Two cameras can carry the same number on the box and behave nothing alike. Megapixels are cheap to print and expensive to actually deliver, so the budget end of the market competes on the specification you can read and economises on everything you cannot.
None of this matters until the night something happens. Then it matters completely, because the whole point of the system is one clip you may need only once.
The sensor is the part you are really buying
Resolution tells you how many pixels there are. The sensor decides how good each one is. Budget cameras use physically smaller sensors, which gather less light, and the result is a picture that looks acceptable at noon and falls apart at dusk — soft, grainy, with colour that smears as the light drops.
This is why a cheap camera can be sold as high resolution and still fail the only test that counts. You can have plenty of pixels and still not have a face, because the pixels are noisy rather than detailed.
Night vision is where the corners get cut
Infrared illumination costs money and draws power. Cheap cameras skimp on both, and it shows in two ways: light that falls off long before the distance the camera is supposedly covering, and a hot spot in the middle with darkness at the edges.
There is a subtler failure too. Infrared LEDs draw considerably more current than daytime operation, and where the supply was specified to the bare minimum, the camera can dim, drop out or reboot exactly when the IR kicks in. A camera that works perfectly all day and misbehaves after dark is very often a power problem, not a camera fault.
- IR range quoted for the lab, not your driveway
- Uneven illumination — bright centre, dark edges
- Undersized power supply that sags when IR switches on
- Washed-out faces from IR bouncing off a dome cover
Lenses and housings
Cheap optics use lower-grade plastic elements, which introduce distortion and softness towards the edges of the frame — the exact part of the image where an approach path usually sits.
Housings tell the same story. A budget enclosure survives its first year, then the gasket hardens, the plastic chalks in the sun, and water finds its way in. On the Kerala coast this is accelerated: salt-laden air attacks cheap mounts and screws, and monsoon rain finds any weakness in a seal. We have replaced two-year-old cameras that failed for no reason other than the housing they came in.
The costs that do not appear on the quote
A cheap system is rarely cheap over its life. Cameras get replaced sooner. Someone has to climb up and do it, and that labour is not free. Firmware stops being updated, so app support quietly degrades. And when a unit fails, matching it becomes a hunt, because the model has already been superseded.
The comparison that actually matters is not price per camera, it is cost over five years including the labour to keep it working — and, more importantly, whether the footage is usable when you finally need it.
- Earlier replacement, plus the labour to do it
- Firmware and app support that fades away
- Models discontinued, so repairs mean mismatched kit
- Footage that exists but cannot identify anyone
Where it is fine to save
Not every position needs the best camera available. General observation views — a wide angle over a compound, a corridor, a store room — do a useful job with mid-range hardware, and spending heavily there is waste.
Put the money where identification happens: the gate, the main door, the billing counter. A sensible system is deliberately uneven, with quality concentrated at the two or three positions that will produce the clip you actually hand to someone.
The short version
Judge a camera by its sensor, its lens, its real infrared range and its housing — not by the number on the box. Spend where identification matters and economise where you are only observing. And ask any installer to show you a night image from the actual distance involved, because that single test separates the two ends of this market faster than any specification sheet.