When selecting a camera, it's important to consider the scene the camera will be looking at and what the end user expects from their CCTV camera. Many times we hear people ask for a very wide angle shot so that they can see as much as possible from one camera to save on cost. That's fine, but in the same breath they also want to read license plates and identify a person from 30 metres, usually at night!
Unfortunately, Hollywood is responsible for much of the misunderstanding of what a CCTV surveillance system is really capable of.
Take a look at a real CCTV scene below. This is a shot from a 4 megapixel IP camera. The native resolution is 2592x1520.
If we use DIGITAL zoom to enlarge the number plate on the white SUV this is what we get.
The actual pixel size of the license plate area from the image above is only 34x23 (we've enlarged it here so you can see it) giving a total number of 782 pixels. Out of the 4 million pixels available, we are only relying on less than 800 to do the job for us. If the car were only one metre further away, you would not be able to discern any characters at all. Digital zoom cannot add missing detail.
But what happens when we use a vari-focal camera and adjust the lens to get a larger view of what's important to us? This is OPTICAL zoom. Sure, you won't see as much of the scene, but you will see the important details more clearly.
Notice that there are now 4233 pixels (83x51) defining the license plate area - a significant increase as you can see below.
Again, we have enlarged the image for easy viewing, but you can clearly see that having the license plate larger in the scene by zooming the LENS (optical) there can be no misinterpreting the results. It's evident that pixel resolution alone is not the only factor in camera performance.
So what's the solution if you need to see a wide view AND get important details of specific areas of the scene? Multiple cameras. Yes, it costs a bit more, but isn't a CCTV system that doesn't do what you want even more expensive?