I saw some folks thinking that voice acting in games was objectively good, always a positive. “What kind of game doesn’t have voice acting in 2025?!” sorta thing
I say that there are positive reasons why one would choose NOT to have VO.
How many design justifications we can come up with?
Join me.
#1: The most obvious: $
Budgets always have trade-offs. A simple scenario might be “if we hire voice actors we don’t have money for loc” or “we’ll have to cut half the music tracks”.
In this scenario they WANT voice acting and can’t justify cost, so not as interesting for our purposes.
#2: A big one is production timelines. In order to ship with VO in multiple languages, you have to:
- Lock down the script
- Translate the script
- Record in every language
Those steps have to happen in order. This work cannot be parallelized. This greatly reduces development freedom
“But Max, isn’t that just cost again?”
I’m glad you asked! Nope!
A lot of game development is squinting at an unfinished game, with missing features and polish, for months or years, trying to extrapolate what it’ll all be like when it comes together.
Have you ever thought “how did they think this was a good idea?” or “How did they not see that this sucks?” while playing a game?
The answer is often: they thought it was a good idea years ago, and it took them years to build it and then see whether it actually was or not.
One of the most valuable things a game production can do is arrange deadlines so that there is time to do polish but also finishing adjustments and fun additions AFTER all the features are stood up.
Multiple iterations. Stuff like doing an “add secrets” pass on every level after they’re finished.
People make better content when 1) They can see the real context it will exist in, and 2) They have the flexibility to have an idea and quickly implement it.
This is why the Dreadnaught in Destiny: The Taken King was good. All the secrets and hidden stories? Added after the campaign was done.
These kind of passes are hard to plan for because you can’t enumerate the content ahead of time. “Add secrets” might be the wrong choice! You need the team to be able to play the game and think “it would be cool if this was here”, and then DO IT while the inspiration is hot. Sans red tape.
Many games in the past simply achieved this by crunching at the end. Which sucks! It’s not a very GOOD way to do it. But this was, say, the Zelda team’s strategy for decades.
AAA games now simply DON’T do this because there are too many interdependencies. You can’t just have an idea and implement.
Sooo long story short: the more freedom you can preserve for the closing stages of production, the better your game will be. Full stop. This is a huge factor in whether games have that human spark in them at the end.
Localized VO? It almost always has to get locked down earlier than other content.
#3: Players can control the pace when they read. They can’t control the pace of voice acting.
#4: Players can imagine different voices and line deliveries for characters if they are not voiced.
I know some people roll their eyes at that one, but, like, it’s a real thing! That kind of personalized experience could be part of someone’s vision. It certainly creates different flavors of fandom.
#5: Sensory overload. You might simply want players to be paying attention to something else in the game, and a LOT of people struggle to process both language and other stuff at the same time.
This is probably more a reason to avoid VO at certain times than altogether.
#6: Have you seen the streamers that play old games and do dramatic readings of the text on screen to their followers? Fun shit. Maybe you want players to do that!
I’ve also known people who did this with friends/siblings to turn SP games into a communal activity.
#7: Sometimes you want people to pay personal attention to prose, with their eyeballs. Reading poetry and hearing poetry spoken aloud are very different things! And some of the people writing our game dialog are fucking POETS.
#8: There are situations that feel natural in just text boxes but feel stilted or uncanny in voice. For instance: presenting a dialog choice. It never feels weird in text boxes, but a lot of games kinda struggle with how to handle that for voiced conversations.
Or just saying really weird shit.
#9: There are things you can represent with text but not in spoken dialog. You can play with line spacing, create the “gap” that comic writers talk about. You can represent untranslatable dialog. You can represent text-based communication. Probably tons of stuff I’m not thinking about.
![A poem by e. e. cummings, called "[l(a]"](http://namelessquality.com/wp-content/uploads/2026/01/bafkreiecn7ddyqehcgsrbh2lxgp6yutvtd56rr7mmjfgoznmpjs5bm3s6a.jpg)
#10: Bad voice acting is arguably worse than no voice acting. Writing for spoken delivery, and directing voice actors, are not skills everyone has or cares to build.
#11: It’s MUCH easier to support player-entered or otherwise variable neopronouns in just text.
I DO think that even voiced games should all be supporting AT LEAST they/them for player-defined characters.
