IP camera buying guide: choose the image your operation actually needs
Resolution is only one input. A usable surveillance image depends on the task, scene width, lens, distance, lighting, motion, placement, compression, storage, network power, and system compatibility.
The most expensive camera can still produce unusable evidence when it is aimed at too wide a scene, mounted too high, pointed into strong backlight, or configured to record at a bitrate the storage system cannot sustain.
A professional camera specification begins with the operational question: what must the video allow an operator or investigator to determine? Once that is clear, the project can define the target area, level of detail, lens, resolution, lighting performance, frame rate, retention, integration, and infrastructure.
Start with the task, not the megapixel count
Before selecting hardware, write one sentence for every camera. Examples: “show whether a person entered the loading dock,” “provide sufficient detail to compare a face with a known person at the reception door,” or “maintain an overview of traffic flow across the yard.” Those are different tasks and may require different camera positions, fields of view, pixel densities, lighting, and recording settings.
Overview, activity, characteristics, identity, object detail, or event verification.
Width, distance, height, movement, lighting, and obstructions.
Lens, field of view, sensor size, focus, and depth of field.
Pixel density, shutter, WDR, low light, motion, and compression.
VMS, PoE, network, storage, cybersecurity, and lifecycle support.
What is happening?
Monitor movement, occupancy, direction, congestion, or general activity across a larger area.
What characteristics are visible?
Clothing, object type, vehicle color, direction, and other scene-specific attributes.
Who or what is it?
Obtain enough visual information for the project’s defined comparison or identification task.
What should be detected?
People, vehicles, line crossing, occupancy, loitering, or another supported analytic event.
Pixel density and DORI: use them as planning models
Megapixels describe the total number of image pixels. They do not describe how many pixels cover the target. When the same resolution is spread over a wider scene, each person, face, plate, box, or object receives fewer pixels. Lens choice and camera placement therefore matter as much as the resolution number.
The 2014 edition of IEC 62676-4 popularized the operational requirements commonly summarized as DORI: detection, observation, recognition, and identification. The 2025 edition updates the operational-requirement framework. DORI remains common industry shorthand, but it should be treated as a planning model—not a promise that a person, plate, or object will be usable under every condition.
Common DORI planning progression
Conceptual—not a distance chartPixel density does not account for everything that affects usable video. Light direction, motion blur, shutter speed, sensor noise, focus, lens quality, compression, angle, occlusion, environmental conditions, and display or export quality can change the result. Validate the intended task in the actual scene.
Scene width is the hidden resolution decision
A camera covering a 12-foot doorway and the same camera covering a 120-foot yard do not deliver the same target detail. The correct question is not “How many megapixels?” but “How wide is the target area at the distance where detail is required?”
Control the capture point
A narrower, well-lit view at a predictable distance can produce better face detail than a higher-resolution camera attempting to cover the entire lobby.
Use the geometry
Frame the aisle, traffic direction, pallet face, or checkpoint. Corridor orientation may use sensor pixels more efficiently when the camera supports it.
Separate overview from evidence
One wide camera may document activity while a second, narrower camera covers the gate, lane, dock position, or other evidence point.
Plan by zones
Long distances, changing light, vegetation, weather, motion, and analytics often require several controlled views rather than one extreme wide shot.
Choose the lens before assuming you need more resolution
The lens defines the field of view and concentrates the scene onto the image sensor. Focal length and sensor size work together: a shorter focal length generally produces a wider view, while a longer focal length produces a narrower view with more target pixels at distance. The same focal length can produce a different field of view on a different sensor size.
Useful for small rooms and overview scenes. Detail is spread across more scene width.
Useful when the target and context both matter, such as entrances or dock positions.
Places more pixels on a distant or controlled target, but covers less surrounding area.
Simple when the scene is known
A fixed lens can be cost-effective and reliable when camera location and target width are already defined. Verify the actual field of view before ordering.
Useful when framing needs adjustment
A manual or motorized varifocal lens provides an adjustable focal-length range. It helps commission the exact scene, but it does not replace a coverage calculation.
Field rule: if the target appears too small, increasing digital zoom after the event does not create missing detail. Narrow the scene, move the camera, select different optics, add a dedicated camera, or increase resolution only after the geometry is correct.
