What blows my mind is that when it comes to costs I feel like voice actors are probably less than 5% of the budget on a video game. Unless they hire a famous actor I can’t imagine this being that worthwhile. It’s just penny pinching.
Imagine the pace you can just dump out new voice lines for items, maps, general quibble etc that you’d never get the budget to bring a bunch of VAs into studio to do for updates
I get they want to keep their talents and jobs. But it’s just not viable for the future and it has nothing to do with cost.
The future is RPG games where the NPC’s can generate their responses in real time and not in text way, but fully voiced.
There is no pre-loading responses. The future is curated content responding to the player.
So to achieve that, its either a fully generated voice like a vocaloid or you train an AI on someone’s voice.
If the voice actors aren’t interested in it being their voice, they’ll find someone or go vocaloid.
It’s not about saving money. It’s about pre-recorded voice lines being dead on arrival.
Think audio books, but choose your own adventure audio books, where all the names/places/things can be curated to the listener. Voice Actor isn’t going to be apart of that.
That’s likely true, but we can write a fair contract that allows for that.
A pay rate for the time spent recording, slightly lower than the normal rate
A pay rate per minute of generated vocal content using user telemetry. For users that opt out of sharing, pay out based on averages applied against their total playtime.
I don’t know about playtime based payment, but maybe % of sales for paid games and paywall generative content in free2play games (with a significant % of purchases for “voice unlocks” going to VAs)
That could be the future, but not anytime soon. I haven’t seen anything AI gen that has enough continuity to make “on the fly” story telling something I’d be interested in.
It’ll be marketed as Skyrim with all LLM text and end up as Oblivion with prefab text chunks.
Even disregarding the fact that current LLMs can’t stop hallucinating and going off track (which seems to be an inherent property of the approach), they need crazy accounts of memory. If you don’t want the game to use a tiny model with a bad quantization, you can probably expect to spend at least 20 gigs of VRAM and a fair chunk of the GPU’s power on just the LLM.
What we might see is a game that uses a small neural net to match freeform player input to a dialogue tree. But that’s nothing like full LLM-driven dialogue.
Obviously you get downvoted, but just as obviously this is the future. Nonsense static responses are useless, having actual responses that match what happens would take immersion to a whole new level.
Oh yeah, that’s why nobody reads regular fiction anymore, it’s all choose-your-adventure these days.
Bots will never be able to act as well as humans do until they are literally sentient. They might sometimes sound like someone naturally speaking, but they will never be able to intonate the correct emotion the way we do. They will never be able to make creative decisions and work with a director. And they will never be able to hold a real conversation. ChatGPT and Deepseek can hold their own longer than Cleverbot did, but they still devolve into nonsense in short order. Zero chance they’ll be able to hold together consistent fiction the way a human writing a script does.
Mate, LLMs are really good for creative stuff. And they improved SO much over the last 2 years. How could you even think to say “they will never”? You can already have conversations with AI, you could a year ago. Now the context window is MASSIVE, you are going to talk for a long time before that runs out at 200k tokens. Let alone methods to condense this down to only relevant information, which would essentially give it infinite memory about what you talk about, easily in the millions of words.
Again, Cleverbot. You could have “conversations” with a computer as far back as 2008. Generative models are better, but still terrible at giving accurate information or even at staying on topic without constant nudging in the right direction. Actual acting is FAR out of their range.
Take the robotic, stilted automatic animations that have been used for decades in games like Skyrim. It’s fine if you don’t care at all about looking natural, and AI can make that trashy bulk animation a lot better, but it will never replace the quirks and characterization provided by hand animated cutscenes like you can find in Uncharted or Red Dead, let alone stylized games like Pikmin and Mario.
What blows my mind is that when it comes to costs I feel like voice actors are probably less than 5% of the budget on a video game. Unless they hire a famous actor I can’t imagine this being that worthwhile. It’s just penny pinching.
