Videos · Compared

Voiceover vs narration: is there a difference?

Voiceover is speech recorded separately and played over the picture by a speaker who is not seen at that moment. Narration is speech that tells or explains the story to the viewer. In most video work the two describe the same track: narration names what the voice does, voiceover names how it reaches the picture.

People use the two words as if they were one, and most of the time nothing goes wrong. The gap shows up in scripts, contracts and accessibility requirements, where asking for the wrong one gets you a different deliverable.

Indunil Asanka · Co-founder

5 min read · Published 13 September 2026

What the words usually mean, field by field
Field	Usually called	What it means there
Film and television drama	Voice-over, written V.O.	Speech from a voice not physically present in the scene, often a character narrating or thinking
Screenwriting	O.S. or O.C.	A character who is in the scene but out of frame, which is not a voice-over
Documentary	Narration	A narrator guiding the story, heard over footage or sometimes seen on camera
Advertising	Voiceover or VO	The announcer's read over the picture, often by a hired voice artist
E-learning and tutorials	Narration	Instructional speech that explains each step as it appears
Accessibility	Audio description	Narration of important visual detail for blind and low vision viewers
Corporate and explainer video	Either word	The offscreen speech that carries the script

Two names for one track

In a typical marketing or training video there is one voice, it is not on screen, and it carries the script from the first scene to the last. Call that track voiceover or narration and everyone on the project will understand. The Columbia Film Language Glossary captures the overlap in one line, defining voice-over as dialogue, usually narration, that comes from an unseen, offscreen voice, character or narrator. It also classes voice-over as non diegetic sound: the audience hears it, but the people inside the story do not. That shared ground is why the two words are swapped so freely.

Where the words part company

The difference is one of angle. Voiceover describes a production technique: audio laid over pictures, recorded apart from them, from a voice that is not in shot. Narration describes a storytelling role: somebody telling the viewer what is happening, what it means or what to do next. So narration is not always voiceover, because a documentary presenter can narrate while standing in frame. And voiceover is not always narration, because a character's inner thoughts or a radio heard over a scene can be a voice-over without anyone telling a story. Scripts mark the distinction carefully, since V.O. and a character speaking out of frame are directed, recorded and mixed differently.

Audio description is narration with one job

Accessibility adds a third meaning that is easy to confuse with the first two. Audio description, as the W3C explains it, is narration that describes important visual information a blind or low vision viewer would otherwise miss, such as on screen text, actions and scene changes, fitted into natural pauses in the audio. A marketing voiceover does not do this by default. Where accessibility rules apply, a video may need both: the voice that sells or teaches, and a described version, or a script written so that the main narration already speaks every piece of visual information that matters.

When the difference matters

It matters in briefs and quotes. Ask a studio for narration on a documentary and they may plan a presenter; ask for voiceover and they will plan an offscreen recording. It matters in contracts, because voice artists are usually engaged and licensed per use, so the word used for an advert read can carry specific usage terms. It matters in scripts, where writers mark V.O. so the editor knows the line plays over other pictures. It matters in translation, since an offscreen narration can be recorded again in another language, while a presenter on camera needs subtitles or a reshoot. And it matters in accessibility reviews, where a reviewer asking whether the video is narrated may really be asking whether visual information is spoken aloud.

Common mistakes

Writing narration that describes exactly what the picture already shows wastes the viewer's attention, so let the voice add what the image cannot. Packing too many words into a short scene forces a rushed read and captions that fly past. Recording before the edit is locked means lines no longer fit their scenes. Burying the voice under music makes the words work harder than they should. Choosing a voice before the tone of the script is settled is another, because a bright read over a sensitive subject undercuts every line. And treating a marketing voiceover as audio description misses the point of description entirely, since it rarely mentions the on screen text and actions a blind viewer needs to hear.

Where it shows up in the product

In the video builder here, the spoken track is written per scene as narration text and synthesised as voiceover, so both words describe the same thing. Voices are Google Chirp3-HD, chosen in four styles, energetic, instructional, warm and neutral, each in two genders, and a tutorial is never given the energetic voice. In each scene's voice dialog you set the narration text, language, voice, a start delay, speed from 0.85x to 1.25x and whether captions show. Captions are generated from that same text, and background music ducks under it. The Lisbon listing example uses a warm Portuguese voice across every scene.

Questions people ask

What does V.O. mean in a script?

V.O. marks a line spoken by a voice that is not physically present in the scene, such as a narrator or a character's thoughts, so the editor knows it plays over other pictures. It differs from a character who is present but out of frame, which scripts mark separately. The label tells the director, the sound team and the editor how that line should be recorded and placed.

Is a presenter talking to camera doing narration?

Yes, if they are guiding the viewer through the story or the subject, but it is not voiceover while they are in shot. Many documentaries and explainers mix the two: the presenter narrates on camera to introduce a section, then the same voice continues over footage as voiceover. For planning, count those as two recording setups, one filmed and one recorded separately.

How many words should narration have per scene?

Fewer than most first drafts. A promotional scene of two and a half to four seconds holds about five to eight spoken words, which is the budget the generator here writes to. An instructional step of five to nine seconds can carry ten to twenty five words, because the viewer is following along. If a line needs more, the scene needs more time or the thought needs splitting.

Should narration repeat the text on screen?

Only for the key message. Reading every on screen word aloud makes the video feel slow, and viewers read faster than a voice speaks. The stronger pattern is short on screen text for the headline or number, with narration adding context around it. The exception is the offer, price or call to action, where hearing and seeing the same words helps them stick.

Does a hired voice artist's recording need a licence?

Usually the recording is supplied under agreed usage terms, which can limit the media, the territories, the period and whether the audio can be reused in other videos. Paid advertising often needs broader rights than internal training. Check the agreement before cutting the same read into a new advert, since reuse beyond the original terms can mean an additional fee.

Make one with videos

The button opens the generator with this use case already described. Change the wording to match your own.

Create a video with OneCraft

Sources

Written and checked by the OneCraft team. Last checked 13 September 2026.