On this page
- 01Before you start
- 02Where it sits in the project
- 03System architectures: IP vs Analog HD vs hybrid
- 04Camera types, form factors and applications
- 05How to specify a camera
- 06What coverage really means — the DORI framework
- 07Recording platforms: DVR, NVR, XVR and VMS
- 08Channels, bandwidth and decoding
- 09Storage: HDDs, retention and RAID
- 10Network infrastructure for IP CCTV
- 11Transmission media and accessories
- 12Design process: survey to selection
- 13Integration and analytics
- 14Check that it works
- 15Common problems and how to fix them
- 16Frequently asked questions
- 17Talk to PSS about CCTV
A CCTV system is a closed-circuit surveillance arrangement of cameras, transmission media, recording platforms, storage, network infrastructure and operator interfaces. It helps organisations monitor activity, investigate incidents, protect assets and keep visibility across buildings, industrial sites, commercial properties and public spaces.
It is three disciplines joined into one: cameras that turn light into a signal, a recorder (or VMS) that stores that signal, and software that lets an authorised operator retrieve a single minute of it months later. The cameras are the visible part — and often the least decisive. Excellent cameras on a badly sized recorder or undersized network produce a system nobody can use.
Coverage is not a camera count. It is the amount of useful detail that survives at the point where something happens: how many pixels land on a face, a number plate or a door handle. That quantity is measurable, and it is what separates a design from a shopping list.
Before you start
- Surveillance objectives per area (detect, observe, recognise, identify) and the IEC edition that will govern DORI numbers in the contract.
- IP, Analog HD (HD-over-coax) or hybrid/XVR — and why, including any coax that can still be reused.
- Camera schedule draft: indoor/outdoor, type, resolution, lens/FOV, mounting height, day/night lighting.
- Recorder path: DVR, NVR, XVR or VMS — channel count, incoming bandwidth, decode limits, storage and retention days.
- Network: PoE standard and total budget, VLAN plan, uplink capacity, copper versus fibre runs over 100 m.
- Integrations: access control, intrusion, BMS, fire/emergency displays, video wall — who supplies each interface.
- Cybersecurity and privacy rules for the site (credentials, remote access, retention, NDAA or local procurement where they apply).
- Who signs commissioning, day/night acceptance tests, and as-built documentation.
Where it sits in the project
CCTV belongs to the ELV/security package, but it never stands alone. It needs mounting points and containment, power and UPS, an IT network with VLANs and an IP plan, rack space, and night-time lighting — a camera cannot record what light never reaches. It often interfaces with access control, intrusion alarm, BMS, fire/emergency displays and video walls.
The site sequence is: containment, cabling, second fix, configuration, commissioning, as-built documentation. Most delays trace back to storage and network decisions taken after the cabling was already closed.
System architectures: IP vs Analog HD vs hybrid
IP CCTV
IP cameras transmit digital video streams over Ethernet. Compatible cameras receive data and power through PoE on Cat5e/Cat6 (or fibre uplinks for longer runs). A typical system includes IP cameras, PoE switches, an NVR or VMS servers, surveillance storage, operator workstations and optional analytics/integrations.
Advantages: wide resolution range (2 MP, 4 MP, 5 MP, 4K); flexible expansion; edge and server analytics where supported; centralised multi-site management; PoE reduces separate camera power cabling; flexible recording and integration architectures.
Design considerations: bandwidth for all streams; PoE budget and uplink capacity; camera/recorder compatibility; cybersecurity and segmentation; large sites may need redundancy, distributed recording or a dedicated surveillance network.
Analog HD (HD-over-coax)
Compatible cameras send HD video to a DVR over coaxial cable. Common technologies include HD-TVI, HD-CVI, AHD and legacy analogue. Formats are not universally interchangeable — camera and DVR must match.
Advantages: can reuse suitable existing coax; often a practical upgrade from legacy analogue; does not normally need an Ethernet switch at every camera for video; supports HD resolutions depending on equipment.
