Upgrade Your NVR Without Replacing Your Cameras

You bought the house. The cameras came with it. Nobody left a password, the previous owner cannot remember which company installed them, and the app either refuses your email or opens into somebody else’s account. Meanwhile there is a box in the basement with a blinking light, quietly recording video you cannot watch.

This is one of the most common calls we get, and for years the honest answer was a disappointing one: replace the lot. That answer has changed. A new generation of recorders runs the artificial intelligence itself, on the video arriving from the cameras, instead of depending on smart cameras to do the thinking. The cameras can be ordinary. The recorder supplies the intelligence. That single shift is what makes a recorder-only upgrade worth doing.

What follows is the whole decision in order: how to tell what you actually own, whether it can be saved, what you gain, what it costs, and where it stops working. We install security camera systems across the GTA and handle a lot of these takeovers, so this is written the way we would walk you through it standing in your driveway.

First, are your cameras digital or analog?

Go and look at the back of the recorder — this one detail decides everything else

Everything in this article applies to digital IP cameras and to nothing else. An analog camera cannot be improved by changing the recorder, because a network recorder has no socket an analog camera can plug into. Check this before you read another word: if your cameras are analog, the rest of this guide is not your answer.

Walk to the recorder and look at the back panel, where the camera cables land.

If the back panel is ambiguous, there is a second test that is even more reliable, and you can run it at the camera end. Count the connectors at each camera and look for power bricks.

An analog camera has two connections: a fat coaxial cable carrying video, and a separate thin pair carrying power. Somewhere near the recorder there will be a power supply brick, or a metal box with a fused output for each camera. A digital PoE camera has one connection: a single network cable roughly five millimetres thick, ending in a clear plastic plug with eight copper pins visible through it. Power and video travel down that one cable together.

So the fastest question a homeowner can answer is this: does each camera need its own power supply? If yes, the system is analog. If the cameras have no power source other than the recorder itself, it is digital, and you are reading the right article.

Two further clues, if you want certainty. The recorder’s own label is usually explicit — “NVR” or “Network Video Recorder” means digital, while “DVR”, “XVR”, “Turbo HD” or “Penta-brid” means analog or hybrid. And the sticker on the camera arm often settles it by model prefix: on Hikvision hardware, DS-2CD is a network camera while DS-2CE is analog; on Dahua, IPC- is network and HAC- is analog.

If it turns out you have analog cameras

You have three options, and we would rather say this plainly than sell you a recorder that cannot help you. Replacing the cameras is usually the right call, because a system old enough to be analog is generally old enough to have other problems too: hazed lenses, weak infrared, sun-damaged cable. A video encoder can digitise analog cameras so a modern recorder will accept them, but it cannot invent detail the camera never captured, and on a four-camera house an encoder tends to cost about what new cameras would. A hybrid recorder that takes both analog and digital inputs is the cheapest upgrade and buys you a modern app, but the AI story is much weaker and the cameras remain the ceiling.

The one genuinely good case for a hybrid recorder is when the analog cameras are recent high-definition-over-coax units with clean lenses and working infrared, and your real complaint is the recorder and the app rather than the picture.

You inherited the cameras but not the passwords. What now?

Resetting equipment you own is routine maintenance, not a workaround

Every manufacturer has an official procedure for exactly this situation, because it comes up constantly — houses change hands, installers go out of business, people forget passwords. The previous owner’s credentials are not protecting anybody at this point. The hardware conveyed with the house and it is yours to reset.

There is a more urgent reason to do it than convenience. An inherited camera may still be enrolled in someone else’s cloud account, still sending push notifications to a phone that is not yours, still emailing snapshots to an address you have never seen, or still uploading to a server somewhere. Until every device has been reset and re-credentialled, you do not actually know who is receiving video of your family. That is the part people underestimate, and it is why we treat this as the first job rather than a loose end to clear up later.

There are four legitimate routes back in:

• The physical reset button. Most IP cameras have a recessed button, commonly under a rubber cover on the cable pigtail or beneath the memory card cover inside the housing. Held while the camera is powered, it restores factory settings and clears the password. This is the fastest route and it covers most inherited cameras. It does require reaching the camera, which is why a soffit-mounted unit at three metres in February is a ladder job rather than a five-minute job.
• The manufacturer’s recovery flow. Some systems email a reset code to an address registered when the device was first set up. For a new homeowner that address usually belongs to the previous owner, so this route tends to fail and fall back to one of the others.
• A manufacturer-issued reset file. Hikvision and Dahua both operate procedures where an installer exports a device code and their support team returns an encrypted reset file, subject to proof of ownership. These files are time-limited and the device must stay powered between export and reset.
• The camera’s own web interface, where credentials survived for one device but not the rest.

