How to Calculate CCTV Storage: A Complete Guide
Accurate storage estimation is the backbone of any CCTV system design. Underestimating storage leads to lost footage and compliance failures. Overestimating wastes budget on hardware that will sit half empty for its entire service life. Both mistakes are made routinely, and both come from the same place: a number pulled from a vendor brochure instead of a calculation built on the actual variables.
Those variables are camera resolution, frame rate, compression codec, recording schedule and retention period. Get all five right and the array you buy will still be sized correctly in year three.
Resolution and codec set the baseline
Resolution has the largest impact on storage consumption. A 2MP (1080p) camera generates roughly 5-8 GB per day at 15 FPS with H.264 compression, while a 4K (8MP) camera can produce 25-40 GB under the same conditions. Upgrading to H.265 (HEVC) compression typically reduces storage by 30-50% compared to H.264, which is why it is the preferred codec for modern deployments.
The ranges above are ranges for a reason. Bitrate is scene-dependent, and the dependency is stronger than most people expect. A camera watching a static corridor at night produces a fraction of the data of the same camera watching a car park with wind-blown trees behind it. Modern encoders make this worse before they make it better. Smart codecs that lower the bitrate on quiet scenes are excellent for the average, but the array must be sized for the busy day, not the average day.
Frame rate and recording schedule
Frame rate is the variable most often over-specified. Above roughly 12-15 FPS, additional frames add storage without adding much forensic value for general surveillance. Higher rates are justified where motion is fast and detail matters between frames, as in cash handling, gaming tables, production lines and vehicle gates, and are hard to justify anywhere else.
Recording schedule and motion detection settings also play a critical role. Continuous 24/7 recording demands far more storage than motion-triggered recording, which can reduce write volume by 50-70% in low-traffic areas. Combining scheduled recording with motion-based triggers lets you optimise storage without sacrificing coverage during critical hours.
Motion recording has a failure mode worth stating plainly. If the detection threshold is set high enough to keep storage down, it will eventually miss the event that matters: a slow approach at the edge of frame, or movement during heavy rain. A common compromise is continuous recording at a low frame rate with a motion-triggered increase, which keeps a full timeline while spending the bitrate where something is happening.
Retention, RAID and overhead
Retention policies vary by regulation and industry. Banks may require 90 days of footage, while retail environments typically keep 30 days. The retention period that applies to a specific site is a legal question, not an engineering one, and it must be confirmed against the current regulation and any contractual obligation before the array is specified.
RAID configuration adds overhead. RAID 5 sacrifices one disk for parity and RAID 6 sacrifices two, but the redundancy is essential. A surveillance array runs its disks continuously, and disk failure is a certainty rather than a risk. Always factor in a 15-20% overhead for filesystem and RAID when sizing the array.
- Size on peak-day bitrate, not on the vendor average figure
- Add 15-20% for filesystem and RAID overhead before you order
- Deduct parity disks from usable capacity: RAID 5 loses one, RAID 6 loses two
- Leave headroom for cameras added after handover, because most sites grow
- Verify the retention requirement against current regulation, not against habit
- Use surveillance-rated drives; desktop drives fail early under continuous write load
One more constraint is easy to miss. Storage capacity and write throughput are different problems. Forty 4K cameras writing simultaneously can saturate a modest array long before the disks fill, and the symptom is dropped frames rather than a clear error. Check sustained write bandwidth against the aggregate bitrate of every camera recording at once.
Does the resolution actually deliver what you need?
Storage is only worth spending on footage that can be used. A 4K camera at the far end of a car park may deliver no more usable identification than a 2MP camera at half the distance, while consuming five times the storage. The pixel density on target (pixels per metre at the point of interest) decides whether footage supports identification, recognition or only detection.
Before committing to a resolution across the site, check the pixel density each camera actually achieves at its target distance with our DORI Distance Calculator
Then estimate the exact array you need, in total terabytes and recommended drive count, from camera count, resolution, codec and retention period with our free CCTV Storage Calculator
Size the array from measured bitrates where you can get them, from conservative estimates where you cannot, and record the assumptions alongside the design. When someone asks in two years why the system only holds 22 days instead of 30, the answer is almost always in one of those assumptions: usually the scene complexity, occasionally a frame rate that was raised after commissioning and never re-costed.