Design considerations: expansion limited by DVR channels and cabling; camera power is usually separate unless PoC is used; resolution, distance and quality depend on technology; analytics/integration vary by model; inspect existing cables, connectors and PSUs before reuse.
| Feature | IP CCTV | Analog HD |
|---|---|---|
| Transmission | Digital IP streams | HD video over coax |
| Typical cabling | Cat5e/Cat6 or fibre backbone | RG59/RG6 coax where suitable |
| Recorder | NVR or VMS | DVR or compatible XVR |
| Camera power | PoE or separate | Separate or supported PoC |
| Addressing | IP address | Usually physical recorder channel |
| Scalability | Network, licences, bandwidth, storage | DVR channels and supported config |
| Best fit | New builds, enterprise, advanced integration | Suitable coax upgrades |
Neither architecture is automatically best. Choose from existing infrastructure, required image quality, integration, budget and expansion plans.
Hybrid / XVR
Hybrid systems mix technologies. An XVR may support selected HD-over-coax formats, legacy analogue and a defined number of IP cameras. Verify supported technologies, analogue/IP channel limits, resolution per channel type, incoming bandwidth, decoding limits and analytics compatibility. Hybrid can reduce upgrade cost — capabilities are model-specific.
Camera types, form factors and applications
Start from the surveillance objective — not the housing shape or advertised megapixel count. Consider environment, target distance, field of view, lighting, mounting height, required detail and analytics.
Dome
Common on ceilings/walls in offices, retail, corridors and receptions. Compact and relatively discreet; selected models offer vandal resistance (e.g. IK10). Check indoor/outdoor rating, lens/FOV, IR reflection from the dome cover, mounting height and artificial lighting. The housing does not guarantee that the lens direction is hidden from every angle.
Bullet
Elongated housing, often outdoor: perimeters, gates, parking, loading bays, yards. Many include IR or white-light illumination. Verify IP rating, temperature range, illumination performance, brackets and site conditions. IP67 describes ingress protection under defined tests — it does not establish impact resistance or suitability for every hazardous environment.
Turret
Adjustable assembly aimed after installation. Used indoors and outdoors at entrances, corridors, warehouses and parking. Some designs reduce internal IR reflection problems seen with certain domes; night performance still depends on placement, illumination and camera design.
PTZ
Motorised pan, tilt and optical zoom (e.g. 25× or 36× on selected models), presets, patrols, alarm positioning, auto-tracking and analytics where supported. Used on large compounds, industrial sites, stadiums and public areas.
Critical limit: a PTZ cannot cover every direction in detail at once. When it turns to follow a target, other areas may go unseen. Keep fixed cameras on critical views; use PTZ for zoom and operator flexibility. Optical zoom recovers detail through the lens; digital zoom only enlarges pixels already captured.
Fisheye, panoramic and multi-sensor
Ultra-wide or multi-sensor panoramic views (often 180°/360°) with dewarping on camera, recorder or VMS. Useful for receptions, open offices, retail floors, warehouses and lobbies. A wide view may reduce camera count for general observation — it does not guarantee identification pixel density across the whole scene.
Multiple imaging modules in one housing are useful at corners, intersections and large spaces. Check sensor count, resolution per sensor, stream count, FOV options, combined bitrate, recorder/VMS compatibility and licensing. Sensor count is not the same as independent recorder channel capacity.
Indoor / outdoor and specialised
Outdoor needs suitable IP protection, temperature limits, humidity/UV resistance, brackets/junction boxes, surge protection/grounding where required, and vandal resistance in accessible locations. Marine, corrosive, hazardous and extreme-temperature sites need equipment approved for those conditions.
Elevator cameras need compact wide-angle units, careful mounting, approved travelling cable or specialised transmission, coordination with the elevator contractor, and no interference with safety circuits. Wireless bridges are not reliable in every shaft.
| Type | Primary purpose | Typical applications |
|---|---|---|
| Low-light / full-colour night | Usable images in limited light | Entrances, warehouses, vehicle areas |
| IR camera | Monochrome night with IR | Perimeters, yards, corridors |
| Thermal | Heat patterns | Perimeter and selected industrial |
| ANPR/LPR | Vehicle plates | Gates, parking, roads |
| People-counting / face-capture | Counts or faces under defined conditions | Retail, entrances, controlled points |
| Explosion-protected / corrosion-resistant | Approved harsh environments | Industrial, marine, oil-and-gas |
Specialised analytics need correct positioning, optics, lighting, software and configuration. High resolution alone does not deliver reliable plate or face capture.