A professional should verify ownership before doing any of this, and any installer worth hiring will. The standard is straightforward: reset devices only at the address where they are mounted, only for the person who can demonstrate they own that address, and write down what was done. We document every device by location, model and serial before touching it, and that document becomes yours at handover.

Can you really keep the cameras and replace only the recorder?

Yes, and on a wired house it is usually the sensible move. The cameras are the part that was expensive to install, because the cost was never really the camera — it was getting cable through a finished wall to a spot three metres up under a soffit. That work is done. It does not need doing again.

The recorder, meanwhile, is the part that ages badly. It holds the software, the app, the detection logic and the storage. A recorder from 2016 gives you a clumsy app, a timeline you have to scrub by hand, and motion alerts that fire at every passing car. The cameras attached to it may be perfectly serviceable 4-megapixel units that still produce a clean image.

Swapping the recorder resets the entire experience of using the system while leaving the expensive physical work untouched. What you should not expect is for it to fix a bad camera. A recorder cannot add detail that was never captured, cannot see through a hazed dome, and cannot brighten a scene the infrared never reached. It changes what happens to the video after the camera produces it.

Will cameras from one brand work with a recorder from another?

Usually for video, often not for anything else

Two standards make cross-brand recording possible. RTSP is essentially a remote control for a video stream: if you know a camera’s address, credentials and exact stream path, you can pull video from it. ONVIF is the broader common language that lets a recorder discover a camera, authenticate, configure it and receive events, without needing to know the manufacturer’s private protocol.

ONVIF is split into profiles, and the split is the whole story:

Two facts about ONVIF explain nearly every disappointment people have with it. The first is that conformance is self-declared — manufacturers run the test tool themselves and submit their own declaration, with no independent lab involved. That is the structural reason “ONVIF compatible” printed on a box and “works with your recorder” are different claims. The second is that conformance attaches to a specific firmware version, so a model appearing in the database does not mean the unit on your wall qualifies.

Some manufacturers are refreshingly blunt about the consequences. Reolink states plainly that “Reolink cameras are the only guaranteed cameras compatible with Reolink NVRs” and that compatibility otherwise “cannot be guaranteed.” Ubiquiti documents that on third-party cameras adopted into UniFi Protect, “Motion detections, PTZ control, and Audio are supported… Other advanced features are not supported.” Both are telling you the same thing: the picture crosses brands, the intelligence generally does not.

So the honest expectation is this. A camera getting a usable picture into your new recorder is likely. That camera keeping every feature it had on its old recorder is not.

Where it goes wrong, and what it costs to fix

Five problems account for most of the difficulty, and none of them are visible before the work starts:

• The cameras are hiding on the old recorder’s private network. If the old recorder had built-in PoE ports, it almost certainly put the cameras on an isolated internal address range of its own. Unplug it and those cameras hold addresses that exist on no network at all. They are powered, alive and invisible. This is the single most common reason a simple recorder swap turns into a service call.
• Unknown fixed addresses set by an installer on a scheme nobody wrote down — same symptom, different cause.
• ONVIF switched off, or needing its own separate account. On several platforms ONVIF is disabled by default and needs a dedicated user created for it. Without that, the camera answers pings, serves its own web page, and still refuses the recorder.
• Proprietary codec modes. Manufacturers layer their own compression variants on top of the standards, and a third-party recorder may decode them badly or not at all. The fix is usually free: log into the camera and set it back to standard H.264 or H.265 before adopting it.
• The old recorder was the power supply. If the cameras were powered from its built-in PoE ports, removing it removes the power for every camera. The replacement needs at least the same number of powered ports, or a separate switch, with the power budget checked against the worst case rather than the idle case. A camera with a heater will run perfectly in September and drop out on the first cold night.

How a recorder adds AI detection to cameras that never had it

The distinction that matters is where the analysis happens. Edge analytics run on a processor inside the camera, on the raw sensor image. Server-side analytics run on the recorder, on a stream the camera has already compressed and sent across the network. Edge is technically the better place to do it, because nothing has been thrown away yet. But edge analytics require smart cameras — and if you had smart cameras, you would not be reading this.

