1. Decide what voiceover adds that the picture cannot
Voiceover is useful when the video needs orientation, a relationship between visible spaces, a small set of verified facts, or a clear campaign action. It is less useful when it merely names every room the viewer can already see or reads a compressed property description over a short montage.
| Audio approach | Useful when | Primary strength | Watch for |
|---|---|---|---|
| Property voiceover | The sequence needs guided relationships or verified context | Explains what to notice and why scenes belong together | Overwriting, unsupported claims, no room for the image |
| Agent on camera | The agent adds current context or answers a real question | Connects a named person to the campaign | Long self-introduction, rambling property walkthrough |
| Natural sound and music | The footage supplies meaningful place or movement | Lets the environment and pace carry the story | Music rights, unclear property orientation |
| Music plus concise text | A short social cut has one simple property or event job | Works without a spoken performance | Dense text, inaccessible design, vague promise |
| Hybrid | An agent intro hands off to property narration or natural sound | Separates identity, property evidence, and atmosphere | Too many voices or audio layers for the runtime |
Write a one-sentence narration job before writing the script: “The voice will orient the viewer from the main living area to the covered terrace and then direct them to current property details.” If the job only says “make the video engaging,” the recording brief is not ready.
2. Choose the agent, voice talent, or approved synthetic voice deliberately
No voice path is universally correct. Choose based on the property campaign, available performance, consent, rights, languages, revision speed, continuity, and who can approve the exact output. Record the decision instead of treating the voice as a disposable preset.
| Path | Useful fit | Production needs | Rights and review questions |
|---|---|---|---|
| Agent voice | Personal campaign, direct expertise, seller-facing consistency | Quiet capture, rehearsal, pronunciation, pickups | Who may edit and reuse the recording? Which listings and channels? |
| Voice professional | Controlled performance, languages, recurring production | Voice brief, audition, session or directed read, revision scope | Term, territory, media, paid use, cutdowns, revisions, archive |
| Approved AI voice | Repeatable drafts or production with explicit authorized voice use | Provider terms, pronunciation controls, version review, regeneration | Whose voice or likeness? What consent, disclosure, storage, and use apply? |
Do not clone, imitate, or imply a real person's voice without explicit authority. Do not select a voice by making assumptions about the kind of resident a property should attract. The performance can support pace, clarity, warmth, restraint, and campaign identity without targeting protected-class preferences or inventing a buyer persona.
3. Write a voice brief the narrator can perform
“Luxury” and “cinematic” are not complete directions. Translate the property and edit into observable instructions: conversational or formal, close or expansive, measured or brisk, where the meaning turns, which terms need pronunciation, and where silence protects a reveal.
Real estate video voiceover brief
Property and campaign: [approved identity / property overview / event / update]
Audience task: [what the intended viewer should understand or do]
Voice job: [orientation / property relationships / context / action]
Performance direction: [conversational / restrained / clear / measured / other specific terms]
Avoid: [announcer tone / false urgency / exaggerated adjectives / imitation]
Pronunciation: [street / building / subdivision / materials / agent / brokerage]
Numbers: [approved reading for address / price / area / dates / times]
Scene markers: [line-to-scene map / required pauses / optional lines]
Required identity or notice: [exact approved wording and placement]
Technical handoff: [file type / sample rate if specified by editor / mono or stereo request / naming]
Usage: [named channels / organic or paid / versions / term / territory]
Approval and pickups: [decision owner / included corrections / deadline condition]
Ask for two or three complete performance directions only when the editor has a real choice to make. Recording ten isolated versions of every sentence creates a selection problem and can produce a final read with inconsistent pace and tone.
4. Prepare the property script for speech and the actual edit
Listing copy is not automatically narration. Spoken language needs a clear subject, natural phrasing, deliberate numbers, and enough time for the property image to register. Lock a rough visual sequence first, then time a real read against it.
- Start from approved facts. Use the current listing record, not memory or a generated summary alone.
- Assign every line a scene. If no frame supports the line, move the fact to the property page or remove it.
- Write one meaning unit at a time. Break long listing sentences into ideas the viewer can follow with the edit.
- Prepare names and numbers. Decide how addresses, measurements, prices, dates, and abbreviations should be spoken.
- Mark optional clauses. Give the editor a clean removal point if the actual performance runs long.
- Protect visual pauses. Leave room after an opening, reveal, room change, or final action.
- Read the entire script aloud. Time the intended performance instead of forcing it into a generic word budget.
