A pan-tilt-zoom camera reads like the answer to a coverage problem on the spec sheet. One device, a wide rotation range, optical zoom, and a price that is only modestly higher than a fixed camera. What the spec sheet does not say is that a PTZ is a camera with one field of view that happens to be movable, and every moment you did not point it somewhere is a moment you did not record.
The coverage you buy is not the coverage you get
A fixed camera records a fixed cone, all the time. A PTZ records one cone, wherever it happens to be pointing. If it is parked on the driveway because that is where you care most, then the side gate is not being recorded while it sits there. If it is running a tour, then each stop gets a slice of time and the gaps between stops are blind.
This is the trade nobody prices. Coverage in PTZ marketing means the set of things the camera is capable of seeing. Coverage in practice means what it was looking at during the two minutes that mattered. Those two numbers differ by roughly the number of preset positions you have.
Small angle maths explains why this bites harder than expected. A camera with a 90 degree horizontal field of view parked in one position covers a quarter of the circle around it. Rotate to a second preset and you have covered a second quarter, but not at the same time. Two PTZ presets do not add up to one 180 degree camera. They add up to two cameras that are each only awake half the time.
Why the playback does not help
After something happens, the first thing anyone does is scrub the timeline. With a PTZ this is where the confusion starts, because most recorders show you video, not state. You see the driveway, then a cut, then the side gate, and if the camera moved automatically there is no annotation telling you why.
Two failure modes follow from this. The first is an event that happened while the camera was mid-travel or pointed elsewhere, which produces no footage at all and looks like a recording gap. The second is an event that happened at the edge of the frame while the camera was swinging toward it, which produces motion-blurred footage of the thing you needed and nothing before or after it.
The fix that actually works is boring: know where the camera was, and why. Set a home position and a return time, so the camera always comes back to the one view you care about after a tour or after manual control. Log manual moves if the recorder supports it. Then, when you scrub, you can reconstruct the state instead of guessing.
Tours are a compromise, and the compromise has a number
A tour cycles through presets with a dwell time on each. If you have five presets and a thirty second dwell, each position is observed for thirty seconds out of every two and a half minutes. That is twenty percent duty per position. Any event shorter than the gap can occur entirely unobserved.
This is fine for deterrence and for general awareness. It is not adequate for anything you would need to prove. The honest way to use a tour is as a supplement to fixed cameras that cover the critical views continuously, not as the primary recorder for them.
Which brings up the pairing that works: one fixed wide camera per critical approach, recording continuously, plus one PTZ for the thing a fixed camera cannot do, which is read a plate, follow a person, or look at something specific after being told to. The fixed camera tells you something happened and where. The PTZ then goes and looks. Neither replaces the other, and a PTZ used without the fixed camera is a camera with a blind spot that moves.
Zoom has the same problem one level down
Optical zoom is the part of the specification that survives scrutiny, and it is genuinely useful. Digital zoom is a crop, and a crop of a 4 MP sensor at 8x is a low resolution image that people describe as blurry and blame on the camera.
The trap is that zoom narrows the field of view at exactly the moment you need context. A plate is readable; the fact that the person walked from the neighbour's gate is not in frame any more. Zooming is a decision to trade context for detail, and it should be made after the wide shot is already recorded, not instead of it.
Latency and the moving-target problem
A PTZ that tracks motion is only as good as the delay between detection and movement. Mechanical travel takes real time, and a fast-moving subject can be out of the frame by the time the camera arrives. Tracking is also where false positives become expensive, because every false detection is a camera pointed at nothing while it is not pointed at the thing that matters.
If you want tracking to be useful rather than decorative, the detection source should be better than the camera's own pixel analysis. A fixed camera, a door sensor, or a PIR that defines where to look gives a much better result than asking the PTZ to notice things while it is also moving.
What to ask before buying
Ask how many presets the camera supports and whether there is a home position with an automatic return timeout. Ask whether the recorder logs PTZ state alongside video, because without it you will be guessing forever. Ask whether the tour dwell time is configurable per position, since a gate and a driveway rarely deserve equal time.
Then ask the question that saves the most money: what is this camera going to watch continuously? If the answer is everything, the honest answer is that a PTZ camera is the second camera you buy, not the first. The first one is a fixed camera pointed at the one approach you cannot afford to miss.
Any camera manufacturer will confirm that the installations people are happy with two years later are the ones where the moving camera has a defined job. A PTZ with a job is excellent. A PTZ bought to avoid deciding where to point a camera is expensive and will disappoint you precisely once, at the worst possible time, and the fixed models in the same range are what make the moving one useful.
Top comments (0)