Server-side analysis is what makes an existing set of cameras worth keeping. The recorder takes in ordinary video and applies its own detection to it, so a camera from 2018 that only ever knew how to report “pixels changed” starts producing “a person entered the driveway.”

Here is the part to check before buying anything, because it is where most disappointment comes from: server-side AI is a metered resource, and it is almost always smaller than the channel count. A sixteen-channel recorder does not mean sixteen channels of AI. Read the AI channel budget in the datasheet and match it to the cameras that actually matter — the driveway and the two door approaches, not the side-yard camera pointed at a fence.

Equally important, many recorders labelled “AI” only relay detection events the camera generated. On a dumb camera there are no such events to relay, so you get nothing. The capability you need is the recorder performing detection on third-party streams, stated by name in the specification.

Which of your existing cameras are actually worth keeping

The useful filter is not brand and not megapixels. It is how many pixels land on a person, in the area you care about, at night. A 4-megapixel camera mounted at 2.8 metres covering a ten-metre driveway is a keeper. A 2-megapixel camera at 4.5 metres on a wide lens covering an entire back yard is not, no matter how good the new recorder is — the person is simply too small in the frame for anything to classify them reliably.

Two practical checks go with that. Recorders frequently run their analysis on the camera’s substream rather than the main stream, and substreams are often configured far below what detection needs — a 4-megapixel camera being analysed should be running at roughly 4 Mbps in H.265 at 25 to 30 frames per second, or about 2 Mbps at 15. And daytime performance tells you nothing about night performance. We test at night, in the rain, at the far end of the zone, because that is when it matters.

Why the Ajax Superior NVR suits a house full of someone else’s cameras

This is the recorder we reach for on takeover jobs, for one specific reason: it runs its own AI on cameras it did not manufacture. Ajax states it directly — the recorders “enable basic cameras to recognize people, pets, and vehicles — with the same accuracy as native Ajax cameras,” and third-party cameras connected this way can “send detection notifications, trigger sirens, or run automation scenarios.”

That is less common than it sounds. Most recorders that advertise AI apply it only to their own brand of camera, and the major competitors say so in their own documentation. It is the difference between a recorder that makes your existing cameras smarter and one that simply records them.

The practical specification, for a house:

• Detection types: people, pets and vehicles, plus line crossing with up to four lines per camera.
• Camera compatibility: ONVIF Profile S and G, plus RTSP, decoding H.264 and H.265. It scans the local network and finds available cameras automatically.
• Storage: two hot-swappable drive bays taking up to 24 TB each. Drives can be swapped without shutting the recorder down. Hard drives are not included.
• Two recording modes: redundancy, which mirrors footage across both drives, or split, which alternates between them for capacity. Retention is settable from 1 to 360 days, or unlimited.
• Local viewing: a 4K HDMI output drives a monitor directly, with live view and archive playback that work with no internet connection at all.
• Model range: 8, 16 and 32 channels, in versions with 8 or 16 built-in PoE ports, or with two gigabit network ports and no PoE for houses that already have a switch.

The limits, stated plainly

Three constraints are worth knowing before you commit, and any installer who does not raise them is not doing the job properly.

The AI has a channel cap. On the 8-channel models, AI analytics cover all eight. On the 16- and 32-channel models, AI on third-party cameras runs on up to twelve channels. A 32-channel recorder is, for AI purposes, a 12-channel recorder. On a house this is rarely a real constraint, but it decides which model to buy on a larger property.

Turning AI on roughly halves decoding capacity. Ajax publishes two separate decode tables for this, and the numbers drop by about half once object detection is enabled on even one camera. It affects how many feeds display at once, not how many record.

It is not fully offline. Recording, playback and local viewing all work without internet, but Ajax specifies that cameras added through the recorder must reach the Ajax cloud server at least once every 30 days for continued offline operation. There is no fee attached to that, but it is not an air-gapped system and we would rather you hear it from us.

One more thing that matters practically: the Superior line is sold only through accredited Ajax partners for sale, installation and maintenance. You cannot buy one from a retailer or a marketplace. SetupTeam is an accredited Ajax partner, so we supply, install and register it — and the registration is in your name, not ours.

Cameras and alarm in one app, with no subscription wall around it

Most households end up with a drawer of apps: one for the cameras, another for the alarm, a third for the doorbell. They do not talk to each other, so an alarm at 3 a.m. means opening one app to see what tripped and another to find out what it looked like.