Scene 1 · opening evidence: [specific property-first hook]
Scene 2 · orientation: [where the viewer is and what connects]
Scene 3 · primary proof: [one verified detail the frame clearly supports]
Scene 4 · second zone: [new property area and its relationship]
Scene 5 · differentiator: [visible, defensible property evidence]
Scene 6 · current close: [approved identity and one next action]
The real estate video script template provides 30-, 60-, and 90-second starting structures, a property brief, and fact-review process. Use the property-first hook guide when the opening still sounds interchangeable with every other listing.
5. Record an agent or voice professional for clean editing choices
Capture consistent, complete takes before chasing individual words. The editor needs a stable performance, clean room tone, and clearly labeled alternatives more than a folder of unrelated fragments.
Before the read
- Approve the script version and keep it visible to everyone in the session.
- Choose the quietest practical space and reduce changing background noise.
- Keep microphone position, narrator position, and input settings consistent.
- Rehearse property names, numbers, line breaks, and scene changes before recording.
- Provide water and time for a natural reset rather than directing a rushed delivery.
During the read
- Record a complete natural take, then give one specific adjustment at a time.
- Ask for pickups as complete phrases or sentences so the edit has matching rhythm.
- Leave clean silence before and after each take and capture room tone.
- Say the take and script version into the slate or preserve them in the file name and notes.
- Stop and correct a property fact; do not plan to hide a factual error with editing.
After the read
- Back up the original recordings before noise reduction, compression, or cleanup.
- Log the selected complete take and every pickup used.
- Listen for clipped words, room changes, mouth noise, background events, and mismatched energy.
- Send the editor the approved script, pronunciation notes, room tone, and selection record together.
If the agent is recording remotely, a quiet, consistent phone recording close to the speaker can be easier to repair than a distant microphone in a reverberant room. Test the exact setup, transfer an original-quality file rather than a heavily compressed message, and let the editor confirm the handoff before the space is reset.
6. Generate AI narration only through an authorized, reviewable workflow
Synthetic narration can make revisions and recurring production easier, but it does not remove the need for script approval, voice authority, pronunciation, performance direction, provider-term review, or inspection of the finished property claim. Treat each generated file as a draft performance.
AI narration record
Provider and model or voice ID: [service / version / preset or authorized custom voice]
Voice authority: [source / consent / permitted uses / restrictions]
Script version: [approved text / date / approver]
Pronunciation controls: [dictionary / phonetic input / alternate render]
Performance settings: [documented controls that affect the final read]
Generated versions: [file IDs / dates / selection]
Human review: [facts / names / numbers / artifacts / tone / scene timing]
Disclosure review: [property / platform / ad / brokerage / local requirements]
Storage and deletion: [source voice / generated files / access / retention]
Approved uses: [listing / channels / versions / paid or organic / term]
- The person or rights holder authorized the voice use and the intended media.
- The selected voice does not imitate or imply an unauthorized person.
- Every name, number, address, date, unit, and property term is pronounced correctly.
- The delivery does not add unintended emphasis that changes the property claim.
- No glitch, missing word, repeated phrase, abrupt breath, or timing artifact remains.
- The generated narration matches the current approved script and visual sequence.
- The provider terms and campaign rights cover the named organic, web, email, MLS-related, and paid uses.
- The team completed the responsible disclosure and recordkeeping review.
The AI real estate marketing disclosure checklist helps inventory the generated change, verify the property, preserve source comparisons, and review each placement. It is an operational framework, not a substitute for the current platform, MLS, brokerage, contract, or legal requirements that apply to the exact campaign.
7. Edit and mix the voiceover around the property story
Select the performance for meaning first. Cleanup cannot rescue a read that emphasizes the wrong feature, races through the reveal, or sounds disconnected from the property. Once the edit is stable, make the voice intelligible across the actual music, ambience, captions, and destination versions.
- Place the complete selected read. Check whether the property route still makes sense.
- Edit for meaning. Use pickups at natural phrase boundaries and preserve a consistent performance.
- Remove unnecessary copy. Let the image carry visible detail instead of accelerating the voice.
- Clean conservatively. Reduce distracting noise without making the narrator brittle or unnatural.
- Balance scene by scene. Keep speech clear as music and footage change.
- Use music under the voice deliberately. Lower or simplify competing elements rather than raising everything.
- Listen on more than one playback system. Check headphones, a phone speaker, and the delivery environment available to the team.
- Inspect every export. Transcoding, alternate crops, and social versions can introduce a different balance or timing problem.