Imagine the pace you can just dump out new voice lines for items, maps, general quibble etc that you’d never get the budget to bring a bunch of VAs into studio to do for updates
Honestly it’s probably an agility thing. You remove the entire. Getting another human to do the work thing
If you believe it hasn’t occurred to them that they won’t have to pay wages any more, I have a bridge to sell you.
Oh I agree with you. I just don’t think they are doing this just because of that.
I get they want to keep their talents and jobs. But it’s just not viable for the future and it has nothing to do with cost.
The future is RPG games where the NPC’s can generate their responses in real time and not in text way, but fully voiced. There is no pre-loading responses. The future is curated content responding to the player.
So to achieve that, its either a fully generated voice like a vocaloid or you train an AI on someone’s voice.
If the voice actors aren’t interested in it being their voice, they’ll find someone or go vocaloid.
It’s not about saving money. It’s about pre-recorded voice lines being dead on arrival.
Think audio books, but choose your own adventure audio books, where all the names/places/things can be curated to the listener. Voice Actor isn’t going to be apart of that.
That’s likely true, but we can write a fair contract that allows for that.
The problemo see with this construct is that it only benefits current actors. There won’t be space for a new generation of actors.
I don’t know about playtime based payment, but maybe % of sales for paid games and paywall generative content in free2play games (with a significant % of purchases for “voice unlocks” going to VAs)
That could be the future, but not anytime soon. I haven’t seen anything AI gen that has enough continuity to make “on the fly” story telling something I’d be interested in.
I challenge you on that.
We’ll see a Skyrim like game using LLM for NPC’s within 3 years, definitely 5.
It’ll be marketed as Skyrim with all LLM text and end up as Oblivion with prefab text chunks.
Even disregarding the fact that current LLMs can’t stop hallucinating and going off track (which seems to be an inherent property of the approach), they need crazy accounts of memory. If you don’t want the game to use a tiny model with a bad quantization, you can probably expect to spend at least 20 gigs of VRAM and a fair chunk of the GPU’s power on just the LLM.
What we might see is a game that uses a small neural net to match freeform player input to a dialogue tree. But that’s nothing like full LLM-driven dialogue.
I think some will exist in that time frame, but I don’t think they’ll be any good, or well received.
IN the near-future of gaming, but not BEING the near-future of gaming.
There’s already mods doing it now that are absolutely doing it better than Bethesda will.
Obviously you get downvoted, but just as obviously this is the future. Nonsense static responses are useless, having actual responses that match what happens would take immersion to a whole new level.
Oh yeah, that’s why nobody reads regular fiction anymore, it’s all choose-your-adventure these days.
Bots will never be able to act as well as humans do until they are literally sentient. They might sometimes sound like someone naturally speaking, but they will never be able to intonate the correct emotion the way we do. They will never be able to make creative decisions and work with a director. And they will never be able to hold a real conversation. ChatGPT and Deepseek can hold their own longer than Cleverbot did, but they still devolve into nonsense in short order. Zero chance they’ll be able to hold together consistent fiction the way a human writing a script does.
Mate, LLMs are really good for creative stuff. And they improved SO much over the last 2 years. How could you even think to say “they will never”? You can already have conversations with AI, you could a year ago. Now the context window is MASSIVE, you are going to talk for a long time before that runs out at 200k tokens. Let alone methods to condense this down to only relevant information, which would essentially give it infinite memory about what you talk about, easily in the millions of words.
Again, Cleverbot. You could have “conversations” with a computer as far back as 2008. Generative models are better, but still terrible at giving accurate information or even at staying on topic without constant nudging in the right direction. Actual acting is FAR out of their range.
Take the robotic, stilted automatic animations that have been used for decades in games like Skyrim. It’s fine if you don’t care at all about looking natural, and AI can make that trashy bulk animation a lot better, but it will never replace the quirks and characterization provided by hand animated cutscenes like you can find in Uncharted or Red Dead, let alone stylized games like Pikmin and Mario.