Unfortunately, the difficulties encountered are often exacerbated by some security consultants whose understanding of the technology is less than complete.
All CCTV systems, of course, must now comply with the terms and conditions of the Data Protection Act. In reality, this covers authorised access to the system (including live displays, and not just recordings), the handling of data, storage times and subject's rights to access (whereby an individual has the right to view recordings in which he or she could be identified).
The consequences of the Data Protection Act for CCTV system end users have been explained in a previous edition of SMT ('Process... and be damned', SMT, September 2001, pp42-43).
A commonly held misconception is that the Data Protection Act requires that all CCTV systems be set up to store one month's worth of recordings. Not so. The Act's CCTV guidelines state that the data storage time must be appropriate to the application, but it's up to the manager(s) responsible for the system to decide – and justify – the storage time required.
For example, seven days might be appropriate for a sports centre, 14 days for an industrial site and 28 days for a town centre scheme. CCTV systems covering financial transactions (at ATM machines, for example) might need three months' storage due to the length of time required for problems to be noted and investigated. Whatever length of storage time is decided upon, it's essential that the management team write it down in their operational policy. This is the Data Protection Act requirement, not the storage time.
Two main points arise from any discussion over storage time that are highly relevant to the end user's system choice. First, your preferred digital video recorder should have features that help you comply with the Act (eg password access, logging and storage time, etc). In addition, you don't need to buy more storage capacity than you feel you need for your application (you should be aware that my sales manager will kill me for making this remark!)
A question of compression
Image compression is possibly the most misunderstood topic among end users with respect to digital video recording systems. Many people glibly refer to MPEG, JPEG, M-JPEG, Wavelet or H261 without having even the vaguest idea what they actually mean or how they differ from one another. This situation is not helped by system manufacturers who always claim that their method is the best.
By definition, image compression systems make alterations to the images during the compression and decompression process. This is also true of analogue recordings, where detail is lost in each stage of the recording and replay process (transmission, filtering, amplification, conversion to electro-magnetic waveform, transfer onto tape and all the way back up this process on replay).
Broadly speaking, image compression methods can be split into three categories: conditional refresh, non-conditional refresh and 'lossless'.
Conditional refresh image compression methods store key images at a lower-than-capture rate, thereafter interpolating the intermediate images by calculation. This type of compression method is used for multi-media systems, video conferencing and some makes of digital video recorder. MPEG, H261 and H263 are examples of this standard, but there are also many types of proprietary method being used.
These types of system have the advantage that they require less storage space, but on the downside they do suffer from poorer image quality and reduced evidential robustness. Lower storage costs are offset by the increased price of the hardware necessary to run conditional refresh compression systems.
Non-conditional refresh compression methods include JPEG and Wavelet compression. Of these, JPEG is the accepted industry standard, although Wavelet is also common. As stated, JPEG is sometimes (incorrectly) referred to as M-JPEG (for 'Motion-JPEG'). This is NOT a different compression method, but is simply a time series of JPEG images which are time lapse or real time video recordings.
It's well worth CCTV end users remembering that non-conditional refresh systems (eg JPEG and Wavelet) compress each image individually, thereby better preserving the integrity of the image data.
Recordings should be stored in such a way that no user intervention is required at any time – precluding the use of any kind of data tape (eg DAT, AIT or digital videotape). Although tape jukeboxes (ie auto changers) are available, they are fiendishly exp
In spite of claims to the contrary by fans of Wavelet, it only offers a minor reduction in file size for CCTV images in general. Both methods compress each image individually, and don't rely on any form of interpolation between images in a time series. Both methods can be described as 'lossy' (ie information is lost during the compression process, the amount of loss dependent on the compression ratio).
Higher compression rates will result in higher loss of information. With normal settings, the loss of detail should be minimal and image artefacts not so obvious. Low compression settings will normally result in images where the compression is totally imperceptible.
For their part, JPEG images have been used successfully in UK courts for over seven years. A new standard – designated JPEG2000 – is currently beginning to take shape, although it's not yet established. Confusingly, end users should note that JPEG2000 is based on a variant of Wavelet compression.
'Lossless' compression systems are not normally used for images (except medical images, such as cardiology, x-ray, ultrasound and the like which are usually all monochrome) as the compression results in very large file sizes. This particular method is often used for computer files, where every single bit of information is critical to the integrity of the file. The commonly-used Winzip is a good example of a lossless compression program.
It's a little known fact, but it's also possible to have lossless JPEG compression. That said, this centres on monochrome images only and, again, results in very large file sizes.
Recordings should be stored in such a way that no user intervention is required at any time – precluding the use of any kind of data tape (eg DAT, AIT or digital videotape). Although tape jukeboxes (ie auto changers) are available, they are fiendishly expensive and are such mechanically complex devices that their long-term reliability must be deemed to be suspect.