When higher resolution is worth it—and when it is not
Higher resolution is valuable when the lens, scene width, target distance, lighting, network, storage, and VMS can use it. It can support wider coverage, more target detail, or digital investigation. It can also increase bitrate, storage, processing demand, low-light tradeoffs, and cost.
| Project condition | What resolution can help | What it cannot fix |
|---|---|---|
| Wide scene with several evidence points | Provide more total pixels, especially when the optics and sensor support the scene | A target that is too small because one camera is covering too much area |
| Controlled entrance or lane | Increase detail and investigation flexibility | Motion blur, poor angle, blocked faces, glare, or insufficient light |
| Low-light moving target | Help only when the full imaging system can maintain usable exposure and noise | Long shutter times that smear movement or aggressive noise reduction that removes detail |
| Long retention requirement | Preserve more recorded detail when storage and bandwidth are designed for it | Insufficient recorder capacity, incompatible codec support, or an unrealistic bitrate cap |
Do not compare resolution alone. Compare the full image chain: sensor, optics, processing, lighting, frame rate, compression, bitrate control, and the target at the required distance.
Lighting can matter more than megapixels
A camera specification should describe daytime, nighttime, transitions, backlight, reflections, vehicle headlights, and expected motion. A static daytime screenshot is not enough.
Bright and dark at once
Entrances, windows, garage openings, tunnels, and outdoor shadows can require wide dynamic range. Do not compare WDR only by a dB claim; scene complexity, motion, artifacts, and processing affect forensic usability.
Exposure versus motion
A camera may brighten a dark scene with a longer exposure, but moving people or vehicles can blur. Evaluate low-light detail on motion, not only on a static scene.
Useful, but not universal
Infrared can provide monochrome video in darkness. Range, target reflectivity, angle, insects, rain, fog, nearby surfaces, and overexposure can affect the result.
WDR scenes
Wide dynamic range scenes contain very bright and very dark areas at the same time. Typical examples include a dark lobby facing a sunlit door, a parking garage entrance, or an outdoor scene with hard shadows. The goal is not the largest WDR number; it is usable detail in both regions without unacceptable artifacts.
Color at night versus infrared
Color can be valuable for clothing, vehicle, or object description, but it requires adequate visible light or specialized low-light performance. IR illumination can support darkness but normally produces monochrome video. White-light illumination may preserve color and deter activity, but it also changes the site, can affect neighbors, and must be planned intentionally.
Shutter speed and motion
Exposure settings affect whether a moving target is sharp. A bright still frame can hide a motion-blur problem. Validate walking, running, vehicle speed, doors, forklifts, and other relevant movement during the hardest lighting condition.
Choose a form factor for the environment—not the label
| Form factor | Best fit | Main planning caution |
|---|---|---|
| Turret / eyeball | Flexible fixed views, often with fewer dome-reflection issues around integrated IR | Can be more visually obvious; confirm weather, vandal, and mounting ratings |
| Dome | Discreet indoor or outdoor fixed coverage with a protected viewing window | Dirt, scratches, condensation, internal reflection, and IR bounce can degrade images |
| Bullet | Visible directional coverage, perimeters, walls, and longer-lens applications | Mounting arm stability, tampering, wind, cable entry, and appearance |
| PTZ | Active operator control, optical zoom, tours, and event response across a wide area | A PTZ views one direction at a time; it should not be the only coverage for every critical fixed scene |
| Multi-sensor / panoramic | Wide situational awareness with several sensors or stitched views | Edge distortion, pixel density by direction, stream count, storage, licensing, and mounting height |
| Specialty camera | Thermal, explosion-protected, corridor, covert, license-plate, or other defined tasks | Use only after requirements, certifications, integration, and limitations are understood |
Form factor does not determine image quality by itself. Compare the imaging, optics, environmental ratings, cybersecurity, analytics, and integration of the exact model.
Frame rate, shutter, and analytics must match the event
More frames per second can provide smoother motion and more temporal samples, but it increases processing, bitrate, and storage. Select frame rate from the event: slow pedestrian monitoring, cash handling, manufacturing motion, vehicle travel, and forensic review may have different requirements.
Analytics also depend on the scene. Confirm the camera model, supported analytic, mounting height, angle, target size, environmental limits, lighting, and expected behavior. “AI camera” is not a complete specification. Validate the event logic and false-alarm tolerance in the actual deployment.
