ve said yes, but after our chat I’m very skeptical.

In short, their take was that AI dialogue is not even close and could even increase costs due to the sheer amount of curation and input required to make it work properly on the dev side.

But who knows? I’m all for it if YS Net could figure it out :)
Click to expand...
That's not true, it doesn't cost anything compared to real voice acting (which is very expensive), but it's true that the result isn't that good compared to real voice acting, especially when you need emotion.
 
The level of voice acting has gone up a lot over the years. What was acceptable in the 90s will no longer be.

My main concern with the use of AI is precisely the voice acting. But I am not very aware of the latest developments in this field anyway. If they don't have money for voices, let them hire only the most famous actors for the main roles and let the AI do the NPCs. We cannot have claims when there is no money. :)

The very fact that they are all voiced is a feat. :)

One of the craziest things about Shenmue 3 is that the lead actor from Final Fantasy XV (Prince Noctis) voices just a normal fisherman on the docks of Niaowo. Waste of talent! :ROFLMAO:
 
Last edited:
I'm all for anything that makes the creation of S4 more efficient for YS Net, reducing costs and increasing the chances of the game being made. AI could be used in some way, but to be honest, the latest UE5 features seem more likely to save time and effort over generic AI tools, as they're already tailored to generating landscapes, character models, etc.

AI voices sound like trash at the mo, maybe on par with the average Shenmue IIx NPC, as mentioned 🤣. Unacceptable in 2023. I'd take duplicate voices e.g. every Elder Scrolls game, over laughable AI voices ✅️
 
I voted for the third option I have no strong feelings one way or the other.

I do not have enough knowledge about Ai AI-generated games to have an opinion.

In general, any additional method to help Y-net make Shenmue 4 a reality is a good thing. But as long as I am not sure this method could go badly wrong I would not consider myself a big proponent of the idea.
 
thedemon1238 said:
One of the craziest things about Shenmue 3 is that the lead actor from Final Fantasy XV (Prince Noctis) voices just a normal fisherman on the docks of Niaowo. Waste of talent! :ROFLMAO:
Click to expand...
I'm not familiar with this actor, but this is surprisingly common. There's often a heck of a lot more consistent work for people playing smaller roles - barks, systemic dialogue, vendors, etc. Actors who break out of the smaller roles often find that business dries up, while those who specialize in systemic dialogue are constantly employed -- kinda the opposite of what you'd expect.
 
OnryoSiva said:
I'm not familiar with this actor, but this is surprisingly common. There's often a heck of a lot more consistent work for people playing smaller roles - barks, systemic dialogue, vendors, etc. Actors who break out of the smaller roles often find that business dries up, while those who specialize in systemic dialogue are constantly employed -- kinda the opposite of what you'd expect.
Click to expand...
There's also the reason of voice acting laws which has a limit of how many main roles every actor can take per year iirc, to allow enough work for everyone. And for the player its good not having lots of different characters sounding almost the same. Voice has a limit for interpretation and when someone gets too many major roles they get lazy and practically don't make effort changing the tone etc.
 
As long as they can get it to work properly I'm here for it.

I feel like this will be one of those things that sounds better on paper though.

"Hey let's save on resources by getting these assets/voices generated". In reality it's probably gonna be some complex web of logic that takes more time than it would've just by outsourcing the work.
 
i made an shenmue 4 cover using AI and some editing

OIG (1).png

And these are the other result i got from the ai


_41562934-0bd0-4de1-b62f-42a9a8a46988.jpg_e4303caa-6c52-48e8-a8ea-f2082c6bb65e.jpg
 
Last edited:
You must log in or register to reply here.
Back
Top