How to specify a camera
Resolution
| Description | Typical pixels | Approx. MP |
|---|---|---|
| 720p | 1280 × 720 | ~0.92 |
| 1080p / Full HD | 1920 × 1080 | ~2.07 |
| 4 MP class | Often 2560 × 1440 | ~3.7 |
| 5 MP class | Model-dependent | ~5 |
| 4K / 8 MP class | Often 3840 × 2160 | ~8.3 |
2 MP remains common for general surveillance; 4–5 MP often balances detail and resources; 4K suits selected critical views. Higher resolution increases processing, network and storage when other settings stay comparable. Sensor size matters: 4 MP on a 1/1.8″ sensor collects more light per pixel than 4 MP on 1/2.8″. A headline 0.005 lux figure is often measured with long exposure — which blurs anyone walking through the scene.
Lens and focal length
Fixed, manual varifocal, motorised varifocal or optical zoom. Representative fixed lengths: 2.8 mm, 4 mm, 6 mm and longer. Shorter focal length → wider FOV; longer → narrower view with more magnification at the same distance. Actual angle depends on lens and sensor — use the manufacturer’s FOV data. For critical coverage, size the lens from required pixel density at the target, not a nominal mm alone.
Low light, IR and full-colour night
Low-light performance depends on sensor, aperture, exposure, gain/noise reduction, processing, IR/white light and target motion. Long exposure brightens but blurs motion; high gain adds noise. For vehicle gates, capture of moving targets at night often matters more than a minimum-illumination headline.
IR enables monochrome night vision. Advertised ranges (30/50/80 m) are typically quoted against a highly reflective target — plan on roughly 60–70% of the printed figure, and never mount a camera where a nearby wall sits inside its own IR beam. Full-colour night cameras use large apertures, sensitive sensors and/or supplemental white light; test under real night conditions — white light may be unsuitable for some sites.
WDR, BLC and HLC
True sensor WDR (e.g. from about 120 dB) helps entrances with bright backgrounds; digital WDR mainly stretches contrast in software. BLC adjusts exposure for subjects against bright light; HLC reduces intense highlights such as headlights, depending on implementation. Headline dB values do not replace a practical image test at the real doorway.
Frame rate and compression
Common recording settings: 10–30 FPS depending on camera, region, resolution and recorder. Higher FPS means smoother motion and more temporal detail — and more bandwidth/storage/decoding load. A corridor rarely needs the same FPS as a fast vehicle gate.
| Codec | Notes |
|---|---|
| H.264 / AVC | Widely supported |
| H.265 / HEVC | Often lower bitrate at similar quality — savings are not fixed |
| Smart codecs (e.g. H.265+) | Scene-aware savings; size storage on busy-hour bitrate |
| MJPEG | Frame-by-frame; much higher bandwidth/storage |
Camera, recorder, VMS and clients must all support the selected codec. H.265 often cuts bitrate roughly 40–50% versus H.264 under suitable conditions — treat that as a planning range, not a guarantee.
IP and IK ratings
IP = ingress of solids/water under defined tests. IK = mechanical impact resistance. IP67 ≠ impact-proof; IK10 ≠ weatherproof. Verify both ratings, temperature limits and accessories for the real installation. Outdoor housings commonly target IP66/IP67 (IEC 60529) and IK10 (IEC 62262, 20 J) where vandal resistance is required.
Integration and procurement notes
ONVIF Profile S (Profile T for events) is the practical interoperability ticket; “ONVIF compatible” alone means little. For US federal and related prime-contractor work, NDAA Section 889 Part B restricts five named vendors (Huawei, ZTE, Hytera, Hikvision, Dahua) — confirm project procurement rules early.
What coverage really means — the DORI framework
Coverage is what an operator can reliably determine from the recorded image — not merely what appears in live view.
Traditional DORI guidance associated with IEC 62676-4:2014 used approximate pixel densities in pixels per metre (PPM / px/m):
| Level | Density (2014) | Intended objective | Where it belongs |
|---|---|---|---|
| Detection | 25 px/m | Presence of a person or object | Perimeter and distant approaches |
| Observation | 62.5 px/m | General appearance and behaviour | Corridors and yards |
| Recognition | 125 px/m | Recognise a familiar person / distinguish an object | Sensitive department entrances |
| Identification | 250 px/m | Greater detail to support identification | Checkpoint, cashier, strong room |
These are planning guidelines, not performance guarantees. IEC 62676-4:2025 replaced the four-level set with seven levels (20/40/80/125/250/500/1500 px/m); in that edition 250 px/m is “characterise”, and identity verification needs 500 px/m. Always quote the edition used in the contract — the same number means different things in each.