Calculate storage from measured or engineered bitrate
Storage is not determined by resolution alone. Scene motion, image complexity, frame rate, compression format, compression settings, GOP structure, noise, WDR, analytics streams, audio, event recording, and bitrate-control mode all affect the recorded rate.
Measured or estimated for the configured stream and scene.
Continuous or event-based recording per day.
Required online video history.
Include primary, secondary, audio, and metadata streams when recorded.
Add resilience, overhead, growth, and recorder requirements.
This decimal approximation assumes 24 hours of continuous recording. Multiply by retention days and camera count, then add headroom for RAID or erasure protection, filesystem and database overhead, exports, spare capacity, failover, and future growth. Verify the recorder vendor’s calculator and tested stream settings.
H.265 can reduce bitrate relative to H.264 in suitable implementations, but actual savings vary by scene and encoder settings, and every VMS, client, analytic, export workflow, and hardware decoder must support the selected codec. AV1 is emerging in surveillance and can offer another efficient option on supported systems. Compatibility comes before theoretical compression savings.
Design the network and PoE budget for maximum demand
A camera is a networked powered device. The design must support its maximum power requirement, not only its typical daytime draw. Integrated IR, heaters, blowers, wipers, audio, illuminators, motors, and PTZ movement can increase demand.
Maximum draw
Use the exact model’s data sheet and the intended accessories and temperature range.
Per-port and total PoE
Confirm standard, port capability, total budget, power supply, allocation behavior, and reserve.
Link and environment
Use compliant cabling, distance, pathways, surge protection, grounding, and outdoor transitions.
UPS and failure plan
Size switch, recorder, storage, network, and camera runtime around the required outage behavior.
Also account for uplink bandwidth, multicast or multiple live viewers, recording traffic, failover, firmware downloads, analytics metadata, and remote access. Segment surveillance appropriately and coordinate the design with the organization’s IT and cybersecurity policies.
Verify VMS, recorder, and ONVIF support by exact model
“Supports ONVIF” is not enough. ONVIF profiles define sets of standardized features, and products must be verified in the official ONVIF conformant-products database. Confirm the exact camera, firmware, recorder or VMS version, profile, codec, event, audio, metadata, edge-storage, PTZ, and analytics functions required by the project.
Profile S is associated with basic IP video streaming, while Profile T supports more advanced video features. Profile G addresses recording and storage, and Profile M addresses metadata and events for analytics. A profile claim does not mean every proprietary feature will work across every vendor combination.
Obtain a tested compatibility statement from the VMS or NVR vendor when a feature is operationally important. Live video alone does not prove that motion events, metadata, audio, edge recording, PTZ control, health monitoring, search, exports, or firmware management will work as required.
Treat camera lifecycle and cybersecurity as buying criteria
IP cameras are computing devices connected to the network. Select manufacturers and platforms with documented vulnerability handling, signed or validated firmware processes, security guidance, product support, and an understandable end-of-support policy.
Common camera-selection mistakes
Start with the task, scene width, target distance, optics, lighting, and motion. Resolution supports the design; it does not replace it.
Use separate overview and evidence views when the target zones, widths, and distances conflict.
Test nighttime, transitions, backlight, motion, weather, headlights, reflections, and the final recording—not only the live view.
A PTZ can be pointed elsewhere when an incident occurs. Maintain fixed coverage for mandatory scenes.
Use measured or engineered bitrate, recording schedule, codec, scene, retention, camera count, overhead, and resilience.
Verify the exact products, profiles, firmware, and required functions in official databases and compatibility documentation.
IR, heaters, PTZ motors, illuminators, and accessories may raise consumption when the system is under the hardest conditions.
Cybersecurity and long-term maintainability are part of the camera specification, not an afterthought.
Camera specification checklist
Send the floor plan, target distances, and retention goal
EhubAmerica can help narrow camera, lens, recorder, storage, and PoE options around the operating requirement before equipment is ordered.
Common questions
Is 4 MP enough for a professional security camera?
It can be, depending on the target width, distance, lens, lighting, motion, and required detail. A well-positioned 4 MP camera covering a controlled entrance can outperform a higher-resolution camera covering an excessively wide scene. Validate pixel density and recorded image quality at the target.
