Back to Blog

Will There Be Footage of the Theft, or Only Clips? What a Camera Records While You Are Away

2026-09-087 min readSolaGuard Team
AOV vs PIRevidence and retentionunattended property

TL;DR

PIR cameras leave a folder of short clips with blanks between them; AOV cameras leave a continuous timeline you can scrub. And a second condition nobody checks: card retention has to outlast your absence, or the night in question is already overwritten. Includes a three-night test you can run on a sample unit.

The short answer: motion cameras leave gaps, continuous recording does not

Come back after a week away, find something missing, and what you get from the camera was decided the day you bought it. A PIR camera records only when its sensor detects moving heat inside a limited cone, so what you inherit is a folder of short clips with blank space between them — and no way to know what happened in the blanks. A camera running AOV (Always-On Video, low-power continuous recording; an industry-generic term, not a brand) writes twenty-four hours a day to a scrubbable local timeline, so you can jump to 02:14 on Tuesday and look, whether or not anything triggered an alert. There is a second condition almost nobody checks before leaving: the card has to hold more days than you were away. If retention is shorter than your absence, the night in question has already been overwritten, and the recording mode stops mattering entirely.

Two different meanings of the word "recording"

PIR camera AOV camera
What lands on the card Event-triggered clips only Continuous video, always
Between events Nothing Recording
Pre-event footage Absent, or one to a few seconds depending on model Present, roll back as far as the card holds
How you search A list of clips by timestamp A timeline you scrub
If the event never triggers No record exists Still recorded
Power profile Very low idle, a spike per trigger Low and steady
Main risk Gaps The card cycling before you review

Both modes are legitimate, and both have their place. What is not legitimate is selling "24/7" and delivering motion detection. It is the most common mis-sell in this category, and it produces the worst conversation in it: someone discovering after the fact that their camera never recorded continuously at all.

The four moments a motion clip usually misses

Moment What is lost Why it matters later
The approach The vehicle parking, how many got out, where they looked first This is the identifiable part; inside the boundary people are usually covered
The first second Sensor wake and focus eat the start of the clip Face and plate tend to live exactly there
Anything outside the sensor cone Activity at the edge of the visible frame You see an empty scene and conclude nothing happened
The exit What they left with, and which way Decides whether there was a second person, a support vehicle, or a return trip

There is a fifth case that only shows up on properties left alone: the walk-through beforehand. Someone comes two or three nights earlier, follows the fence, touches nothing. On PIR that may trigger nothing at all, or produce a clip nobody watches. On a continuous record, scrubbing backwards from the night of the loss makes the pattern visible — and that pattern is often what turns a generic report into an actual line of enquiry.

Retention: the arithmetic to do before you leave

Continuous recording is a ring buffer. When the card fills, the oldest footage is overwritten. Evidence therefore has an expiry date, and that date is set by card capacity, resolution and the bitrate the camera writes at.

Standard configuration is microSD up to 128 GB — larger only on custom orders — and the retention in days has to be asked for in writing, for the exact model and resolution you are buying. No honest supplier gives one number valid across a whole catalogue, because there isn't one.

The rule is simple:

Card retention ≥ length of absence + time to notice.

The second term is the one people forget. You rarely spot a loss while unloading the car. You spot it the next morning, in the shed, counting stock. If you are away nine days and the card covers seven, the relevant night is gone before you go looking for it.

Keeping the footage from expiring

  1. Review remotely halfway through. Opening the app mid-absence does not extend retention, but it lets you pull down a suspicious segment before it cycles out.
  2. Spend your card budget where evidence is worth most. A quiet internal yard does not need the same detail as the vehicle gate. Lowering bitrate at low-value points buys days at the points that matter.
  3. Download on any signal at all. A gate left open, a padlock in a different position, tyre marks on the verge: that is enough reason to save the last few nights that same day.
  4. Confirm overwrite behaviour. Check that the camera cycles cleanly and keeps recording rather than stopping when the card is full.

Test it yourself before you commit to a quantity

One sample unit and three nights settles the argument without relying on a datasheet.

  • Night one — the gap audit. Leave it running and next morning ask to view 01:00 to 04:00 as one stretch. If the system hands you a list of clips instead of a continuous span, it is not continuous recording.
  • Night two — the approach test. Walk slowly and straight at the camera from the boundary. Then check whether the whole walk was captured or only the last few metres.
  • Night three — the edge test. Move along the outer edge of the visible frame rather than through the middle. Check whether any record exists. This is where promises and behaviour diverge most often.

Then measure real retention: on day four of operation, check how far back footage is still available. That number, measured on your site at your resolution, beats any table.

What you will be asked for afterwards

Insurers and police do not ask for "a video". They ask for continuity and time. Concretely: timestamped footage showing the state of the asset before the incident, the incident, and the state after, with no unexplained cuts in the middle. A folder of six nine-second clips separated by twenty-minute holes invites the question nobody wants: what happened during those twenty minutes? A continuous record simply does not raise it.

The same logic applies to disputes that are not thefts at all — a delivery that supposedly never arrived, damage to a fence blamed on a contractor, a subcontractor's timesheet that does not match the site. Those are settled by a timeline, not by clips.

The power objection, honestly

The classic argument against recording all the time is battery drain. It held when continuous meant sensor, encoder and radio at full power. AOV changed that arithmetic: the camera records continuously at a low standby draw and uses AI scene analysis to decide how much detail to store moment by moment — more while something is happening, less while the scene is still. And the continuous file lives on the card, not in the cloud: the 4G link carries alerts and the segment you ask for, not the whole day.

It is also worth stating the other side of the ledger. A PIR camera on a site with tall grass, loose animals or branches in the wind can fire dozens of times a night, and every wake costs energy. A low steady draw is not automatically worse than many unpredictable spikes.

Which mode belongs where

Position Recommended mode
Gate, vehicle access Continuous, highest resolution available at that point
Store, workshop, plant room Continuous, with two-way audio if deterrence matters
Long perimeter fence Continuous on stretches reachable from a road; event-based where no vehicle can get near
Low-traffic interiors Event-based, to preserve retention elsewhere

The SolaGuard catalogue is 150 SKUs, wholesale prices quoted per order, 1080p–2K as standard — 4K only as a custom or OEM build — IP66 on outdoor models, and five to ten minutes to install per camera. Models by use case are at /en/products/, and the technical detail on continuous mode is at /en/guides/aov-solar-camera-guide/.

Four questions before you pay

Ask them, and get the answers in writing for the exact SKU: does it record continuously or on events; how many days of retention on the card at the resolution I will actually use; how many days does it run under overcast skies; and what happens when the battery drops below threshold — does it stop recording or reduce quality? A manufacturer answers all four. SolaGuard builds with Shenzhen Leksell, TÜV Rheinland certified and established in 2012, plus CE, FCC and RoHS; samples in 7–10 days by air, volume in 25–40 days by sea, MOQ 10 for a trial and 100 for OEM/ODM.

To get written retention figures for a specific model: WhatsApp +86 188 1869 9870 or info@solaguard.net.

Interested in SolaGuard cameras?

150+ SKUs, dealer orders from 10 units · OEM from 100, 24-month warranty.

Get Quote

B2B terms

  • Dealer margin 20% from ~10 units · 30% from ~50 units
  • OEM / private label / custom spec: MOQ 100 units
  • 24-month warranty · Samples in 7–10 days by air

Browse the matching models

Solar 4G cameras →

Topic overview: AOV Solar Camera: Complete Buyer's Guide — Always-On Video Technology Explained [2026] →

Interested in SolaGuard Cameras?

Contact us for B2B pricing and product samples.

WhatsAppGet quote