When someone tells us their old system was useless, they are nearly always describing night footage. The daytime recordings were fine. The one that mattered was at 2am, and it showed a shape.
The causes repeat from building to building, and every one of them is decided during design rather than at the parts counter. None are fixed by buying a higher megapixel count.
The light is behind the subject
A porch light, a security floodlight or a lit doorway directly behind someone turns them into a silhouette. The camera exposes for the brightest thing in frame, and the brightest thing is not the face you wanted.
This is the most common single cause and usually the cheapest to fix, because the answer is generally moving the camera a few feet or changing its angle rather than buying different hardware. It is only cheap if it is caught during design. After installation it is a re-mount and often a new cable run.
The infrared is bouncing back
Infrared illuminators reflect off anything close and pale: a wall, a soffit, a downpipe, a gutter, even heavy rain or falling snow. The reflection floods the sensor and the middle of the frame goes flat white.
A camera mounted tight under an eave will frequently do this to itself, and it is the reason dome housings need particular care outdoors: any dirt or scratching on the cover scatters the camera’s own IR straight back into its lens.
The lens is too wide for the distance
A wide lens covering an entire yard puts very few pixels on a person at the far end of it. In daylight there is enough contrast to work out what happened anyway. At night there is not, because the sensor is already struggling and there is nothing left to recover.
Identification is a lens and distance problem long before it is a resolution problem. A 4K camera with the wrong lens at the wrong distance produces a very detailed picture of an unidentifiable person.
The camera is behind glass
An interior camera looking out through a window is a recurring disappointment. At night it reflects the room back at itself, and if it has IR it illuminates the glass rather than the scene. If the job is outside, the camera belongs outside.
Nobody looked at the site at night
This is the honest root cause of all of the above. Conditions at three in the afternoon tell you almost nothing about the same wall at eleven at night. Where a system will be judged on night footage, and most are, the assessment has to happen after dark at least once.
That is one of the practical arguments for a local installer: a return visit at 10pm is a scheduling detail when the crew is twenty minutes away and a line item when it is not.
What actually helps a camera see at night
Deciding each position against a specific job, so the lens matches the distance. Placing cameras so existing light works with them rather than against them. Adding light where the answer genuinely is more light, which is less often than people assume. And accepting fewer, better-placed cameras over even coverage.
More cameras spread evenly around a building is the approach that produces the most footage and the least usable evidence, and it fails most visibly at night.
Night performance is decided at the design stage, not in the app, which is why Elite Security looks at a site after dark before specifying it. More like this in the blog; for a quote rather than an explanation, start with commercial CCTV.