#12: Player defined names. Those are only really possible w/o voice, and it can be very awkward when games try to let you do it with voiced dialog.
A final note from me: Voice acting IS super valuable for accessibility. Low-vision players may need it. It’s not always an option and there may be SOME artistic visions where it would so contradict the goals that it doesn’t work. But DO consider how to make your game work for this audience.
From Zoe Quinn (@unburntwitch.com):
The most obvious one for me: maybe you want to put so many words in the game with so many small dynamic permutations within the same lines that voice acting all of it would be an objectively stupid choice. Like if I want to swap out 3-4 words every other line to have dialogue be responsive to many different player choices, do you record every permutation of the entire line to get natural sounding reads? Do you record weird inserts of the dynamic portions of the sentences? It gets really whack really quick. And impossible if you want user-entered language versus bespoke words.
Another design reason: maybe the bulk of your game isn’t the kind of dialogue you would hear diagetically in the first place. Maybe communicating via text (like Mystic Messenger) is part of the design. Maybe the dialogue of the story is told through found emails or diaries.
From Mr. Lucha (@luchalutra.bsky.social):
Ok, after looking over everything you have, the only thing I can think of you didn’t cover (or I didn’t see) was that a voiced experience is a directed experience. You are told how to feel by the reception and perspective of what the dialogue is doing. It’s manipulative in that regard. for example:
Because a character is voiced a certain way, (tying into what you said about player agency), you are effectively enforcing the author into the scene whether deliberate or not. It doesn’t allow for the nuance that a players on headcannon can create, and the plausible deniability of silence.
and that silence is so, so important, for many scenes, settings, and circumstances. and you don’t have to look far into what I mean by that, look how a player reads a character’s dialogue, after something hits them, and only them, on a specific level, a specific tier of their own personhood.
If in Signalis, the main character was voiced, you get all her emotion (or lack there of) and her curated experience that you subconsciously inject into the scene b/c of the author behind the scene.
But if a scene where she would be resolute, the player injected shaky uncertainty, that’s special.
You can’t “get” that with a voiced role.
Gray Fox’s last stand in MGSX, strong, strong performance. but it’s being told to you and the player that gray fox is inaccessible in any other reading because the player has no interjection on it. He’s a person, but he’s not the player’s.
So you get ONE strong read of a character, but you lose out on the thousands of permutations and possible shifts that a player’s on voice and interjections can create for them.
I hope all that makes sense, but that’s one of my main arguments on the matter.
Finally, if you want the best example, just watch how people read undertale characters. It’s a tapestry of expression, only possible with the slight guidance of audio cues in the tone of voice, but the player creates the character themselves.
From Cortana (@cortanav.bsky.social):
Even just partial VO from a design standpoint is better.
Imagine if Pathologic had full VO for everything. It would be a slog.
Oh god. Imagine if MORROWIND had full VO. I mean, we’ve seen where that lands us, I guess. And it wouldn’t have been feasible for storage capacity at the time.
From Edgecrusher (edgecrusherntb.bsky.social):
It can be better for portraying aliens as alien, instead of making them feel too human!
I’ve been enjoying that about the Psion’s little dialogue boxes (with no voice acting) in Destiny 2’s most recent expansion.
From Eran Arbel (@sabrerunner.bsky.social):
Procedurally generated text. E.g. Dwarf Fortress/Rimworld/etc…
From Marcus N. (@moewenfels.bsky.social):
In a way those VO “noises” are similar to pixel sprites.
Because in both cases they are very abstract and your own imagination fills the gaps.
And it fills the gaps with what you might like best. Turning characters into different experiences for every player.
No VO might also be more universal IF the player doesnt mind reading.
Every VO will have a certain style and people will react differently to it.
But if you read something, you’ll read it in the way that you prefer. Like a particular style or tone.
From Michael Williams (@rocketmoose.bsky.social):
Do you want to add more content for your game down the road? (DLC, seasonal content, sequels, etc). Adding voice acting means you need to hope those actors will be available again, or have an alternate plan (sound-alikes, re-recording, etc). Are you localizing? Then you’ve got that for every region!
From Bruno Dias (@brunodias.bsky.social):
Obviously many practical reasons but a creative one is that voice acting is one thing that grounds a video game character in a real person (the VA), thus giving them a real voice, dialect, accent, etc. Not something you necessarily want! Sometimes your character is stylized, archetypal, etc, etc.