A CCTV clip can identify far more people than the individual making a request. Faces, vehicle registrations, staff ID badges, computer screens, voices and the layout of secure premises may all be visible in a single recording. Knowing how to redact CCTV footage therefore means more than applying a blur effect. It requires a controlled process that protects third-party information while preserving enough context for a fair, lawful and usable disclosure.
For UK organisations responding to DSARs, FOI requests, litigation disclosure, regulatory enquiries or internal investigations, the decision must be capable of later explanation. The footage released should contain no reversible redactions, and the record should show what was withheld, why, by whom and under what authority.
How to redact CCTV footage for a defensible disclosure
Start by preserving the source recording. Export or secure a copy in its original form, record the source system, camera reference, time range and relevant timestamps, and restrict access to authorised reviewers. The original should not be edited in place. It may be needed to verify the accuracy of the redacted version, respond to a challenge or evidence the integrity of the disclosure process.
Next, establish the scope of the request. A DSAR may concern footage of the requester, but it does not automatically entitle them to unrestricted copies of every third party captured alongside them. Consider the date, location, cameras, duration and whether audio is included. A request for “all CCTV” is rarely a practical or proportionate instruction without further scoping, particularly where systems retain footage from multiple locations or long periods.
Review the clip frame by frame where necessary. Identify personal data and other restricted material before editing begins. The most obvious issue is a bystander’s face, but redaction decisions should also account for vehicle registration marks, names shown on uniforms, screen content, access-control information, audio, medical details and security-sensitive areas.
The appropriate treatment depends on the purpose and legal context. For a DSAR, the organisation must consider the requester’s right of access alongside the rights and freedoms of others. For FOI, personal information and other exemptions may apply. For legal or regulatory disclosure, the relevant rules, directions and confidentiality obligations will shape what can be provided. The practical redaction method may be similar, but the rationale must be specific to the request.
Use true redaction, not visual concealment
A black box, pixelation or blur is not automatically a secure redaction. If the original pixels, audio track or underlying video data remain available, the information may still be recoverable. The same concern applies where an editor merely places an overlay over the footage and exports a project file or layered format.
A defensible workflow creates a new disclosure copy in which the concealed content cannot be restored. Redactions must be permanently applied to every affected frame, including frames where a person enters or leaves the scene. If audio contains third-party personal data or confidential information, it should be removed or silenced rather than simply hidden in the player interface.
Metadata also needs attention. File names, embedded camera labels, GPS data, comments, timestamps and export history can reveal information that is not intended for disclosure. Check what travels with the final file and retain the fuller technical record securely within the case file instead.
Track people carefully across the full clip
Video redaction is difficult because the subject moves. A face may be partially obscured in one frame, then become clear as the person turns or walks closer to the camera. Automated object tracking can reduce manual effort, especially across longer clips or multiple cameras, but it needs human review. Tracking may drift, fail during occlusion, or mistakenly follow a similar-looking person.
Set the redaction area generously enough to avoid exposing identifying features at the edges. Review transitions, cuts, reflections in glass, mirrors, vehicle windows and picture-in-picture views. A redaction that works in the main image can still fail where the same person appears in a reflection or another camera angle.
Where footage is low quality, do not assume a face is anonymous. A combination of clothing, gait, location, companions and timing may identify someone to people who know the circumstances. Equally, excessive redaction can remove the very context the requester is entitled to see. The objective is not to obscure more than necessary. It is to make a reasoned, proportionate decision that can be defended.
A controlled workflow for CCTV redaction
CCTV disclosure work benefits from defined roles. One person may conduct the initial review, another may apply the redactions, and a suitably authorised reviewer may complete quality assurance. Smaller teams may not be able to separate every task, but they should still document checks and avoid relying on memory or informal messages.
A practical case record should capture the request reference, footage sources searched, retention issues, relevant dates, applicable legal considerations and the outcome for each clip. For every redaction, record the timecode or frame range, the type of information withheld and the reason. This creates an auditable decision trail rather than a folder containing an unexplained edited video.
Use a clear set of review stages:
- confirm the source footage and request scope;
- identify all people and restricted information in the relevant sequence;
- apply permanent visual and, where required, audio redactions;
- review the exported disclosure copy independently against the source; and
- retain the source, redacted version and decision record securely under the case reference.
The final review is not a formality. Play the entire exported file, not just selected moments. Check that every redaction remains in place during movement, fast-forwarding and playback on a standard media player. Confirm that sound has been handled correctly, the clip begins and ends at the approved points, and the file contains no hidden tracks or metadata that undermine the decision.
Common failures that create disclosure risk
The first failure is using general-purpose editing software without confirming how it handles layers, project data and exports. A visible mask can look satisfactory while leaving the underlying content intact. The second is releasing a clip before checking all frames. A single missed frame can disclose a face, registration mark or screen display.
Another common failure is treating the redaction as self-explanatory. If a requester complains to the organisation, the ICO or a court, the team should be able to show the basis for its decision without reconstructing events from emails and spreadsheets. A defensible record is particularly valuable where several staff members have worked on the same request or where the footage forms part of a wider investigation.
Time pressure also causes avoidable errors. Statutory timescales do not reduce the need for care, but they do make early scoping essential. Identify CCTV sources promptly, secure footage before routine deletion, estimate the review effort and seek clarification where a request is genuinely broad or unclear. If an extension, refusal or partial disclosure is being considered, the legal basis and communications should be managed through the organisation’s established information-rights process.
When specialist tooling is justified
For a short, simple clip with one third party, manual work may be manageable if the organisation has a secure and validated process. The risk changes quickly when requests involve hours of footage, multiple cameras, moving subjects, audio or several overlapping legal issues. In those cases, disconnected tools and unstructured notes make consistency difficult to maintain.
A purpose-built platform such as Redaktr supports irreversible video redaction alongside structured reasons, controlled access and a complete audit trail. That matters because the output is not merely a modified video file. It is a disclosure decision with evidence behind it.
The most useful question before releasing CCTV is not “does the blur look right?” It is whether the organisation could explain, months later, exactly what was disclosed, what was protected and why. If the answer is clear from the case record as well as the footage itself, the process is doing its job.