There is a moment in most camera upgrades where somebody points at the ceiling and says the walls have to come open. It is usually the most expensive sentence in the quote, and it is sometimes true.
The building has an older camera system. Each camera runs on coaxial cable — the round, single-conductor cable with a screw-on connector — back to a box in a closet. The cameras are tired, the pictures are soft, and the owner wants better. The proposal that comes back prices new cameras, a new recorder, and new cable to every position, and the cable line is often the largest one.
It is also the line that deserves the most scrutiny, because the coax already in your walls has real remaining value — and a ceiling on that value. Both halves matter. An installer who tells you to rip it all out has not looked. An installer who tells you to reuse everything has not asked what you want the system to do in five years.
This article is about how to tell the difference.
The recorder is the part that names the system
Start with the two acronyms, because they are the clearest way into the whole subject.
A DVR is a digital video recorder. It expects analogue cameras. Each camera sends a video signal down its own coaxial cable to the DVR, and the DVR does the digitizing, encoding, and writing to disk. The intelligence lives in the box. The camera at the end of the cable is a lens and a sensor that produces a picture and pushes it down a wire.
An NVR is a network video recorder. It expects network cameras — IP cameras. Each camera is a small computer that digitizes and encodes its own image and sends it as data over a network cable to a switch, and from the switch to the recorder. The intelligence lives at the camera. The NVR receives streams, stores them, and manages them; it is a storage and management appliance rather than a video processor.
That difference — where the picture becomes data — decides nearly everything else, including what the cable in your wall is being asked to carry.
What the cable actually carries
On a traditional DVR system, the coax carries an analogue video signal, and only that. One camera, one cable, one channel, one direction. Power for the camera comes from somewhere else: either a local outlet at the camera, or a second pair of conductors bonded alongside the coax in what installers call siamese cable. There is no data path in either direction beyond the picture. The camera cannot be addressed, configured over the wire, updated, or asked a question. It transmits, and that is all.
On an NVR system, the network cable carries data in both directions, plus power. That single change produces every capability owners associate with a modern system: the camera can be configured remotely, firmware-updated, and made to run analytics on its own processor so it sends an event rather than just a picture. It can carry audio both ways. And it can be powered from the switch through Power over Ethernet, so the camera position needs no electrician and no outlet, and the whole fleet rides through a power blip on one battery backup in the closet.
The honest summary: coax carries a picture. Network cable carries a conversation.
HD-over-coax is a real middle path, not a consolation prize
Here is the part that is left out of a lot of quotes.
Somewhere in the last decade the industry produced a family of formats — usually sold under the umbrella "HD-over-coax" — that push high-definition analogue video down ordinary coaxial cable. A compatible camera at one end, a compatible recorder at the other, and the cable in between stays exactly where it is.
This is not a marketing veneer on old technology. It genuinely lifts image quality well beyond what the old analogue standard could carry, over the cable you already own, with no walls opened. For a building where the runs are sound and the pathway is difficult — poured concrete, a heritage facade, a finished ceiling in a tenanted suite — that is a meaningful saving and a legitimate answer.
A second option sits alongside it: media converters, sometimes called Ethernet-over-coax. These are small devices that put an Ethernet network onto existing coaxial cable — one at the camera end, one at the recorder end — so you can run genuine IP cameras over the coax already in the wall. Some models also pass power through the coax, recovering part of what PoE would have given you.
Both paths are real, and both are used well by competent installers. Neither is free of consequence.
Where reuse caps you
Every reuse strategy buys you the cable and inherits its limits. These are the four that owners feel later.
Analytics. On an IP system, the camera decides what it is looking at — a person crossing a line, a vehicle stopped where vehicles should not stop, someone at a door after hours — and sends an event a human can act on while it is happening. That processing lives in the camera because the camera is a computer. An HD-over-coax camera is not; it produces a picture and the recorder does what it can with it centrally. Some analytics can be performed at the recorder, but the recorder is dividing its attention across every channel, and it never has access to what the camera saw before compression. If the reason you are upgrading is "I want to be told when something happens," this is the limit that matters most, and it is the one least likely to be raised in a quote.
