There is a specific kind of bad night footage that owners describe the same way every time. The daytime picture is fine — sharp, clean, you can read a jacket logo. Then the light goes, the camera flips to night mode, and the image turns milky. Not dark. Foggy. Like someone breathed on the lens. Faces wash out, headlights bloom into shapeless white, and the middle of the frame goes soft in a way that no amount of squinting fixes.
The usual explanations get offered in order: it's raining, the lens is dirty, the camera is old, you bought cheap cameras.
Sometimes one of those is true. But there is a mechanism that produces exactly this picture, that has nothing to do with the quality of the camera, and that almost nobody mentions when they are selling you one. The camera is being blinded by its own light — and the thing doing the blinding is the clear plastic bubble in front of it.
What is actually happening inside the bubble
Almost every camera that sees in the dark carries its own illumination. A ring of infrared emitters sits around the lens, invisible to your eye, flooding the scene with light the sensor can see. That is why a night image looks black-and-white: the camera has stopped reading colour and started reading the reflection of the light it is producing itself.
Now consider where those emitters sit on a dome camera. The lens assembly and the emitter ring are both mounted inside the housing, and a single clear hemisphere — the bubble — covers both of them. Every photon the camera emits has to pass through that bubble on the way out.
Most of it does. But no transparent material passes light perfectly. At each surface the light crosses, a fraction reflects instead of transmitting. On a curved surface sitting a few centimetres in front of the lens, some of that reflected fraction has nowhere to go but back inward — and the lens is right there, aimed at it, wide open because it is night and the camera is starved for light.
So the sensor receives two things at once. It receives the scene, which is dim. And it receives a haze of the camera's own illumination that never left the housing, which is bright and close and spread across the whole frame. The result is a lifted black level and a loss of contrast everywhere — the milky look. Photographers call this class of problem flare. It is not an artefact of your recorder, your cabling, or your compression settings. It happens in the first few centimetres of the optical path, before any of that exists.
Two consequences follow, and they explain a lot of complaints that get misfiled as something else.
Your camera does not become less sharp at night — it becomes less contrasty. The detail is still being resolved. It is buried under a wash of stray light. That is why turning up sharpening in the settings never helps, and why the footage that looked acceptable on a monitor turns out to be useless when someone needs to identify a person from it.
A bright object makes it dramatically worse. Anything highly reflective inside the beam — a wet driveway, a white van, a snow-covered walkway, a retroreflective safety vest — throws a large amount of the camera's own light straight back at it, and the bubble takes another slice of that and reflects it inward too. This is why a camera can look fine over grass and go to mush the moment a car pulls into the frame.
Why the same camera gets worse over the years
The important part for anyone maintaining a building: this is not a fixed property of the camera. It degrades.
A dome bubble is a piece of optical plastic mounted outdoors, and it ages like one. Rain deposits mineral film. Dust and pollen bond into a haze that is not visible from the ground in daylight. Ultraviolet exposure gradually clouds the material itself. Fine scratches accumulate — from wind-blown grit, from pressure washing, and very often from someone wiping the bubble with a dry paper towel or a shop rag, which is an excellent way to grind dirt into plastic.
Every one of those changes does the same thing: it takes light that used to pass cleanly through and scatters it in all directions, including backward. A hazed or micro-scratched bubble is a far more efficient reflector than a new one. The camera's daytime image barely changes, because in daylight the scene is thousands of times brighter than the scatter. At night the scatter is the brightest thing in the optical path.
This is the reason cleaning a dome bubble counts as real maintenance rather than housekeeping. It is not cosmetic. It is the single cheapest thing that restores night performance on a domed site, and it is skipped on most buildings because nobody ever framed it as maintenance. If it is worth writing down anywhere, write it into the same schedule as your lamp and filter changes, and specify a soft cloth and clean water rather than whatever is under the sink. And when a bubble has gone permanently cloudy or crazed, replacing the bubble alone is usually possible — a far smaller job than replacing the camera, and one worth asking about before anyone quotes you new hardware.
The gasket nobody mentions
Manufacturers know all of this. Good dome cameras are designed against it: the emitter ring is separated from the lens by a light-blocking barrier, and a foam or rubber gasket seals the gap between the lens barrel and the inside surface of the bubble. The job of that gasket is to make sure no light can travel from the emitter side of the housing to the lens side without going out into the world first.
That gasket is the part that gets ruined during service.
It is soft, it is easy to pinch, and it sits exactly where a technician's fingers go when the camera is being re-aimed, refocused or cleaned. If it is left twisted, pinched under the lens barrel, dislodged to one side, or simply not reseated when the bubble goes back on, the light seal is broken and the camera acquires a direct internal path from its own emitters to its own lens. The failure signature is distinctive: a camera that was fine last month is now hazy at night, and often the haze is worse on one side of the frame than the other, because the gap is on one side.
The same thing happens when a bubble is fitted the wrong way round, when a replacement bubble from a different model is substituted because it physically fits, or when the foam has aged into a hard, compressed ring that no longer touches the plastic.
This is worth knowing because it turns a vague complaint into a specific question. "The night image got worse after the last service call" is not a complaint anyone can act on. "The night haze started after the cameras were cleaned in March — has the light gasket been reseated on that camera?" is a question that produces an answer, and the answer is often a five-minute fix rather than a replacement.
