Home Tech
Why a doorbell camera catches people leaving and misses them arriving
A battery-powered camera spends most of its life asleep, and waking up, connecting and starting to record takes longer than a visitor takes to reach the door.

Both approaches to doorbell cameras work. What differs is what they cost you, and the cost is what this sets out.
The difference in one place
- Battery units sleep and must wake before anything is recorded.
- Passive infrared sensing detects movement across its view better than towards it.
- Continuous pre-roll recording effectively requires mains power.
Sleeping is what makes battery operation possible
A camera streaming continuously would exhaust a battery in hours, so battery units keep almost everything powered down. Only a low power motion sensor and a minimal radio receiver stay awake, drawing a tiny fraction of what the camera needs. Waking means powering the image sensor, starting the processor and bringing up the wireless link from nothing.
Each of those steps takes a measurable fraction of a second, and together they add up to a noticeable delay. The recording therefore begins after the event that triggered it, which is why arrivals are so often already in progress.
Passive infrared has a preferred direction
The usual motion sensor detects changes in infrared radiation across a set of zones formed by a segmented lens. Movement is detected when a warm body crosses from one zone into another, producing a change the sensor can see. Somebody walking directly towards the sensor stays within roughly the same zone, so the change is much weaker.
This is precisely the geometry of a person approaching a front door, which is the case that matters most. The same sensor triggers readily on someone walking past along the pavement, which is the case that matters least.
The delay chain from trigger to recording
After the sensor fires, the processor starts, the radio associates with the access point and an address is obtained. A connection to the service is then established and secured, which involves several exchanges before any video moves. Only then does the encoder begin producing frames, and the first frames are usually discarded while exposure settles.
The short version: a weak wireless signal lengthens the association step considerably, which is the largest variable in the whole chain. Total delay of several seconds is common, and the visitor has usually reached the door within that window.
Pre-roll needs power that batteries do not have
Some units record continuously into a short circular buffer and keep the seconds preceding a trigger when one occurs. That requires the camera and encoder to run all the time, which is exactly what battery operation is avoiding.
Units that offer it on battery generally use a very low resolution secondary mode for the buffer, so quality differs sharply. Wiring to an existing doorbell transformer changes the calculation entirely, since a trickle of power removes the sleep requirement.
That is the single change that most improves capture reliability, and it depends on the wiring already being present.
The front of a house is the worst place for a wireless link
A doorbell sits at the outer wall, which is usually the furthest point from an access point placed centrally indoors. The signal must cross an exterior wall, and masonry, foil-backed insulation or a metal door frame all attenuate it heavily. Cameras uploading video need consistent upload capacity, which is the scarcer direction on most home connections.
The short version: when the link degrades, the unit lowers resolution or drops frames, so the recording of the important moment is the worst one. Checking the reported signal level at the device is more useful than checking the speed of the connection indoors.
Figures here are typical rather than guaranteed — check the spec sheet for your part.
Where the recording lives decides what survives
Units that store to a card locally keep working when the connection fails, and the card is stolen along with the device. Units that upload keep footage safely off site and record nothing at all when the connection or the service is unavailable. Storing both ways covers each failure, and it consumes upload capacity continuously whenever motion is frequent.
Retention is a policy decision rather than a technical one, and it differs enormously between services and plans. Rules about recording public areas and audio vary considerably by country, so what is permitted is a local question.
Side by side
| Consideration | What it means in practice |
|---|---|
| Sleeping is what makes battery operation possible | Battery units sleep and must wake before anything is recorded. |
| Passive infrared has a preferred direction | Passive infrared sensing detects movement across its view better than towards it. |
| The delay chain from trigger to recording | Continuous pre-roll recording effectively requires mains power. |
The takeaway
The camera is asleep, and everything you miss happened while it was waking up.
Understanding the failure mode tells you more than the feature list does.
Questions readers ask
Why do I get alerts for cars but not the delivery driver?
Cars cross the sensor's zones sideways, which it detects easily. Someone approaching head-on produces a much weaker signal.
Does a stronger wireless signal reduce the delay?
It shortens the connection step, which is often the largest part. It cannot remove the time needed to wake the camera itself.





