Fisheye for overview at low cost; multi-sensor where detail at distance matters, because each imager has its own full resolution.

A fisheye spreads one sensor's resolution across the entire scene, so anything far from the camera is small and soft.

A multi-sensor camera has several imagers, each covering a sector at full resolution, which gives usable detail much further out.

Multi-sensor units are larger, heavier and considerably more expensive, and they produce several streams to record.

Fisheye suits a shop floor, an office or a small warehouse where knowing what happened and roughly who was involved is sufficient.

Multi-sensor suits large yards, car parks and concourses where identification at distance is required.

It is the correction of the circular fisheye image into conventional rectangular views, and doing it at the viewing client preserves the whole scene for later.

The raw fisheye image is unusable as it stands - a circular, heavily distorted view.

Dewarping in the camera sends conventional views to the recorder, which is simpler and throws away everything outside those views.

Dewarping at the client records the full circular image and lets an operator pan around it afterwards, including in recorded footage.

That retrospective capability is a real investigative advantage and is usually worth the extra client-side processing.

It requires the recording and viewing software to support the specific camera model, so confirm compatibility before purchase rather than after.

Often yes for coverage, not always for detail - and rarely for bandwidth.

One mounting point, one cable and one licence replacing four is a genuine installation saving.

Coverage is also better in one respect: there are no blind spots between separately aimed cameras.

Detail is the trade. A fisheye at the centre of a large space gives far less identification capability at the edges than a dedicated camera aimed there.

Bandwidth and storage often do not fall, because the panoramic stream is large and continuous.

The usual sensible design is a panoramic camera for overview plus a small number of fixed cameras at the points where identification matters, such as entrances and tills.

Centrally and high, on the ceiling for a circular view or on a wall for a hemispheric one - and the position cannot be corrected later by aiming.

A ceiling-mounted fisheye sees the area beneath and around it, with detail falling off toward the edge.

A wall-mounted unit gives a 180-degree view of the space in front, which suits a room with one main aspect.

Obstructions matter enormously: a column, a beam or high racking blocks a whole sector and cannot be worked around by re-aiming.

Mounting height sets the trade between coverage and detail - higher covers more and resolves less.

Model the coverage before installation. A panoramic camera in the wrong place is a much bigger problem than a conventional camera in the wrong place.

Considerably more than a conventional camera, and a multi-sensor unit produces several streams rather than one.

High-resolution sensors running continuously generate large amounts of data, and panoramic cameras are high-resolution by definition.

Multi-sensor cameras are effectively several cameras in one housing as far as the network and the recorder are concerned.

Recorder licensing frequently counts each imager separately, which is worth confirming before assuming a licensing saving.

Modern compression and intelligent bit-rate control reduce the load substantially in static scenes, and much less in busy ones.

Calculate storage from the actual scene activity and retention period rather than from a general figure, and check the network path can carry the stream at peak.