Avoid treating a single loudness value as proof that the mix works everywhere. Destination processing, dialogue dynamics, music density, phone playback, and alternate versions all affect intelligibility. Keep a clean voice stem, music stem, mix master, and final approved exports so a corrected fact does not require extracting speech from a compressed video.
The real estate video music guide covers the mood brief, licensing record, editorial placement, mix, and post-publication claims. The editing workflow carries narration through rough sequence, review, versioning, quality control, and archive.
8. Caption the final narration and package every version
Caption the final approved audio, not an earlier script or draft voice. Correct automatic output against the recording and property source sheet, then decide which destinations need a separate caption track, an open-caption export, or both.
| Deliverable | Purpose | Final check |
|---|---|---|
| Approved voice master | Clean reusable narration source | Script version, take, facts, artifacts, rights |
| Corrected caption file | Selectable synchronized text where supported | Words, names, numbers, timing, language, version |
| Open-caption social export | Text rendered into a named video version | Crop, contrast, interface overlap, synchronization |
| Clean video master | Future captions, languages, corrections, crops | No unintended text or compressed-platform source |
| Music-only or alternate-language version | Approved campaign or destination need | Story clarity, rights, identity, matching property facts |
| Approval and usage record | Connect source, voice, script, mix, files, and permissions | Approver, date, destinations, term, update owner |
W3C's current WCAG 2.2 guidance includes captions for prerecorded synchronized media and explains that captions represent necessary speech and non-dialogue audio. Review the W3C caption criterion as an accessibility reference, then determine the obligations and delivery formats that apply to the organization and destination. The real estate video captions guide covers transcript correction, open and closed versions, on-screen property text, post copy, and maintenance.
9. Approve the complete narration inside every final export
A correct audio file can become an incorrect property video when it is paired with the wrong shot, an old price card, an altered generated frame, or a stale destination. Review the exact published package, not only the script and standalone voice stem.
- The narration matches the current approved property record and campaign moment.
- Every spoken property claim has visible support or an appropriate verified source.
- The opening, scene order, voice, music, captions, text, and CTA tell the same property story.
- Names, address, price, measurements, dates, times, status, agent, and brokerage are correct.
- Pronunciation, emphasis, pauses, edits, and generated speech artifacts have been reviewed.
- The voice and performance rights cover the exact files, channels, paid uses, languages, term, and territory.
- Required identity, accessibility, altered-media, advertising, and placement reviews are complete.
- Mobile playback keeps speech intelligible without overpowering music or abrupt level changes.
- Caption files and open-caption versions match the final approved voice word for word.
- Someone owns rerecording, regeneration, copy updates, replacement, and retirement when a material fact changes.
Voiceover approval record
Listing and campaign: [property / version / purpose]
Source record as of: [date / authoritative property record]
Script and voice: [version / narrator or voice ID / authority]
Selected audio: [file / take / generated version]
Final exports: [horizontal / vertical / branded / clean / captions]
Destinations and rights: [named uses / paid or organic / term / territory]
Approved by: [responsible reviewers / date]
Change triggers: [fact / price / status / event / destination / rights]
Update owner: [person / contact path / replacement process]
Review a cinematic, narrated property preview
Westby turns an active listing's photos and details into a cinematic video preview with property narration. Enter the property address or listing link, then review the visual sequence and spoken direction before deciding whether the video fits the approved campaign.
Create a free listing-video preview.
Frequently asked questions
Should a real estate listing video have a voiceover?
Use voiceover when spoken guidance adds property relationships, context, or a clear next step that the images and concise text cannot carry alone. Music-only, natural sound, captions, or an on-camera agent may be better when narration would repeat the visuals, crowd a short edit, or create claims that are difficult to maintain.
Who should narrate a real estate video?
Choose the agent, a voice professional, or an approved synthetic voice based on the campaign, performance needs, rights, consent, review capacity, languages, revision path, and desired continuity across listings. The voice should serve the property story rather than imitate an identity or style the team is not authorized to use.
How do you make a property voiceover sound natural?
Write for speech, rehearse against the actual scene order, mark meaning and pronunciation instead of every breath, record several complete options, keep property names and numbers easy to say, and leave room for the visuals. A rushed read usually needs less copy, not faster delivery.
Do narrated property videos need captions?
Plan a corrected synchronized caption track for spoken and meaningful audio, then choose closed captions, open-caption exports, or both based on the destination. Accessibility obligations depend on the context, but captions also provide a usable text version when audio is unavailable or unclear.