Lost in Translation: Why Literally Translated Scripts Break Voice Actors
- TooSix Media Group
- Apr 7
- 7 min read
When Squid Game became Netflix's biggest series launch in history - hitting number one in 94 countries simultaneously - it didn't just change what the world was watching. It accidentally started one of the most important conversations the voice acting and localization industries have ever had.
In September 2021, Korean-American comedian and bilingual speaker Youngmi Mayer posted a video pointing out problems in the English subtitles and dub. The clip racked up over ten million views. The examples she gave weren't small. Character motivations shifted. Class commentary vanished. Emotional beats flatlined. The writers of Squid Game had built a story loaded with layered meaning - and in English, much of it simply didn't survive the crossing.
Here at TooSix Media Group, working out of Seoul with Korean and English voice talent, we've seen versions of this problem from the inside of the booth. Not in ways that make headlines, but in the quieter, more persistent way that shapes every session where a script moves between languages: the performer reads the translated line, something feels wrong, and when you trace it back, the problem isn't the acting. It's the script.
The Fundamental Misunderstanding: Translation Is Not the Same as Localization
Most people who commission Korean-to-English voice work assume translation and localization are the same thing. They're not. Translation converts words. Localization converts meaning - and when it comes to voice performance, it also converts rhythm, breath, and emotional timing.
A translated script is a starting point. A localized script is a performance-ready document.
The difference matters enormously for the voice actor. When a performer receives a script, they aren't just reading words - they're looking for where the emotional weight of each line lands, how much breath a sentence requires, where their voice rises or falls, and what mood the end of the sentence is leaving the listener with. Mess with any of those invisible structures, and even the best actor in the world will struggle to make it sound natural.
Korean and English create these problems structurally, by design.
The Grammar Problem: Korean Is a Language That Builds Toward Emotion
English is a Subject-Verb-Object language. The action comes early: "She loved him." You know the verb immediately. The emotional core arrives upfront and the sentence runs out from there.
Korean is Subject-Object-Verb. The verb - and therefore the emotional punch - comes at the end: "She him loved." The sentence builds. You don't know where it's going until it arrives.
This isn't a minor grammatical footnote. It completely changes how a voice actor performs.
In Korean, an actor knows to build through a line, because the meaning is still being constructed. The breath stays open, the energy accumulates, and the delivery resolves at the end of the sentence. In English, the actor commits to the emotional color early, because the verb has already told you what's happening.
When a Korean script is literally translated into English without restructuring for this difference, what lands in the booth is a sentence whose rhythm is fighting itself. The actor ends a line where the energy should be building, or peaks too early and has nowhere to go. It sounds flat. Not because of the performance - because the architecture of the sentence is pulling in the wrong direction.
We've seen this slow down sessions noticeably. Talented performers going take after take, knowing something is off, unable to articulate it, because the problem is upstream in the script room, not in the booth.
The Honorifics Problem: Social Registers That Simply Don't Exist
Korean has an elaborate system of speech levels - formal, polite, informal, intimate - that shift depending on age, relationship, and social hierarchy. These aren't optional stylistic choices. They're grammatically built into every sentence.
When a younger character speaks to an elder in Korean, the level of formality is encoded in the verbs themselves. When two close friends speak to each other in banmal (informal speech), the intimacy is in the grammar. The relationship between characters is constantly being expressed through language structure, not just through what is said.
English has no equivalent system.
The standard localization shortcut is to replace honorifics with first names, or to use terms like "sir," "old man," "babe," or "mister" as rough approximations. The Vice analysis of Squid Game found exactly this: the Korean term oppa - used by women for older male figures they're close to - was rendered as "babe" in one scene and "old man" in another, with "mister" appearing elsewhere. The term hyung, used by men to address older close males, was simply replaced by first names throughout the English dub.
