Dubbing and subtitling are two different arguments about what a performance is

Comparisons of localisation as performance craft usually pick a winner. This one picks the circumstances, which is more useful.

What dubbing has to solve

A dubbed line must fit the duration of the original and match visible lip movements, particularly on closed and open sounds. That constrains the translator before meaning is considered, since a semantically perfect line that does not fit is unusable. Adaptors therefore rewrite rather than translate, finding phrases that carry the sense and land on the right mouth shapes.

The performance is then recorded by an actor watching the picture, matching timing while generating their own reading. Calling it a copy of the original misdescribes the work, which is closer to performing a constrained new script.

What subtitles have to solve

Reading speed limits how much text can be displayed, so subtitles are compressed rather than complete transcriptions. Standard practice caps characters per line and duration on screen, which forces the removal of a substantial share of the words.

Here is what the choice buys: what goes first is repetition, names in address, discourse markers and anything the image already conveys. That means subtitle viewers receive an edited version of the dialogue, which is rarely acknowledged in arguments about fidelity. Good subtitling is an act of judgement about what can be lost, and it is skilled work with very little visible credit.

The attention cost

Reading occupies visual attention, so a subtitled viewer is spending part of every scene not looking at the composition. Experienced subtitle readers become extremely efficient, but the cost does not fall to zero, particularly in dense scenes.

By the middle of the season, dubbing removes that cost entirely and returns the whole frame to the viewer, which is a real benefit rather than a concession. It substitutes a different loss, since the original performer's vocal work is replaced completely. Neither trade is obviously correct, which is why national preferences differ so consistently and so strongly.

Why traditions diverged

Countries developed different norms early, based on literacy rates, market size, and the cost of producing dubbed versions. Large markets could support full dubbing industries, and those industries became established professions with their own stars. Smaller markets subtitled because it was cheaper, and audiences grew up expecting it, which made it the accepted form.

On a second viewing, both patterns are now deeply embedded and are experienced as preferences rather than as historical accidents.

Neither reflects a judgement about the medium, though arguments between the two camps frequently assume otherwise.

Where each fails

Poor dubbing produces flat, evenly recorded voices with no acoustic relationship to the spaces the characters are standing in. That is a mixing failure rather than a performance one, and well-made dubs place voices in the scene convincingly. Poor subtitling compresses too aggressively, translates idiom literally, or lags the speech, all of which are common and avoidable.

Machine-generated subtitles fail characteristically on names, overlapping speech and anything requiring context. Judging either practice by its worst examples is easy and misleading, since both have excellent practitioners.

Budget and streaming figures are rarely audited and frequently strategic.

Watching with both

Playing a scene dubbed and then subtitled reveals immediately how much dialogue the subtitles removed. It also reveals what the dub rewrote, since jokes and idiom are usually replaced rather than translated.

Structurally, notice whether dubbed voices carry the room's acoustics, which is the clearest indicator of a well-made version. Watch how subtitles handle overlapping dialogue, which is the hardest case and where quality differences are starkest. The comparison makes the craft visible and settles the argument about neutrality quite quickly.

Side by side

The takeaway

Both routes into another language remove something; the useful question is which loss you prefer.

Notice what is kept out of frame. That is usually where the decision was made.