Currently, hard disks are by far the best storage option for digital video recorders. That scenario will remain true for some considerable time until new storage technologies are developed. Hard disk costs are falling and the storage capacities are increasing dramatically. Make sure your chosen digital video recorder can handle large disk capacities if you need them (note that even larger disk capacities should be integrated in the recorder, and should not require the added complication and space of external boxes).
We would consider a typical average disk capacity for a 16-camera system to be about 360 Gb, with a larger system nearer 1,000 Gb.
Assessing data transport features
At the end of the day, it's advisable that CCTV end users look for a simple, straightforward user interface which is easy to control and yet has all the features you need. Most important of all, it must be easy to train members of the security team to operate the system.
If you are going to use your recordings for legal action, or even just for an investigation, you will almost certainly have a need to extract recordings and use them elsewhere. The most obvious example is handing over recordings to the police for criminal investigations.
Such a procedure is relatively easy with analogue tapes, whereby you simply render the original tape (although it is then effectively no longer available to you). With digital systems an identical copy is usually made on some form of digital media, and it's the copy that's handed over.
The data transport feature is a critical part of the function of a digital recorder, and the method and process must be smooth and easy to carry out. As manufacturers ourselves, we are constantly astonished at how few digital video recording systems have a well thought-out data transport system. What's the point of having great recordings if the police can't access them?
In real terms, this illustrates the futility of recording onto DAT or digital videotapes. The police cannot use the recordings, thus a VHS tape must be generated from the system. A good digital video recorder will provide this facility, but this loses image quality and ALL the automatic audit trail data. It's far better to maintain the recordings in digital format, and offer all the fast access and search features that these provide.
The data transport feature is a critical part of the function of a digital recorder, and the method and process must be smooth and easy to carry out. What is the point of having great recordings if the police can’t gain easy access to them?
We are beginning to see DVD drives being specified in tender documents for this very purpose. Although it's a relatively simple matter to use a DVD drive instead of a CD drive to realise the benefit of the higher capacity offered by DVD, we don't recommend this at the present time. There are a number of DVD recording standards (DVD-RAM, DVD-RW, DVD+RW, etc) fighting for dominance, so it's likely that there'll be huge compatibility problems for some time to come.
It's also unreasonable to expect that your local police force will be fully informed about the differences in DVD recordable disk standards, and for them to have easy access to just the right kind of drive. The message for end users is that they should stick with CD!
Currently, the only credible medium for this is CD-Rom. All police forces have Windows-based personal computer systems with CD-Rom drives, and so do all other authorities which might have cause to view CCTV data. The means to replay the data on the CD must also be included on any disks generated, and this must be supplied free of charge for maximum flexibility and practicality.
Service, support and expansion
In general, installed CCTV systems will be designed to work for five to ten years. Those of you who've purchased a set-up in the past will also know they're expensive. It follows that the availability of service and support must be a critical part of the buying decision.
Think about where the equipment is manufactured, as well as the availability of knowledgeable local support. How familiar is your chosen installation company with the kit? Are they committed to it, or do they just offer the cheapest box for every job? Do they work closely with the manufacturer? As an end user you stand a much better chance of solving problems and realising the best possible service if they do.
If your security team is prone to a high churn rate – and, let's face it, many are – then you need to ensure the long-term availability of training on the system. You also need to make certain of an equally long-term availability of upgrades, spare parts and service for your system. With this in mind, take stock of the track records of both your installer and manufacturer.
In an ideal world, your digital CCTV system will be expandable. Cameras are frequently added to CCTV installations as the existing ones prove their worth. A good digital video recording system should scale-up to large numbers of cameras without you having to discard any equipment, and without you having to add further screens and control devices.
All systems will record times and dates when monitoring, of course, which is a crucial part of any recording. The trouble is that clocks can be incorrectly set (or not set at all), they can drift or they can be affected by power failures. These remarks hold true for both analogue and digital systems. The traditional solution is to set the time correctly, and put procedures in place to check the time and date settings periodically.
That said, the better digital system recorders should offer an option of automatic synchronisation to an external reference clock signal which is always accurate. We recommend GPS, as the Rugby Time Clock signal does go off air periodically.
Further, if multiple recorders are used then they should automatically synchronise their clocks to the one with the reference clock signal.
One final pitfall for end users to watch out for when it comes to clocks and digital recorders is to check that the bi-annual clock changes required by daylight saving time are handled correctly. Some digital video recorders record over the overlapped hour in the autumn clock change. Remember, too, that the time and date form a vital part of the audit trail for the system, the audit trail being an obligatory feature for any digital recording set-up.
The audit trail must be tamper-proof, but easily accessed. It must be copied (preferably on an automatic basis) onto any removable media along with the recordings themselves. A rock solid audit trail such as this is a vital component of Data Protection Act compliance.
Source
SMT
Postscript
Alastair McLeod is managing director of Visimetrics (UK)
No comments yet