Camera resolution vs storage — the real math behind “4K everywhere.”

The “4K everywhere” camera specification is now a default request in commercial CCTV projects, driven by falling 4K camera prices and the reasonable assumption that more pixels means better evidence. For some applications this is correct. For others, it produces a storage cost two to four times higher than necessary without improving the forensic value of the footage for the actual use cases the system will serve.

The storage math behind camera resolution is not complex, but it is rarely done at design time. When it is done after the system is sized and the NVR storage is ordered, the result is either an undersized system that doesn’t meet the required retention period or an oversized system with a capital cost that wasn’t justified by the forensic requirements. The retention requirements that define the floor — what insurance carriers specify, what premises liability defense requires, and what local ordinance mandates — are covered in the video retention policy post; this post covers how resolution choice affects the storage capacity needed to meet those retention targets.

The bitrate math — where storage cost actually comes from

Storage cost is a function of bitrate, not resolution directly. The bitrate determines how many bits per second the camera produces and, therefore, how much disk space a day of recording requires. Resolution is one input to bitrate; frame rate, codec, scene complexity, and motion activity are the others.

A 64-camera system with all cameras at 4K H.265 at 8 Mbps average requires approximately 16 TB of storage per day. A 30-day retention system requires 480 TB of raw storage — roughly $15,000–$25,000 in NVR hardware at current pricing, before cameras, switches, and labor. The same system at 1080p H.265 at 2 Mbps requires approximately 135 TB for 30 days — about one-third the storage cost, for footage that serves most forensic use cases equally well.

Scene complexity and motion recording — the variables that move the real number

Fixed bitrate encoding at a manufacturer’s published typical bitrate is a starting point, not a guaranteed figure. Scene complexity — the amount of motion and detail change from frame to frame — drives bitrate up with variable bitrate (VBR) encoding. A camera watching a quiet parking lot at 2 AM may encode at 0.5 Mbps. The same camera during a busy afternoon may hit 8 Mbps. Storage planning that uses only the low-motion figure produces a system that fills up faster than expected during busy periods and starts overwriting footage before the retention target.

Motion-triggered recording can reduce actual storage consumption by 60–80% versus continuous recording for general-coverage cameras. For applications where continuous recording is required — insurance policies increasingly specify it for certain camera locations — that reduction disappears. For general coverage areas, motion-triggered recording at high resolution can make 4K viable at a storage cost that would otherwise require 1080p for the same retention period.

Where 4K actually justifies the storage cost

4K provides meaningful forensic value over 1080p in specific applications where the additional pixel density matters to the investigative outcome:

• Wide-area coverage with identification requirement: A single camera covering a large lobby, parking structure, or outdoor plaza where individuals need to be identified from the full-frame shot. At 1080p, the pixels-on-face to total-frame-pixels ratio may be too low for positive identification at the scene width. 4K doubles the linear pixel density, which can change an unidentifiable face to an identifiable one without reducing the coverage area.
• License plate recognition from general security cameras: The pixel density requirement for LPR from a standard security camera viewing multiple lanes is often met only by 4K or higher resolution at typical capture distances. The LPR placement math post covers the pixels-per-foot calculation at capture distance — the same geometry that determines whether a general security camera can serve double duty as an LPR capture point.
• Retail and occupancy analytics: Behavior analytics engines that count people, detect dwell time, or map traffic patterns benefit from higher resolution for accurate body detection across large floor areas.

Where 4K doesn’t improve the forensic outcome

4K adds storage cost without adding forensic value when:

• The camera watches a narrow hallway or door frame where the subject fills most of the 1080p frame already. 4K in a 4-foot-wide corridor produces more pixels per face, but the face was already recognizable at 1080p.
• The camera is in a low-light environment without adequate illumination. Resolution is limited by light level, not pixel count. A 4K camera in inadequate light produces a higher-resolution dark frame that is no more useful than a 1080p dark frame. Camera manufacturer selection also intersects with NDAA compliance requirements in certain facilities — the NDAA Section 889 compliance post covers which manufacturers are available and which are prohibited from federal-funding-adjacent facilities, constraining the 4K product options in ways the resolution spec sheet doesn’t reflect.
• The use case is motion detection or perimeter alerting, not identification. A camera watching a fence line for intrusion events doesn’t need 4K to detect a person crossing; it needs adequate coverage area and reliable motion detection, both of which work at 1080p.

The specification approach that produces the right resolution mix

The right process starts with the forensic use case for each camera, not a site-wide resolution preference. For each camera in the design, two questions determine the right specification: What does this camera need to identify or detect? And at what distance from the camera does that identification or detection need to happen? Those parameters define the minimum pixel density at the subject, which determines the minimum resolution given the coverage area. Working backward from use case to resolution produces a design where some cameras are specified at 4K because the use case requires it and most cameras are at 1080p or 4MP because that’s what the use case actually needs. Our CCTV design process for Atlanta and Southeast commercial buildings includes the per-camera use-case analysis that produces a resolution mix aligned with both the forensic requirements and the storage budget — rather than a single resolution applied uniformly because it was on the specification template.