Approximate maximum distance for a required density:
D(m) = (horizontal pixels × focal length mm) ÷ (required px/m × sensor width mm)
Worked example — 4 MP (~2688 px horizontal), 1/1.8″ sensor ~7.6 mm wide, 4 mm lens: identify at 250 px/m → (2688 × 4) ÷ (250 × 7.6) ≈ 11.3 m; recognise at 125 px/m → ≈ 22.6 m; detect at 25 px/m → ≈ 113 m. Horizontal FOV ≈ 2 × atan(7.6 ÷ (2 × 4)) ≈ 87°.
Useful coverage also depends on mounting height/angle, lighting and WDR, motion blur, compression, obstructions and the recorded resolution/FPS — not only the camera’s capture resolution. A camera mounted too high may capture the top of a head; a camera facing a bright doorway may silhouette people without true WDR.
Overview vs identification: wide-angle cameras give context; dedicated cameras at doors capture detail. That split is usually more effective than one camera asked to do both. Practical corridor guidance often lands around one camera per 25–40 m; entrances at roughly 2.2–2.6 m mounting height, offset from the door line so faces are captured before the turn — always validate against the required DORI level and site geometry.
For each camera document: purpose, target, critical area, required detail level, lighting, lens/position, recording stream settings, and how acceptance will be tested (site survey, FOV simulation, day/night test).
Recording platforms: DVR, NVR, XVR and VMS
DVR
Receives compatible analogue/HD-over-coax video on physical inputs and encodes for recording. Check channel count, supported formats, max resolution/FPS per channel, drive bays and capacity, codecs, simultaneous playback, display outputs and integrations. Advertised channels do not guarantee every channel at every resolution/FPS.
NVR
Receives IP streams over the network. Check max channels, incoming recording bandwidth, resolutions/FPS, codec and camera interoperability, HDD bays/capacity, RAID if available, live-view/playback decoding, network throughput, analytics and integration.
XVR / hybrid
Model-specific mix of coax and IP channels. Verify allocation rules, resolution limits, bandwidth and storage before specifying.
VMS
Software to manage cameras, recording servers, playback, users, events and integrations — on dedicated servers, appliances or distributed platforms. Capabilities often include multi-site management, roles, alarm handling, evidence export, analytics, access-control integration, maps, failover and health monitoring. A VMS does not automatically include all required storage hardware, OS, databases or licences.
| Feature | DVR | NVR | VMS |
|---|---|---|---|
| Typical source | Analog HD / coax | IP cameras | Supported IP + recording platform |
| Architecture | Dedicated recorder | Dedicated recorder | Software / distributed |
| Expansion | Channel/model limits | Channels + bandwidth | Licences + platform design |
| Multi-site / integration | Model-dependent | Model-dependent | Often broader when licensed |
Channels, bandwidth and decoding
Channel count alone does not define recording capacity.
Example: 32 cameras × 4 Mbps average = 128 Mbps incoming. The NVR must support at least that load with headroom for peaks and secondary streams. A 160 Mbps rating may look adequate for 128 Mbps — still check peak bitrates and the datasheet’s operating limits.
Decoding limits how many streams the device can show live or in playback at given resolutions. A recorder may accept many 4K streams for recording but only display a few at native resolution, using sub-streams for multi-view grids. Distinguish capture resolution, recording resolution, live-view decode, playback decode, display output and simultaneous stream count — especially for video walls.
Outgoing bandwidth (remote live view, playback, export) may be specified separately. Central monitoring centres must size the path between recorder, core network and clients.
Storage: HDDs, retention and RAID
Use surveillance-rated drives (e.g. WD Purple, Seagate SkyHawk families) only after checking the recorder’s compatibility list, capacity limits and workload rating.
Simplified continuous-recording estimate (decimal GB):
Storage GB ≈ Total bitrate (Mbps) × hours/day × retention days × 0.45
Equivalently, GB/day per camera ≈ bitrate (Mbps) × 10.8.