The check you can run yourself, tonight
You do not need equipment for this, and you do not need to go up a ladder — which you should not do anyway.
Watch the dusk transition on live view. Sit with the camera's live image as the light fades. The camera will switch to night mode at some point, and you will see it: colour drops out and the infrared comes on. If the picture is crisp in colour and goes milky within a second or two of the switch, the haze arrived with the camera's own illumination. Fog and drizzle do not appear instantly on a switch. Internal reflection does.
Compare cameras on the same night. If your site has both domes and any other form factor — turrets, bullets, anything where the light source is not behind the same window as the lens — pull them up side by side at the same moment. Genuine weather affects everything on site. If the domes are milky and the others are clean under the identical sky, the sky is not what you are looking at.
Check whether a bright object triggers it. Watch a domed camera as a vehicle enters the frame at night. If the whole image degrades when the vehicle arrives and recovers when it leaves, you are watching the camera's own light come back at it.
Compare against the same camera's day image. If daylight footage is sharp and night footage is soft rather than merely dark, the sensor and lens are working. The problem is in the illumination path.
Any of these gives you something concrete to hand to whoever maintains the system, which is the actual point. It is the difference between "the night footage is bad" and "the night haze appears at the mode switch and only on the domes."
So why does anyone still fit domes?
Because the flare is a trade, not a defect, and the other side of the trade is real.
Vandal resistance. A dome puts a hard, curved shell between the world and the camera. A curved surface is difficult to get purchase on, difficult to strike squarely, and difficult to knock out of aim with a swing. There is nothing protruding to grab. On a parkade level, a stairwell, a school corridor, a shelter, or anywhere the camera is within arm's reach of someone who resents it, this matters more than the night image does.
The aim is concealed. From below, you cannot easily tell where inside the bubble the lens is pointing. Someone who wants to know exactly what is covered has to work for it. A turret's aim is legible from across the street.
It stays where you put it. The lens is enclosed and the bubble locks over it. It cannot be nudged by a ladder, a pressure washer, a passing forklift, or a curious hand.
It looks like it belongs on a ceiling. This sounds trivial until you are the person specifying cameras for a lobby, a hotel corridor, a clinic, or a strata common area, and the tenant committee has opinions. A flush dome reads as building infrastructure. A turret reads as surveillance equipment. Both are honest; they are simply not the same message, and the message is part of the brief.
Turrets have their own costs, too. The exposed lens face collects dirt and needs wiping. The visible aim can be adjusted by anyone who can reach it. They protrude, which is a snag point in tight spaces. And they announce themselves — which is a benefit when deterrence is the goal and a problem when discretion is.
Choosing, honestly
The rule that survives contact with real buildings is short: put domes where the camera needs protecting, and turrets where the image needs protecting.
Domes earn their place indoors, in reach of the public, in ceilings, in corridors, in lifts, in parkades, and anywhere the housing is more likely to be attacked than the picture is to be needed at night. Indoors, with ambient lighting on all night, the flare mechanism is largely moot — there is enough light that the camera never has to turn on its own emitters.
Turrets earn their place outdoors on dark perimeters, over parking areas, on loading bays, on approaches and gates — anywhere the identifying detail you will one day need is going to be captured in the dark, by infrared, with wet pavement and vehicle paint throwing light back.
There is a third answer that gets overlooked: separate the light from the camera entirely. An external illuminator mounted away from the lens, or plain white lighting on the area, removes the mechanism completely and improves the picture for human eyes at the same time. On a site where you have both a night-image problem and a lighting problem, one fixture can address both.
And the honest limit on all of this: form factor is one variable among several. Lens choice, mounting height, camera exposure settings, the reflectivity of the surfaces in frame, and whether anyone ever tuned the night settings at all will each move the picture as much as the housing does. If your night footage is poor, the bubble is a strong first suspect on a domed outdoor camera at a dark site — it is not the only one, and anyone who tells you a single change will certainly fix it has not looked.
What to ask for on your next quote
Ask which form factor is proposed for each outdoor camera, and why. "Domes throughout" is a default, not a decision. There is nothing wrong with the answer; there is something wrong with there being no reason behind it.
Ask whether infrared is on-camera or separate on the dark exterior positions. This is the question that most directly affects whether you get a usable face at night.
Ask for bubble cleaning to be a named line in the maintenance schedule, with the cleaning method specified. If it is not written down, it does not happen, and the degradation is invisible until the night you need the footage.
After any service visit, look at the night image. A gasket disturbed during a cleaning or re-aim shows up immediately on live view and is trivial to correct while the technician is still on site — and considerably less trivial six months later, when nobody remembers the visit.
The short version
The foggy night image is often the camera photographing its own infrared light reflecting off the inside of its dome. It gets worse as the bubble ages, hazes and picks up fine scratches, and it can appear abruptly after a service visit if the internal light gasket was not reseated. A turret separates the illuminator from the lens and largely removes the mechanism.
Domes are not a mistake. They resist attack, conceal their aim, hold their position, and disappear into a ceiling — and on indoor and in-reach positions those are the right things to optimize for. The mistake is choosing a housing without anyone naming the trade, and then blaming the picture on the weather.
