How Security Cameras Work | Light, Sensors & Storage Explained

A security camera works by capturing light through a lens, converting it into an electronic signal via a CMOS or CCD image sensor, processing that signal into a digital video stream, and then storing or transmitting the footage locally or to the cloud.

Whether you’re setting up your first system or upgrading an old one, understanding the core process helps you pick the right hardware and avoid expensive mistakes. The chain from lens to live view involves three critical stages: the sensor that catches the light, the processor that compresses the data, and the storage or network that holds the result. Each stage has its own trade-offs, and knowing them separates a system that works from one that frustrates.

How Does A Security Camera Actually Capture Video?

The job starts the same way a digital camera works. Light enters through the lens and hits a grid of millions of tiny light-sensitive cells called photodiodes, which sit on an image sensor. Each photodiode corresponds to one pixel in the final image. The sensor measures the intensity of the light hitting each cell and converts that measurement into an electrical voltage. A Sensor DSP (Digital Signal Processor) then turns those voltages into a raw digital image, adjusts color and brightness, and sends the data to the compression engine.

CMOS vs. CCD Sensors: What’s The Difference In A Security Camera?

Almost all modern security cameras use a CMOS sensor. CCD sensors, once the standard for higher image quality, are now rare outside specialized industrial cameras. CMOS sensors consume less power, generate less heat, and can be made cheaper at high resolutions — which is why you see them in everything from a $30 indoor cam to a $400 4K bullet camera. CCD sensors historically handled low light better, but modern CMOS designs have closed that gap so completely that CCD is no longer an advantage worth chasing.

Video Compression: Why H.264 And H.265 Matter For Storage

Raw video from a 4K sensor would fill a hard drive in hours. Codecs like H.264 and H.265 shrink the file size by only storing the parts of the frame that change between images—your lawn stays the same, so the codec records the full lawn once and then only notes the moving car. H.265 (also called HEVC) halves the bitrate of H.264 at the same quality, which means you can store twice as much footage on the same drive or stream 4K over a slower connection. Most cameras released after 2020 support at least H.265, but double-check before buying if you plan to mix new and older cameras on the same NVR.

How Do Wireless Security Cameras Transmit Video?

Wireless cameras send the compressed video stream over your home Wi-Fi network using the 802.11 protocol, almost always on the 2.4 GHz band. Budget models often lack 5 GHz support, which matters because a router set to 5 GHz only will leave that camera with no connection. The camera is assigned an IP address just like your phone or laptop, and the stream travels over TCP/IP to either a local NVR, a cloud service, or directly to your phone app. A common mistake is placing a Wi-Fi camera too far from the router — the signal drops, and the recording stops without warning.

Motion Detection: PIR, Microwave, And Video-Based Methods

Security cameras use three main technologies to decide when to record, and each has a different strength.

  • Passive Infrared (PIR) sensors detect a change in infrared radiation — basically heat from a person or animal moving across the sensor’s field of view. They are the most common and the most power-efficient, but they struggle if the object is the same temperature as the background.
  • Microwave sensors emit a low-energy signal and measure how it reflects back. They can detect movement through thin walls, but that also means they trigger false alarms more often.
  • Video-based algorithms compare consecutive frames pixel by pixel. If enough pixels change value, the system triggers an event. This is how most modern IP cameras do person, vehicle, or animal detection — the camera’s onboard processor runs a small neural network to decide what changed.
Detection Method Best For Common Weakness
Passive Infrared (PIR) Saving battery life; general motion alerts Misses objects at ambient temperature
Microwave Covering large open areas False alarms from movement behind walls
Video-based AI Person/car/package recognition Higher power consumption; more expensive
Thermal Imaging Total darkness; long-range detection Very high cost; low detail for identification

Night Vision: How Security Cameras See In The Dark

Most security cameras use infrared (IR) LEDs that flood the area with light invisible to the human eye. The camera’s sensor picks up that reflected IR light and renders it in black-and-white. For total darkness without any visible glow, thermal imaging cameras detect the heat emitted by objects — a person at body temperature shows up clearly against a cold wall. IR is by far the more common and affordable option, but it has a limit: if the camera faces a bright light source like a streetlamp, the IR wash can create a glare that wipes out the whole image.

If you’re choosing between models right now, our roundup of the best camera for security installations compares real-world testing data on night vision range, motion detection accuracy, and storage options across the current top picks.

Analog vs. IP vs. Wireless Cameras: Which System Architecture Fits?

The type of camera you choose determines how the video signal gets from the lens to your eyes. Analog systems send a raw analog signal over coaxial cable to a DVR, which digitizes and compresses the footage. Analog is cheaper per camera but tops out at resolutions equivalent to around 1080p — you cannot get a true 4K image from an analog CCTV camera, even if the box says “HD.”