Example — 24 cameras at 4 Mbps, 24 h/day, 30 days: 24 × 4 = 96 Mbps → 96 × 24 × 30 × 0.45 = 31,104 GB ≈ 31.1 TB before design allowances and RAID overhead. Another check: 32 cameras × 4 Mbps × 30 days ≈ 41.5 TB usable before 10% headroom and RAID parity.
Variables that change the result: resolution, FPS, average bitrate, codec/quality, scene motion, continuous vs event recording, audio/metadata, platform reservations. Statements like “one 4 MP camera needs 15–20 GB/day” are examples under particular settings — not universal. Validate with manufacturer calculators and commissioning measurements.
Event-based recording can reduce storage when motion/alarms are accurate; critical areas often still need continuous recording. Recorders with 1–8+ bays have per-drive and total capacity limits — do not assume every retail drive size is supported. RAID 0/1/5/6/10 behaviour and usable capacity vary; RAID is not a substitute for backup or evidence export. Agree retention days (15/30/60/90…) in writing and align with privacy/regulatory policy.
Network infrastructure for IP CCTV
PoE
| Standard | Common name | Max PSE power / port |
|---|---|---|
| IEEE 802.3af | PoE | 15.4 W |
| IEEE 802.3at | PoE+ | 30 W |
| IEEE 802.3bt Type 3 | PoE++ | Up to 60 W |
| IEEE 802.3bt Type 4 | PoE++ | Up to 90 W |
Power at the camera is lower than the PSE rating because of cable loss. PTZ cameras with heaters/IR often draw 25–40 W — check per-port capability and total PoE budget, not port count alone. Example: 16 × 12 W = 192 W camera demand; 4 × 30 W PTZ = 120 W — a 60 W total budget cannot run those four PTZs.
Ports, uplinks, distance and security
Size ports with expansion spare. Aggregate bitrate example: 24 × 5 Mbps = 120 Mbps toward the recorder — a gigabit uplink may suit that load only after allowing peaks, live viewing and other traffic; larger groups need fibre or higher-capacity core links.
Standard copper Ethernet channels are normally limited to 100 m (TIA-568 / cabling standards). Longer runs use fibre, intermediate switches, compatible extenders or manufacturer extended-PoE modes (sometimes at reduced data rates — not equivalent to standard 100 m gigabit). Fibre does not power a PoE camera by itself — the remote end needs a switch or local power.
Prefer managed switches for enterprise CCTV (VLAN, PoE monitoring, QoS, STP, SNMP, diagnostics). Put cameras on a dedicated VLAN/segment; change defaults, restrict services, update firmware, limit remote access, sync time securely and monitor health. Segmentation does not replace hardening of cameras, recorders and servers.
Transmission media and accessories
| Component | Function | Typical use |
|---|---|---|
| Cat5e/Cat6 | IP video + PoE | IP cameras |
| Coaxial (RG59/RG6) | Analog / HD-over-coax video | DVR systems |
| Fibre + SFP | Long-distance / backbone | Inter-building links |
| PoE injector / extender | Power or extend a PoE link | Selected long runs |
| Video balun / media converter | Adapt supported interfaces | Coax/UTP or fibre-Ethernet |
| Surge protection, rack, patch | Protection and organisation | Outdoor and comms rooms |
Design process: survey to selection
Site survey: layout and boundaries; entrances/gates/loading; high-risk points; existing cabling/network; day/night lighting; mounting locations/heights; power and comms rooms; environment; recording/monitoring rooms; integrations; expansion.
Purpose per camera (examples): main entrance — facial/visitor detail (angle, WDR, height); vehicle gate — vehicles/plates (lens, shutter, illumination); perimeter — detect/monitor (distance, night, weather); elevator — cabin activity (wide angle, safe cabling); cashier — transactions (close-up, FPS, privacy).
Camera schedule fields: ID, location, indoor/outdoor, type, resolution, lens/FOV, height, objective, day/night illumination, IP/IK, analytics, transmission, switch/port, bitrate/FPS, retention, recorder/VMS allocation.
Select the recording platform from camera count, resolution/codec/FPS, average and peak bitrates, incoming bandwidth, simultaneous view/playback, storage/retention, redundancy, operator displays and integrations. Document expansion allowances for ports, PoE, uplinks, channels, storage, licences, rack, cooling and power.
Integration and analytics
- Access control — associate video with door-forced, denied-access and timeline events via connectors, APIs or vendor modules.
