The Second Script: Editing Brand Video Captions for the Ear and the Eye

A brand video may begin as a piece of speech, but its captions are not simply that speech typed out. They are a second script, written for a different sense and a different pace. The viewer is listening and reading at once, often on a phone, frequently with the sound low or off. The editorial task is therefore not word-for-word accuracy alone. It is to produce captions that can be read comfortably at speed while still sounding like the person who spoke them.

That distinction matters for UK brands. A caption that preserves every false start and filler may be faithful to the recording but poor as reading material. A caption that tidies every sentence into polished corporate prose may read smoothly but erase the speaker’s rhythm, humour and regional character. Good subtitling sits between the two: close enough to the voice to be credible, clear enough to be useful.

Begin with the voice, not the waveform

Before editing a single line, listen to the whole video without pausing. Note the speaker’s habits: typical sentence length, recurring phrases, where the humour falls, whether the tone is dry or warm, and which regional words or pronunciations recur. These notes become the reference point for every later decision. If the speaker says “a bit of a faff” rather than “an inconvenience”, that is not a flaw to be corrected. It is evidence of voice.

The first transcript is merely raw material. The editorial pass begins when you ask what the speaker sounds like on a good day, and how that impression can survive in two-line blocks on screen.

Read at the speed of the viewer

Caption readability is governed by time as much as by space. UK viewers are accustomed to concise subtitles, and online platforms offer little patience for dense blocks of text. A comfortable caption usually occupies no more than two lines, with each line holding a natural phrase rather than an arbitrary character count. Guidelines vary, but the principle is consistent: the eye should be able to take in the line before the next one arrives.

Line breaks should follow meaning. Break after a clause, before a conjunction, or at a natural pause in speech. Avoid leaving a single word stranded on a second line, and avoid breaking between an article and its noun or a preposition and its object. Punctuation should guide the eye, not decorate the screen.

Punctuation is pacing

In captions, punctuation does more than satisfy grammar. It tells the reader where to breathe. A full stop creates a pause; a comma creates a lighter one; a dash can signal an aside or a change of direction. Semicolons and elaborate subordinate clauses rarely earn their place on screen. They slow the reader down and make the caption feel more formal than the speaker does.

Contractions deserve protection. “Do not” may be correct, but if the speaker said “don’t”, the caption should usually say “don’t” too. Expanding contractions can make a lively speaker sound like a policy document. The same applies to sentence fragments. If the voice is conversational, a fragment may be the most faithful representation of it.

Filler: cut the noise, keep the character

Not all filler is equal. False starts, repeated words and meaningless verbal tics can usually go. They add nothing for the reader and consume valuable screen time. But some fillers are doing real work. A deliberate “well” or “look” can signal a shift in argument. A regional idiom can carry warmth or local identity. A hedge such as “sort of” may be central to how the speaker softens a claim.

The test is whether removing the filler changes the impression of the speaker. If cutting it makes the speaker sound more assertive, more corporate or simply less like themselves, it may need to stay. The aim is not a sanitised transcript but a readable representation of a particular voice.

Regional flavour without confusion

UK brands often work with speakers who have distinctive regional speech, and captions should not sand that away. Standard spelling is usually the right choice: phonetic transcription of an accent can look patronising and is difficult to read. But word choice is different. If a speaker says “nowt”, “wee”, “bairn” or “canny”, the caption should generally keep the word rather than replace it with a bland equivalent.

Where a term might genuinely obstruct comprehension for a national audience, the answer is not to flatten the voice. It is to consider the context, the intended audience and whether the video itself supplies enough support. Captions serve accessibility, but accessibility is not the same as homogeneity.

When captions should diverge from the audio

It is normal for captions to condense. Redundancies, off-topic asides and repeated phrases can be removed without loss of meaning. What should not happen silently is correction of factual slips. If a speaker gives the wrong date or misnames a product, the editor should flag it to the client rather than quietly fixing it in the caption. Captions represent speech, and a brand’s credibility depends on consistency between what is heard and what is read.

A short editorial workflow

A reliable sequence is to transcribe with timestamps, read the transcript for sense, mark the voice traits that must survive, edit into caption blocks, then read the captions aloud at the speed of the video. The final check is against the brand’s voice, not only against the audio. Does this read like the speaker on form? Would a regular viewer recognise the person behind the words?

A short style sheet helps: how contractions are handled, which regional terms are retained, how laughter or music is indicated, and whether speaker labels are needed. Consistency across a series is part of voice.

Captions are not a transcript. They are an editorial second script, written to be read in motion. The standard to aim for is simple: clear at speed, faithful in voice, and never so polished that the speaker disappears.