When is 4K or 8 MP worth it?
It is useful when the project needs more total pixels for a wider scene, more target detail, or investigation flexibility and when the optics, sensor, lighting, bandwidth, VMS, decoder, and storage can support it. It does not fix poor angle, focus, motion blur, glare, or inadequate lighting.
What is more important: lens or megapixels?
They work together. The lens and sensor determine the field of view, while resolution determines how many pixels are distributed across that view. A narrower correctly framed view may place more useful pixels on the target than simply increasing megapixels.
Should I choose a fixed or varifocal lens?
Use a fixed lens when the camera position and required field of view are known. A varifocal lens is useful when commissioning requires adjustment or several possible target widths must be accommodated. Model the exact focal range and sensor combination before buying.
Does WDR matter at entrances?
Often, because an entrance can contain a dark interior and bright exterior simultaneously. Verify WDR on moving people in the actual light direction and time of day. A larger published dB number alone does not guarantee a more usable image.
Is color night vision always better than infrared?
No. Color can provide useful descriptive information but requires adequate visible light or suitable low-light performance. IR supports darkness but generally produces monochrome video. The correct choice depends on the evidence requirement, motion, scene, lighting policy, and environmental conditions.
How much storage does one IP camera need?
Use average recorded bitrate, recording hours, retention days, camera count, and all recorded streams. As a quick decimal estimate, 1 Mb/s recorded continuously uses about 10.8 GB per day. Add recorder overhead, redundancy, spare capacity, failover, exports, and growth.
Does H.265 always cut storage in half?
No. Savings depend on the encoder, scene, motion, noise, image settings, GOP structure, bitrate control, and compared H.264 configuration. Confirm VMS, client, hardware-decoder, export, and analytics compatibility before selecting the codec.
Is ONVIF compatibility enough to mix camera and recorder brands?
Not by itself. Verify the exact models in ONVIF’s conformant-products database, identify the required profile, and confirm the specific functions with the recorder or VMS vendor. Live video may work while metadata, analytics, audio, edge recording, PTZ, or health events do not.
Should I choose dome, turret, or bullet cameras?
Choose from the environment, mounting, vandal risk, visibility, optics, IR behavior, service access, and aesthetics. Form factor alone does not determine image quality. Compare exact models and installation conditions.
Can a PTZ replace several fixed cameras?
It can provide active coverage and optical zoom across a large area, but it views one direction at a time. Critical scenes that must always be recorded generally still need fixed coverage or a multi-sensor design.
How do I size a PoE switch for cameras?
Confirm the PoE standard and maximum power requirement of every camera and accessory, then verify per-port capability and the switch’s total PoE budget with the installed power supply. Add design reserve and account for IR, heaters, PTZ motors, and worst-case conditions.
What should be tested before accepting the installation?
Review the final recorded video at every target during representative day, night, backlight, motion, and weather conditions. Confirm field of view, focus, pixel density, exposure, WDR, IR, frame rate, bitrate, retention, analytics, time synchronization, failover, exports, health monitoring, and user access.
Primary technical references
- Axis Communications: Pixel density based on IEC 62676-4:2025
- Axis Communications: Pixel density and the DORI model
- Axis Communications: Lenses in surveillance
- Axis Communications: Wide dynamic range
- Axis Communications: IR in surveillance
- Axis Communications: Bitrate control for IP video
- ONVIF: Profiles and feature overview
- ONVIF: Conformant Products database
- Cisco: PoE, powered devices, and switch power allocation
- Hanwha Vision: Long-term firmware support policy
The camera-buying rule to remember
Start with the evidence requirement, constrain the scene, select the lens and position, verify pixel density and lighting, then engineer the recording, network, storage, power, integration, and cybersecurity around the final image.
More pixels can be useful. They become valuable only when the complete system delivers those pixels on the target, under the conditions that matter, for the retention period the operation actually needs.
When you are ready, browse video-surveillance equipment or send EhubAmerica your camera schedule and site requirements.
We help integrators and enterprise teams match cameras, recorders, networking, power, and storage to project requirements. Final system design, installation, privacy compliance, cybersecurity, commissioning, and acceptance remain with the qualified project team.