Lens Preview

See how clearly a camera records a person, a face, a vehicle or a licence plate at a given distance. Pick a camera, set the distance (and the zoom, for varifocal and PTZ cameras), and the preview redraws our reference photos at the exact number of pixels that camera puts on the subject.

See how clearly a camera records a person, face, vehicle or licence plate at any distance.

Lens figures for this model are calculated from its sensor and focal length; they are not on the datasheet.

Detail on the subject

—

DORI level

—

Licence plate

—

Camera’s view at this distance

—

DORI distances for this camera

How far away each level of detail reaches at the current zoom. Highlighted rows are met at your distance.

Identify

Identify a stranger beyond reasonable doubt (250 px/m)

—

Recognize

Recognize someone you already know (125 px/m)

—

Observe

See clothing, colours and what they are doing (62.5 px/m)

—

Detect

Tell that a person or vehicle is there (25 px/m)

—

Full camera image

The preview above is zoomed in. This sketch shows where that area sits in the whole picture, at its real size.

How the preview works

The reference photos were taken at 2,000 pixels per metre for the person and 1,500 for the vehicle, measured from the 12-inch licence plate in each shot. The tool works out how many pixels per metre the camera puts on the subject at your distance, shrinks the photo to that, then enlarges it again, the same as zooming into a recording on the NVR.

It then adds what the camera’s own image processor does: noise reduction smooths fine texture such as skin, fabric and pavement, and sharpening draws a light outline around edges. We matched this to real footage from Uniview IPC3624SR cameras at about 30 feet. Untick “Camera image processing” to see the clean version.

Pixel density comes from each camera’s datasheet DORI distances, which Uniview measures at the centre of the image. DORI is the IEC 62676-4 standard: 25 px/m to detect, 62.5 to observe, 125 to recognize and 250 to identify.

This is the best case: daylight, sharp focus, a still subject in the centre of the image. At night on infrared, with heavy compression or a moving subject, expect less. Wide-angle and fisheye lenses also lose detail toward the edges.

For licence plate recognition, Uniview recommends the plate be 90 to 300 pixels wide in the image, about 130 ideally.