Resolution headroom. The ceiling on an HD-over-coax system is set jointly by the format, the recorder, and the physical condition of the run — cable type, length, connector quality, splices, interference along the path. The ceiling is therefore a property of your building, not a number from a brochure, and not one a better camera raises later. On a network system, headroom is governed by bandwidth and storage, both of which you can add without touching the wall.
Power. PoE is quietly one of the biggest practical gains of a network system: no outlet at the camera, no electrical trade at each position, one battery backup for everything. Plain coax gives you none of that on its own. Either the old power conductors are still live and sound and you reuse them, or you power each camera locally, or you buy converters that carry power over the coax. Each of those is workable; each adds either a device or a trade at every camera position, and the cost of that has a habit of eating the cable saving quietly, one position at a time.
Two-way audio and speech. Talking through a camera to a person on your property — the thing that turns a recording into an intervention — needs a return path. A one-way analogue link does not have one. If speaker challenge is part of what you are buying, it is a network-camera capability, or it is a separate device and a separate cable.
There is one honest point in the other direction, and it is not small: a camera with no network connection cannot be attacked over a network. Network cameras are computers, and they inherit every obligation that comes with computers on your network — segmentation, password hygiene, firmware maintenance, and a plan for the day a vulnerability is published. That obligation is manageable and it is the price of the capability, but it is a real cost of the modern path.
The asset you actually own is the pathway
Here is the reframe worth taking away from this whole subject.
When people say "the coax is worth something," they usually mean the copper. Often the more valuable asset is the pathway — the conduit, the sleeves through fire separations with their fire stopping already done, the routes above the ceiling, the core drills, the mounting points with a box and a bracket already at the right height. That is the part that cost real money to create and that nobody wants to create twice.
A pathway with pull-through room is worth far more than the cable inside it, because it means new cable can replace old cable without opening anything. In that building, the argument for HD-over-coax falls away — you can simply pull network cable through the route the coax is using and get the full modern system for the labour of a pull.
Which is why the first question is never "DVR or NVR." It is: what is actually in the wall, and can anything else get through it?
How to decide, position by position
The right answer is usually mixed, and mixed is fine. Camera systems are not required to be homogeneous — hybrid recorders that accept both analogue and network inputs exist for exactly this reason, with the caveat that their channel counts are fixed at purchase, so a hybrid box makes today easy and can make expansion awkward. Decide with your eyes open.
Ask for five things before anyone quotes cable:
1. A per-run cable survey, in writing. Cable type as printed on the jacket, approximate run length, condition of the terminations, whether the run is in conduit or free-air, and whether the conduit has room for another pull. An installer who has walked the building can produce this. One who has not will estimate, and the estimate is where the surprise lives.
2. Whether the pathway can take new cable. Ask it as its own question, per run, and separate it from the cable question. This is the single most valuable answer in the survey.
3. What each camera position is for. There is a real difference between a position that needs to tell you a person is there and a position that needs to identify who. Reuse is a much easier decision on a wide overview of a yard than on the one camera covering a till, a door, or a gate.
4. What the system must do that it cannot do now. If the answer is "see better," reuse is genuinely on the table. If the answer includes alerts, analytics, remote configuration, or speaking to someone on site, say so at the start — those requirements are what make a network path worth the cable, and hearing them late is what produces a system that is replaced twice.
5. How long you will own the building. A five-year horizon and a twenty-five-year horizon justify different answers, and neither is wrong. The trap is not reuse. The trap is reusing without knowing you have set a ceiling, and then paying the rewire anyway three years later — having bought a recorder in between that the new system cannot use.
The short version
A DVR system puts the intelligence in the recorder and sends a one-way picture down coax. An NVR system puts the intelligence in the camera and sends a two-way conversation, plus power, down network cable. HD-over-coax and media converters are real bridges between the two, and in buildings where the pathway is genuinely difficult they save real money without costing you the picture.
What they cost you is headroom: analytics at the camera, resolution you can raise later, PoE, and a return audio path. If none of those are on your list, reuse is a sound decision and you should not be talked out of it. If some of them are, the cable saving is a discount on the wrong system.
Neither answer is universal. The building decides — and the building will tell you, if someone walks it with a light and writes down what is in the wall.