Because the recorder here is part of a full security platform rather than a standalone video product, the cameras and the alarm system live in the same app. An alarm from a motion detector highlights the relevant camera feed automatically, and a notification takes you straight to the footage of that moment instead of dropping you on a timeline to search. Detectors can trigger recording; cameras can trigger sirens and automation.

One app covering both is not, on its own, unusual — Ring, ADT, Vivint and SimpliSafe all manage it. What is unusual is what surrounds it. On most of those platforms the app is the hook for a subscription: Ring charges a monthly fee before you can arm or disarm your own alarm from your phone, ADT keeps your footage in the cloud with no local recording at all, and the systems that do support third-party cameras generally gate them behind their most expensive tier.

Here the same app covers the alarm and the cameras — including cameras from other manufacturers running through the recorder — with local recording and no recurring charge attached to either. That combination is genuinely hard to find in Canada. If you are already replacing the recorder, it is worth knowing the same platform extends to a fully wireless alarm system that retrofits into a finished house without opening walls.

And if you want someone watching it around the clock

Everything above is self-monitored: alerts come to your phone, footage stays on your drives, and nothing recurring is charged. Some households want more than that — a professional monitoring station that receives alarms day and night and dispatches a response when you are asleep, away or simply unable to answer the phone.

That service does carry a monthly charge, and it is the one recurring cost we think is genuinely worth considering. It is worth being precise about the difference: you never pay monthly to record or watch your own video. A monitoring subscription buys a staffed control room responding to alarm events, which is a different product entirely from paying a camera manufacturer to keep your footage on their servers.

Local recording means no monthly fee, and no account to inherit

A recorder with its own hard drives stores footage in your house. There is no recurring charge to record, no recurring charge to watch, and no company holding your video on their servers under terms they can change. This is the structural difference between a recorder and a cloud camera, and over five years it is not a small one.

Here is what the subscription route actually costs in Canada, taken from the manufacturers’ own Canadian pricing pages:

Those figures assume prices hold flat for five years, which they have not historically done — several of these plans have risen in the last eighteen months, so treat the totals as conservative. Against a multi-camera plan, a local recorder pays for itself somewhere between the first and third year and then costs nothing further.

One honest caveat in the other direction: if you have exactly one camera, the arithmetic flips. A single-camera cloud plan at about $60 a year takes many years to justify a recorder, and we will tell you so rather than sell you hardware you do not need.

There is also a non-financial argument, which several of our customers care about more than the money. Footage on a drive in your basement keeps working when the internet drops, is not subject to a vendor changing their policy or discontinuing a product line, and is not sitting on someone else’s servers. The counterpoint is equally real: a recorder can be stolen along with everything else, which is why we place them out of sight and, on higher-risk properties, mirror the footage across both drives.

How much footage will you actually keep?

Retention depends on camera count, resolution and whether you record continuously or only on detected events. The arithmetic is simple enough to do before buying drives: multiply each camera’s bitrate by the hours, add them up, and divide the drive capacity by the daily total.

A worked example for eight cameras recording continuously:

Recording only on detected events extends those numbers considerably, but by how much depends entirely on the scene. A camera facing a quiet back yard and one facing a busy sidewalk differ by an order of magnitude, so we would rather size the drive generously than promise you a number we cannot stand behind. Retention also needs verifying after installation by going back and checking how far the archive actually reaches, rather than trusting the calculation.

Use surveillance drives, and do not let anyone talk you out of it

This is the corner most often cut, and the specifications make the case better than any argument. A standard desktop drive such as a Seagate BarraCuda is rated by its own datasheet for 2,400 power-on hours per year and a 55 TB annual workload. A surveillance drive such as a WD Purple or Seagate SkyHawk is rated for 8,760 hours — every hour of the year — and 180 TB.

Now compare that to the load. Eight cameras recording continuously at 4 Mbps write about 126 TB per year. That is more than double a desktop drive’s rated annual workload, on a drive only rated to be switched on for 27% of the year. Even four cameras at 4 Mbps exceed it. A desktop drive in a recorder is not a bargain, it is a scheduled failure — and it fails silently, so you find out when you go looking for footage that was never written.

What a recorder upgrade costs, and what the work involves

There are two parts to the number, and we quote them separately so you can see what you are paying for.

Equipment is quoted per project. It depends on channel count, whether you need built-in PoE ports, and how much storage you want — drives are a meaningful part of the cost and are sized to the retention you actually want rather than to a round number.