- Intrusion — alarm may call a camera view, bookmark recording, notify operators or drive a PTZ preset.
- BMS — selected health/status points only; do not assume the BMS replaces CCTV management or recording.
- Fire / emergency — associate displays or bookmarks where specified; never interfere with life-safety operation or introduce unapproved control dependencies.
- Video walls — size decoders/workstations for simultaneous high-resolution streams; channel count ≠ decode capacity.
Analytics may include motion, line crossing, intrusion, area entry/exit, person/vehicle classification, people counting, queue monitoring, left/removed object, LPR and facial analytics where legally permitted. Performance depends on position, geometry, light, target size, config and environment. Edge vs server analytics and licences differ — define measurable acceptance criteria and test on site.
Check that it works
- Each camera meets its stated DORI objective in day and night tests at the target distance.
- Recording resolution, FPS and codec match the design; playback and evidence export succeed.
- Incoming bandwidth and decode capacity stay within the recorder’s rated limits under peak load.
- Storage retention matches the written days after measuring actual average/peak bitrates.
- PoE per-port and total budget cover connected cameras with headroom; VLAN addressing and time sync are correct.
- Offline-camera and storage-failure alarms notify as designed; integrations/analytics meet acceptance criteria.
- As-built drawings, camera schedules, network data, licences and operating docs are complete.
Common problems and how to fix them
IR fog / white dome night image
Dirt/moisture on cover; internal IR reflection; missing isolation gasket; wall too close to IR beam. Clean cover, remove films, seat gasket, adjust position, review IR/exposure.
Offline IP camera / IP conflict
Duplicate IP, wrong VLAN/mask, cable/PoE faults, plug-and-play NVR private camera network misconfigured. Check addressing, switch counters, PoE and status pages (failed ping alone is not proof if ICMP is disabled).
Analog HD loss / flicker
Bad BNC/coax, distance, ground loop, weak PSU, wrong format. Inspect connectors, voltage at camera, grounding; isolators treat symptoms, not root grounding faults.
Stuttering playback
Network congestion, incoming/outgoing bandwidth, decode limits, HDD errors, client limits, codec mismatch. Review loads, try sub-stream, check HDD health; H.265 does not automatically fix stuttering.
Shorter retention than designed
Underestimated bitrate, added cameras, schedule/RAID changes. Recalculate from actual average/peak bitrates and usable capacity.
PoE camera dead
Budget, standard mismatch, cable loss, bad port. Verify camera watts, per-port and total budget, cable and logs.
Third-party “online” but missing features
Partial ONVIF, unsupported codec/resolution, proprietary analytics, missing licences. Test required functions, not only live video.
Design mistakes that look like product faults
4K + 2.8 mm at 40 m sold as “identification” (~detection only); storage sized from catalogue defaults; no NTP; no camera-to-switch-port map; default passwords left on.
Frequently asked questions
Can third-party IP cameras connect to an existing NVR?
Often yes via compatible ONVIF profiles — basic video does not guarantee proprietary analytics, PTZ or every event type. Verify exact models and required functions.
How do I calculate HDD storage?
From camera count, average bitrate, hours/day and retention (formula above). Validate with a manufacturer calculator.
What is the maximum distance for an IP camera?
Standard copper channel ≈ 100 m. Longer: fibre, intermediate switch or approved extender/extended PoE under stated conditions.
Does CCTV need internet?
No for local recording and viewing on a proper LAN. Internet may be needed only for selected remote/cloud features.
Is 4K always better than 2 MP?
No. Lens, distance, lighting, position, compression and motion blur matter as much as megapixels.
Can a 32-channel NVR record 32 cameras?
Only if bandwidth, resolution, codec, FPS, decoding, storage and network also fit.
Are all surveillance HDDs interchangeable?
No — check capacity, workload, interface and compatibility lists.
Can one PTZ replace multiple fixed cameras?
Not when continuous multi-direction coverage is required.
What is the most important design factor?
Defining what each camera must capture and what the footage must prove — then sizing optics, coverage, recorder, network and storage from that.
Talk to PSS about CCTV
PSS designs, installs, and maintains CCTV surveillance systems, from camera schedules and DORI objectives through recorders, PoE and storage sized for real site loads. Contact the PSS team to book a first site survey.
