Voiceover: how to write narration scripts for brand videos
How an agency or freelancer writes a voiceover script that sounds natural for client videos, with rules, narrator selection, and mistakes to avoid.
A voiceover is narration heard in a video without the speaker necessarily appearing on screen. It is used in ads, corporate videos, tutorials, reels, and presentations. It can accompany existing footage, though its suitability depends on the content and the piece’s purpose.
Well written, a voiceover complements the visuals and guides the viewer without distracting. Poorly written, it literally repeats what's already on screen and overloads the viewer with redundant information. This article covers how to write the script, choose a narrator, and avoid the mistakes that force re-recording.
When to use a voiceover
- When what you're showing needs context that isn't visually obvious.
- When you want to connect shots that would otherwise feel disconnected.
- When the person will not speak to camera or the content needs separate narration.
- When you're building pace: a narrator can speed up or slow down an edit as needed.
- When preparing several versions and separate narration helps adapt the message without repeating the entire shoot.
When NOT to use a voiceover
- When the visuals already tell the story. Repeating is noise.
- When a direct statement contributes something that would be lost by replacing it. A real person can also record voiceover; appearing on camera is not the only way to provide testimony.
- When the format calls for intimacy (personal reels, founder videos, community content).
- When the brand competes on authenticity and the voiceover feels artificial.
Rules for writing a voiceover script
1. Write to be heard, not read
Text that reads well usually sounds stiff when narrated. Short sentences, natural contractions, conversational rhythm. Before approving a script, read it out loud. If a sentence is difficult to say, review its construction before recording.
2. Avoid literally repeating what's on screen
If someone is typing on a computer on screen, the voiceover doesn't need to say "and here they're typing on their computer." That's redundancy. Better: say something that adds context or emotion the image doesn't convey.
3. One idea per sentence
Sentences with five subordinate clauses are the enemy of narration. The viewer needs to process while watching. One idea per sentence helps both pace and comprehension.
4. Mark pauses and emphasis
In the script, indicate where the narrator should breathe, when to pick up the pace, and which words to stress. A script with no markings forces the narrator to improvise, and improvisation rarely lines up exactly with the edit.
5. Match the tone to the brand
Define the register and listen to a sample before choosing a narrator. A formal brand can use a warm voice; what matters is whether the performance fits the message and audience.
6. Estimate realistic timing
Time a reading of the script at the intended pace, in the intended language, with the planned pauses. The word count that fits varies with the performance and with space reserved for visuals, music, or silence. Adjust the script to that reading instead of treating a words-per-minute figure as a rule.
How to choose a narrator
- Demo in the real context: ask the narrator to record 30 seconds of your specific script, not just to listen to their generic demo.
- Tone consistent with the brand: professional ≠ stiff; warm ≠ casually informal.
- Direction-friendly: a narrator who takes direction and offers variations makes it easier to adjust the performance.
- Recording quality if remote: many narrators record from home. Verify that the acoustic setup is professional.
Common recording mistakes
- Reverberant sound from a poorly treated room.
- Uneven pace because the script was written without thinking about cadence.
- Clipping or distortion from poor level balance.
- Inconsistent tone between takes if you record in separate sessions.
- No alternates. Asking for only one final take without options limits choices in editing.
Mix and post-production
The voiceover doesn't end when it's recorded; that's where it begins. The mix decides how loud it sits against music and SFX:
- Voice intelligibility: adjust music and ambience during speech and listen to the full result. A fixed 5–10 dB difference does not guarantee clarity in every mix. Audition’s ducking controls let you adjust attenuation and transitions for the material.
- Light compression to even out the volume without sounding artificial.
- EQ: boost intelligibility without turning the voice nasal or radio-like.
- De-essing: reduce overly prominent "S" sounds, especially with narrators who have very sharp diction.
Related concepts
Frequently asked questions
What's the difference between a voiceover and a translation voice-over?
A voiceover narrates from outside the scene: the narrator isn't part of what's on screen. A translation voice-over is laid on top of the original voice of someone speaking on camera, which can still be heard faintly underneath, as in many dubbed documentaries and interviews.
How much script fits in a 30-second video with a voiceover?
Less than you'd think, because some of those seconds belong to the visuals, the music and the pauses. The reliable way to know is to read the script aloud at the intended pace and time it. If it doesn't fit, cut words rather than reading faster, which sounds rushed and is harder to follow.
How do you write a voiceover script for a reel or short video?
Open with a line that hooks viewers in the first two seconds, keep to one idea per shot and end with a concrete action. Write the script in two columns, visuals on the left and narration on the right, so you can check that the voice adds information instead of describing what's already on screen.
Can you record a professional-quality voiceover at home?
Yes, if the room is treated: a closet full of clothes or a corner hung with blankets absorbs reverb better than an empty room with a better microphone. Record at a consistent distance, leave headroom so the level doesn't clip, and listen to the take on headphones before calling the session done.