What's lost isn't just politeness markers. What's lost is relationship depth. The moment you replace a word like hyung with a first name, you've erased the signal that tells the audience: these two people are close, and this closeness has history. The voice actor playing the scene in English has to try to perform that intimacy without the linguistic scaffold that carries it.
The Cultural Meaning Problem: Words That Don't Have English Counterparts
Some of the hardest translation moments aren't about grammar at all. They're about concepts that Korean has words for, and English simply doesn't.
The clearest example from Squid Game - and one Youngmi Mayer specifically highlighted - is the scene involving the word gganbu, a term for a close childhood friend with whom you share everything, with no concept of individual ownership between you. The English dub rendered this as "we share everything." The more literal translation would be something closer to "there is no concept of mine and yours between us."
Those two things feel similar on paper. In performance, they are completely different lines. One is a statement of shared practice. The other is a declaration of a specific kind of bond - almost a philosophy of relationship - that the character is invoking at a pivotal moment.
The voice actor reading "we share everything" is performing one scene. The actor who understands what gganbu actually means is performing a different, much richer one.
Then there are episode titles. The first episode of Squid Game, known in English as "Red Light, Green Light," is titled in Korean after the mugunghwa flower - South Korea's national flower - and the children's game named for it. The English title is a functional replacement. It tells you what game is being played. The Korean title tells you something about identity, nationhood, and childhood all at once. A narration line written around the Korean title and its meaning cannot be translated literally without losing everything that makes it meaningful.
What This Looks Like in the Recording Booth
When a Korean script arrives in our studio that hasn't been properly localized - only translated - the signs are recognizable within the first few pages.
The sentences end in the wrong place for an English performer. The emotional payoffs arrive too late or too early. Social relationships that should be obvious from the dialogue feel undefined. Characters who are supposed to sound warm come across as neutral. Characters who should be sympathetic register as cold.
The most common mistake we see in translated scripts is what might be called the summary problem: the translator understood what the line meant, and rendered its functional meaning accurately, but stripped out the way that meaning is carried. The line "I am very smart, I just never got a chance to study" - the correct translation of Han Mi-nyeo's line in Squid Game - carries a specific class commentary, a character's pride, and her bitterness, all in one breath. The translated version, "I'm not a genius, but I still got it worked out," says something similar but performs nothing like it. A voice actor reading that line can't perform the original intent, because the original intent is no longer on the page.
Localization repairs this. It asks not just what does this mean but how does this meaning need to arrive in English so that a performer can deliver it and an audience can feel it.
For Brands and Producers: Why This Matters Beyond Entertainment
This conversation might seem like it only applies to streaming drama, but the same problem appears in corporate localization, e-learning content, advertising, and branded video.
Korean organizations producing content for English-speaking markets, and English-speaking brands entering Korean ones, are both vulnerable to the same gap. A translated script that retains the SOV rhythm of Korean, or that removes the implied social hierarchy of a professional conversation, will produce voice performances that feel slightly wrong to native English speakers - even if they can't explain why.
The budget saved on proper localization is often spent twice over in reshoots, re-recordings, and additional direction time trying to coax a natural performance out of a script that was never built for one.
The Standard to Aim For
The benchmark for Korean-to-English localization isn't accuracy. It's performability. A localized script should arrive in the booth ready to be recorded - with the right emotional architecture, the right breath points, the right relationship signals, and the cultural meaning intact even if the literal words have changed significantly.
Director Bong Joon-ho described English subtitles as a "one-inch barrier" that, once crossed, opens audiences to entirely new worlds of cinema. That's true. But for voice actors and producers working on the other side of that barrier - building the audio world that audiences enter - the work is more granular than that. It's line by line. Sentence by sentence. Word by word.
Translation gets you across the barrier. Localization makes what's on the other side feel like it was always meant to be there.
At TooSix Media Group, we work with Korean and English voice talent, and we've seen both sides of this challenge firsthand. If you're producing bilingual content and want to talk about what a properly localized script looks like before it reaches the booth, get in touch.
Comments