Camera Recording Incident Review and Chain of Custody

Incident review and chain of custody for camera recordings: clock synchronization, locking recordings, export format, SHA-256 hash verification and a handover record.
In this article8 sections

Three days after an incident, someone searches for the recording and one of two things happens: looped recording has deleted the relevant hours, or the time on the footage found is ten minutes off from the access control logs. A camera recording can be evidence for an internal investigation, an insurance claim, a customer dispute or legal proceedings. But footage recorded with the wrong time, filmed off a screen with a phone, or whose handling history is unknown loses much of that value. Below we explain how to set up incident review and export so that integrity and traceability are preserved. This isn’t legal advice. You can assess how to present the source, date and integrity information of a camera recording within the framework of audit evidence types and reliability.

What is a chain of custody?

A chain of custody documents who handled a piece of evidence, when and how, from the moment it was collected. For digital evidence, ISO/IEC 27037 provides guidance on identifying, collecting, acquiring and preserving evidence; NIST SP 800-86 describes integrating forensic techniques into incident response. For a video recording, the goal is to be able to show that the file presented is the same as the original recording in the system and that its time information is correct.

In practice, you need to be able to answer three questions: Was this file taken from the original recording? Has it changed since it was taken? Who has handled it? If one of the answers is missing, the chain is broken.

The answers to these questions come from technical settings and written records. If the system doesn’t keep per-user logs, no form can answer “who exported it?”; without a form, even the best log doesn’t show who handed it to whom. Both must be set up together.

What must be ready before an incident

A chain of custody isn’t set up on the day of the incident. If the following aren’t configured before the incident, the post-incident procedure stays incomplete no matter how meticulous it is.

  • Correct time: All cameras, NVRs and servers should be synchronized to a reliable NTP source, and clock drift checked periodically. A wrong clock makes it impossible to match events across different cameras and access control logs.
  • Personal user accounts and logs: Recording playback and export actions should be logged per user (access and audit trail).
  • Export permission: Should be given only to defined roles; a reason and approval should be required.
  • Integrity feature: The VMS/NVR’s digital signature, watermark or hash features should be enabled.
  • Written procedure and form: Who requests, who searches, who exports, who it’s handed to.

A common mistake with time is cameras taking NTP from the NVR while the NVR itself isn’t tied to any source. The whole system is internally consistent but minutes off from real time. Cameras, NVRs and the access control system should be tied to the same corporate NTP source. Daylight saving and time zone settings should also be checked per device.

Steps after an incident

  1. Request and logging: Incident number, requester, time and place of the incident, relevant cameras.
  2. Protection: Locking or bookmarking recordings in the relevant time range to protect them from looped deletion. This step should be done in the first hours.
  3. Search: Finding the relevant moments with smart search; combining multiple cameras and access control events on a timeline (smart video search).
  4. Export: In the original format or the VMS’s signed format, with a player. The exported range should cover a little before and after the incident.
  5. Integrity record: The file’s hash (e.g. SHA-256) is calculated and written on the form; if the VMS’s signature verification tool can be used, its result is recorded.
  6. Handover record: To whom, on what medium and when the file was handed over; signature of the recipient.
  7. Retention and destruction: How long the incident file and protected recordings will be kept, and their destruction at the end of the process (retention and destruction).

The protection step is usually skipped. In a system with 15 days of retention, if an incident is noticed on day 14 there isn’t even time to search; lock first, then search. In export, generic formats like MP4 make sharing easier but don’t carry the signature and watermark information of most VMS products. Keep the original or signed format as the primary copy, and give out the generic-format copy only for viewing.

Verifying integrity: hash practice

A hash is the file’s fingerprint. If a single byte changes, the value changes completely. Calculate the hash right after export, write it on the form, and have the receiving party do the same calculation and compare values. On Windows, the PowerShell command Get-FileHash -Algorithm SHA256 .\incident-2026-114.mkv is enough; on Linux, sha256sum incident-2026-114.mkv.

If there’s more than one file, list each file’s value on a separate line. If files are compressed and handed over as a single archive, record the value of both the archive and the files inside it. The hash shows the file hasn’t changed since. That the recording is original is shown by the VMS’s signature verification and access logs; the two complement each other.

Fields the handover form should include

Field Example
Incident no. / requester 2026-114 · OHS Department
Cameras and time range DOCK-03, DOCK-04 · 14:05–14:25
System clock verification NTP offset < 1 s (time of check)
Exporting user Personal username
File name, size, hash SHA-256 value
Recipient / date / medium Signature, encrypted USB or secure share

Don’t leave the clock verification line blank. Comparing the camera’s and recorder’s clocks with a reference clock at the time of export and writing down the offset makes the time statement defensible.

Privacy and third parties

An exported recording may also include people unrelated to the incident. Depending on the purpose of sharing, masking may be needed; not sharing recordings with unauthorized people and using a secure transfer method also matter under KVKK (KVKK and camera recordings). A masked copy is a new file. The original must be preserved, and the masked version recorded with a separate hash.

Common mistakes

  • Not protecting the recording on the day the incident is noticed, and losing it to looped recording
  • Filming the screen with a phone instead of exporting the original recording
  • Exporting with a shared user account
  • Stating a time without checking clock offset
  • Circulating the file by email and losing track of how many copies exist
  • Limiting the exported range to the moment of the incident and losing what came before and after
  • Handing over in a signed format without including the player, so the other party can’t open the file
  • Calculating the hash but not having the recipient verify it

Frequently asked questions

Is filming the screen with a phone enough?

It can be used for quickly sharing information but doesn’t replace the original recording; resolution, time information and integrity are lost. The original recording must be exported from the system.

Is an exported recording valid in court?

Evaluating evidence is a legal matter and up to the court. What can be done technically is to document the recording’s integrity, time and whose hands it has passed through.

How long should we keep an incident recording?

For the duration of the relevant investigation or legal proceedings; when the process ends, it’s destroyed according to the organization’s policy. An incident recording must be protected separately from looped recording.

To set up the procedure together with your system’s export and integrity features, see our smart search and incident review service.

Support

Contact us for support or a discovery session on this topic.

Contact form
[email protected]

Eretek Video Surveillance

IP cameras, recording and central management infrastructure. We plan blind spots, retention periods and the camera network around the risks of each site.

Knowledge Center

Other articles