IP cameras convert the image to digital right at the camera and send the compressed stream over an Ethernet cable or Wi-Fi. They support much higher resolutions, person-detection AI, and easier remote access. Wireless cameras are just IP cameras that use Wi-Fi instead of a physical network cable, but they still need power — either a plug, PoE, or a battery.

NVR systems require IP cameras because they expect a digital stream. DVR systems require analog cameras. Mixing the two types will not work without a converter.

System Type Connection Max Typical Resolution Central Recorder
Analog CCTV Coaxial cable 1080p equivalent DVR
IP (wired) Ethernet (PoE) 4K and above NVR
Wireless (Wi-Fi) 802.11 (2.4/5 GHz) 4K NVR, cloud, or SD card
Cellular 3G/4G LTE (SIM card) 1080p-2K Cloud or SD card

Where Does The Footage Actually Go? Local vs. Cloud Storage

Every camera needs somewhere to write the video. Local storage means a hard drive inside a DVR or NVR, or a microSD card inside the camera itself. Local storage is faster, has no monthly fee, and keeps working if the internet goes down.

Cloud storage uploads clips or continuous video to a remote server. It protects footage if the camera is stolen or damaged, and you can view it from anywhere without setting up a VPN. The downside is the subscription cost — typically $3 to $10 per camera per month — and the bandwidth needed for 24/7 4K streaming. Many modern systems support a hybrid setup: record locally for reliability and back up motion events to the cloud.

Common Setup Mistakes And How To Avoid Them

  • Network frequency mismatch: Buying a 2.4 GHz-only camera when your router runs a 5 GHz-only network. Check your router settings before ordering.
  • IR facing a light source: Placing a camera with IR LEDs aimed at a streetlamp or porch light — the glare washes out the image. Position the camera so the lens points away from direct light.
  • Ignoring the 512 GB limit: Not all cameras support a 512 GB microSD card even if the slot fits one. Confirm the maximum capacity in the spec sheet or manual.
  • Analog resolution expectations: Expecting 4K clarity from an analog CCTV camera. Analog tops out well below 4K; buy IP cameras if resolution matters.
  • Skipping periodic formatting: Leaving a microSD card in the same camera for a year without reformatting increases the risk of corruption. Format the card inside the camera every 2-3 months.

Putting It Together: The Decision Sequence For A Working System

  1. Decide where each camera will go and whether you have power there.
  2. Choose wired (PoE) or wireless based on distance from the router and whether you can run Ethernet.
  3. Pick IP cameras unless you already own an analog DVR — IP gives you higher resolution, AI detection, and future flexibility.
  4. Set up the NVR or cloud account before mounting cameras to verify connectivity.
  5. Test night vision and motion detection zones before finalizing the camera angle.

FAQs

Do security cameras record all the time or only when motion happens?

That depends on how you configure them. Most systems support both continuous recording and motion-triggered recording. Continuous recording fills storage faster but guarantees you catch everything. Motion-triggered recording saves space but can miss events if the detection zone is set too narrow.

Can a security camera work without internet access?

Yes, but only if it records to local storage. Analog cameras connected to a DVR, or IP cameras writing to an NVR or microSD card, continue recording normally during an internet outage. Cloud-based systems lose remote viewing and off-site backup when the connection drops, but local recording on those cameras still works in most models.

What is the difference between a DVR and an NVR for security cameras?

A DVR works with analog cameras. It receives raw analog video over coaxial cables and digitizes the footage inside the recorder. An NVR works with IP cameras, which send a pre-digitized, compressed stream over Ethernet or Wi-Fi. The two are not interchangeable — an NVR cannot process analog video, and a DVR cannot process IP camera streams.

How much bandwidth does a 4K security camera use?

A single 4K camera streaming at 15 frames per second with H.265 compression uses roughly 8 to 12 Mbps of upload bandwidth. If you have four cameras, that is 32 to 48 Mbps — enough to saturate many residential upload connections. Lower the frame rate to 10 fps or switch to H.265+ (a variable-bitrate version) to cut the bandwidth by about half.

Do I need a special router for wireless security cameras?

Not usually, but a standard ISP router can get overloaded if you add more than 6 or 8 cameras. The bigger issue is the 2.4 GHz vs. 5 GHz compatibility — if your router broadcasts only on 5 GHz, budget cameras that only see 2.4 GHz will never connect. A dedicated access point or a mesh system with a separate 2.4 GHz SSID solves the problem.

References & Sources

Please use a real email you check. If it's fake or mistyped, your message won't reach us and we can't reply — wrong addresses are rejected automatically.