One of the most interesting jobs we independent CCTV consultants are hired to undertake is the comparison of digital video recorders (usually abbreviated to DVR) especially on behalf of our customers who might be non-technical people. We enjoy getting into the nitty-gritty of testing the performance and foibles of various manufacturers’ DVR offerings, while the aforementioned customers usually avoid it like the plague! It’s an ‘Anoraks Only’ zone.
This work has highlighted the third of the annoying aspects of CCTV; particularly regarding DVRs and their software-driven interfaces hosted on a pc screen. The problem is that digitized CCTV images (both live and recorded) are displayed to the CCTV operator by the machine’s GUI (graphical user interface)(pronounced “goo-ee”, if you were wondering) on a computer screen. That display is made up of an array of pixels, each capable of showing almost any colour you care to name. Wouldn’t it make sense to achieve best picture clarity for the viewer by arranging for each pixel in the footage to correspond with a single pixel on the display? Surely, yes, because the alternatives reduce picture quality. Here’s how they do it:
If the image is displayed using fewer pixels than the footage actually contains (we technical people call that ‘smaller’) then many of the picture details simply get ‘ignored’ by the display because the software driving the display has to recreate the image (‘downsample’) so that each of the display’s pixels knows what colour to show. The downsampling software can sully the CCTV image too.
By the same argument, if the image is displayed using more pixels than the footage actually contains (we technical people call that ‘bigger’) then many of the picture details simply get created by the display equipment. Significantly, the software driving the display has to recreate the image (‘upsample’) so that each of the display’s pixels knows what colour to show. The upsampling software can sully the CCTV image too. It is simply fabricating its detailed information.
In order to arrange the CCTV images on the GUI, be it a single image or maybe 25 images in a ‘mosaic’ of ‘cameos’, the software almost invariably shrinks or expands the images as described above, thereby reducing clarity. Sometimes the software will ‘auto-fit’ the array of images to fill the screen and change their aspect ratios in order to fill it, giving rather unhelpful distortion of the footage. Sometimes it will allow the operator to arrange and squash the images that they wish to watch into the available screen space by dragging and sizing floating ‘windows’ around the screen, leading to the same unhelpful distortion.
When testing DVRs for my customer I asked each product’s attendant salesman if their GUI included a button that simply set footage to the ideal 1:1 pixel assignment or, at least, something to force the correct aspect ratio to be restored. Their machines didn’t. In fact, only when I spoke to the actual designer of one well-established DVR did I get the response, “What a good idea! We’d not thought of that.” In the words of the average American teenager… “duh?!”
Ok, it’s easy for me to play the smart-arsed armchair expert here, but surelythe people who design these things should be thinking about these things. It strikes me that far too little thought often goes into these products from the point of view of the people who actually use them to do their job. So, the lesson here? Ask this question of your prospective DVR supplier. Heck, if you’ve already got one, ask them too and see the look on their face 😉 and suggest that they get one up on their competitors by putting this into their next software update.
Until next time; stay focused.