Installation and setup starts at $150 per camera, as a one-time fee. That figure moves with the specifics: how high the camera is mounted, how awkward the position is to reach and work at, and how many cameras are in the project. A recorder swap on cameras that are already wired, reachable and healthy sits at the lower end of that range, because nobody is pulling new cable. A camera at the peak of a three-storey elevation, or one that needs re-aiming and re-terminating, costs more — and we would rather price that honestly on site than quote you a flat number and revise it afterwards.

What the setup fee covers is the work that makes the system actually yours: auditing and documenting every camera, testing each cable run, resetting and re-credentialling every device, putting the cameras on their own isolated network segment, updating firmware, configuring detection zones per camera, setting retention, and handing you the admin account.

What we do when we take over an orphaned camera system

This is the sequence, and each step exists because skipping it causes a specific problem later.

1. Audit and document everything first. Location, mounting height, field of view, make, model, serial and current address for every camera, with a still frame from each. You cannot plan a migration against an inventory you do not have, and this document becomes yours.
2. Test every cable run individually. A marginal run works perfectly with an old 2-megapixel camera and fails intermittently the moment something higher-resolution goes on it. Finding that before the new recorder arrives is the difference between one visit and three.
3. Reset and re-credential every device, with ownership verified and recorded. This is the step that severs any remaining connection to the previous owner.
4. Put the cameras on their own network segment. Cameras are long-lived, rarely patched devices running a web server, physically reachable from outside the house. Canada’s Cyber Centre recommends segregating this class of device from the main network, and cameras are the textbook case.
5. Disable UPnP and do not port-forward the recorder. Remote access should go through the manufacturer’s encrypted relay or a VPN, never an open port pointed at a recorder’s web interface. The Cyber Centre’s guidance on UPnP is to disable it entirely.
6. Change every default password and update firmware, recording the versions. Camera vulnerabilities are not theoretical — one well-known flaw affecting a wide range of recorders and cameras gave remote attackers full control, and unpatched devices with default passwords remain a genuine route into a home network.
7. Configure detection zones and schedules per camera, excluding the road, the neighbour’s windows and the tree that moves in the wind. This is where the AI purchase is either realised or wasted.
8. Hand over the admin account to you, in writing, along with the device inventory, retention setting, firmware versions and network details. We keep a separate service account or none at all.

That last point is the whole reason this article exists. The situation you are in — locked out of cameras on your own house — was created by an installer who kept the admin account. We are not going to create the next one.

One note on audio, because it is the thing people get wrong

Video of your own property is straightforward. Audio is not. Canada’s rules on recording conversations turn on the consent of someone actually taking part in the conversation — owning the microphone does not give you standing to consent on others’ behalf. Recording a conversation you are part of, such as speaking to a delivery driver through a doorbell, is a different situation from a camera quietly capturing two neighbours talking on the sidewalk. Our default is to leave audio disabled unless there is a specific reason to enable it. This is general information rather than legal advice, and worth a conversation with a lawyer if your situation is unusual.

When the cameras should be replaced instead

Sometimes the answer is that the cameras are not worth keeping, and we would rather say so before you spend money on a recorder. Any one of these is a strong argument for replacement; two or more makes keeping them a false economy, because the labour to remount is the same either way and labour is most of the bill.

• Standard-definition or 720p cameras. There are not enough pixels landing on a person at a realistic distance for either you or the software to work with.
• Hazed, crazed or yellowed dome covers. Polycarbonate degrades under UV, producing a permanent soft-focus veil that becomes a total white-out at night as infrared scatters inside the dome. Worth checking whether replacement bubbles are available first — on some professional domes they are.
• Cracked, sun-rotted or animal-chewed cable, or corroded outdoor terminations. A run that tests marginal in September fails in January.
• Dead or dim infrared. Infrared emitters degrade. A camera whose night image has gone dark at range is an expensive daytime-only camera.
• End-of-life firmware. If the manufacturer has stopped issuing updates, the camera is permanently vulnerable and will never gain the ONVIF support a new recorder wants.
• Locked to a dead ecosystem. Kit cameras that only work with their original recorder, or cloud-bound cameras with no local protocol. Keeping these means keeping the old recorder, which defeats the project.

If you are replacing cameras anyway, the calculus changes and it is worth reading our take on when UniFi Protect is the right choice — for a system built from scratch, where you are choosing the cameras as well as the recorder, it is often the better platform. This article is about the other situation: the cameras are already on the wall and staying there. It is also worth reading what Ring cameras do and do not give you compared with a wired system, if a subscription camera is the alternative you are weighing.