A character turns away before answering. The shoulders remain tense. The reply is quiet, but it is not relaxed.
Now imagine replacing the original dialogue with a fluent line in another language. The words may be correct. The timing may fit. Yet if the new voice sounds comfortably at ease, something about the person on screen has changed.
The dubbing actor is not starting with an empty character. Another actor has already established a body, an attention, a rhythm of response, and a relation to the scene. The new performance has to enter that arrangement.
In the first note, I called acting an inverse process: working back from a scripted line toward the conditions under which a character could speak it, then returning to the words as an action. Dubbing makes that work more complicated. Part of the answer is already visible in someone else's body.
Preparing conditions, not prescribing a sound
Reading Darryl Hickman's The Unconscious Actor: Out of Control, in Full Command gave me a starting point for thinking about this problem. I read it through a tension between deliberate preparation and responsive performance: an actor makes choices, but the response should not appear to be merely the execution of a preselected effect.
The application to dubbing here is my own. Rather than begin by choosing an emotional sound, I want to begin with the conditions that could produce it.
What is the character trying to do? What resistance do they encounter? What are they preventing themselves from saying? What has the other person just made possible—or impossible?
Those questions do not remove technique. They give technique a purpose. A change in breath, pressure, or pace becomes part of an action, rather than a decoration applied to the words.
The actor needs enough control to work within a recording session, but enough responsiveness for the line to belong to what is happening in the scene. This is a tension to manage, not a choice between planning everything and leaving everything to impulse.
Two bodies, one perceived act
There are at least two bodies involved in live-action dubbing: the body visible on screen and the body producing the new voice.
The viewer does not normally see the second body. Nevertheless, its breathing, posture, effort, and orientation can participate in the recorded performance. The voice is not disembodied at its source. It is separated from the body the viewer sees.
The work is to make that separation hold together perceptually. The audience should be able to hear the translated line as something the visible character is doing.
I use re-embodiment for this process: interpreting a performance already established through one actor's body and making it work through another speaking body, in another language.
This does not mean reproducing every muscle movement. Nor does it mean literally transferring the original actor's bodily state. It means reconstructing the conditions of speaking closely enough for the new voice to connect with the existing performance.
The mouth is one constraint, but not the only one. Gaze, gesture, posture, silence, distance, and the timing of another person's reaction are also part of the event. A dub can match the mouth while contradicting the body.
Five layers of the work
I find it useful to separate five layers. This is a working model for these notes, not a standardized account of every studio's practice. The layers overlap, and a decision in one can require revisiting another.
Reading the visible performance
First, look at what has already happened on screen: eyes, hands, posture, breathing, the opening and closing of the mouth, changes in distance, and moments when nothing is said.
Camera distance and editing also matter. An action that reads one way in a close-up may have a different weight in a wider shot. The performance is not a collection of facial expressions detached from the film around it.
Reading an action, not just an emotion
Second, ask what the character is doing through the line.
“Sad” is not yet an action. Trying to keep someone from leaving is. Suppressing a reply because rejection is expected is another. These formulations connect the voice to an object and a possible consequence.
They are interpretive proposals. A look does not give direct access to the character's inner state. The reading has to remain coherent with the rest of the scene and the work.
Handling the constraints
Third, consider what the translated version permits: word order, the amount of information in a line, the available time, visible mouth movements, recording conditions, and the exchange with other performers.
The question is not simply whether the translated words fit. It is whether they fit without changing the function of the line. An adaptation that gains perfect timing by making a cautious character sound decisive has solved one problem by creating another.
Working through the speaking body
Fourth, let the reading of the scene inform the dubbing actor's own conditions of speaking.
An on-screen effort to hold back anger need not become a louder voice. It might require pressure held in the breath, a reduced range, or words released more carefully. An averted gaze might be reflected in the way a phrase withdraws rather than in a generic tone of shyness.
These are possibilities to test, not a dictionary in which each gesture has one prescribed vocal equivalent. The appropriate choice depends on the character and the scene.
Listening to the joined event
Finally, return the voice to the image and listen.
Does the person on screen seem to be saying the line? Does the speech preserve the effort, restraint, direction, and relation that the scene establishes? Does the other person's response still make sense?
A technically clean recording is not the endpoint of this evaluation. The object being evaluated is the audiovisual event.
The speaking body inside the booth
The recording booth brings a practical difficulty. Movement may have to be limited to maintain microphone position, avoid unwanted noise, and fit the working setup. The conditions that make a recording usable can also encourage the performer to concentrate on the script and the sound alone.
In my Japanese draft, I described this as a “cage of signs”: a situation in which the path from written symbols to vocal effects can become too direct.
The phrase is a metaphor, not a judgment that microphone work necessarily removes embodiment. A constrained body is still a body. The practical question is how to keep it involved without undermining the recording.
Breathing, weight, orientation, and imagined distance can help the performer connect the line to a situation. The actor need not physically enact every visible movement. What matters is whether the conditions of the new performance support the action already present on screen.
The restriction does not make the work less like acting. It makes the problem of translating action into voice more concentrated.
When a useful code travels too far
Voice conventions are not inherently empty. A recognizable sound can efficiently establish age, authority, energy, or genre. In animation, the voice may participate in constructing a character alongside drawings and movement. In live-action dubbing, much of the character's performance is already present.
Neither medium has only one method. The risk I want to isolate is more specific: a vocal code developed for one context may become a default applied before the performance in another context has been read.
A performer may deliver a convincing soldier, villain, or flirt while failing to deliver this soldier, this antagonist, or this person's way of approaching someone else.
The result can be skillful at the level of the convention and mistaken at the level of the character. That is why increasing vocal expressiveness does not automatically improve a dub. The direction of the expression matters.
A question for training
My concern with Japanese voice-performance training is not that specialist skills should disappear. Timing, microphone technique, clear delivery, and the ability to respond quickly are real production requirements.
The question is how those skills are related to the construction and reading of a character.
One possible sequence begins with a character's actions, body, relations, and world, then adapts the performance to a voice medium. Another begins with recognizable vocal features and builds the character through them. In practice, both may be present.
These are analytical sequences, not complete portraits of Japan and the English-speaking world. A claim about national differences in training would require specific institutions, curricula, and periods of comparison. What I want to retain here is the structural question: when a vocal convention is taught, are the conditions it condenses taught as well?
If not, a performer can inherit the answer without inheriting the question. The familiar sound becomes a substitute for understanding why this person speaks that way.
Toward representation
The body on screen is also socially situated. Its posture, pronunciation, dress, distance from others, and ways of responding can be read in relation to age, class, region, gender, ethnicity, occupation, and history.
Connecting a new voice to that body therefore involves more than matching movement. It also participates in how the character becomes understandable as a person in a social world.
Re-embodiment leads to representation because the body being read was never neutral to begin with.
The next note asks what happens when the new voice preserves, removes, or adds those distinctions—and why reading them belongs inside the craft, rather than after it.
Method note: “re-embodiment,” the five-layer model, and the applications to training are the author's proposals. They do not describe a surveyed industry-wide standard. The reading of Hickman is the author's interpretation; passage-level references remain in progress. This note does not establish that particular performers were trained by, or derive their methods from, him.
Previous: From Script to Utterance · Next: Representation Is Part of the Craft