Jon Hill, Video Product Specialist at Bewator, a Siemens company, has almost 15 years’ experience of working with CCTV and security systems. In his present role, he provides technical support on CCTV-related matters for the whole of the Bewator sales team throughout the UK, and as well as managing the security system installation arm of Siemens Building Technologies on a day-to-day basis.

Jon’s current period of employment with Bewator started in 2002, when he joined the company as Southern Area Sales Manager. Having achieved substantial success in this role, he was invited, two years later, to join the then newly formed Systems Team, where he worked on system design in conjunction with consultants, installers and end users. Jon was promoted to his present position in the final quarter of 2007.

Previous experience in the CCTV and security industry for Jon stretches back to 1994 when, after a period of working as an avionics fitter on the UK fleet of Chinook helicopters, he took up a position with Photon Security Systems as a wireman working on the manufacture of control systems and camera heads. With this company, he ultimately rose to become production manager responsible for a team of staff.

When Photon Security Systems sold the rights to its Surcha camera range to Bewator, Jon joined Bewator for the first time, and worked in the development laboratory on documentation and verification testing. Subsequently, he moved away from Bewator for a short period, but returned after gaining further invaluable sales experience.

What exactly is video analytics?

Answer: It’s the use of software algorithms for the analysis of CCTV images to detect alarm conditions, such as an intruder moving into a restricted area. The major benefit of this technology is the potential for automating the sometimes laborious task of monitoring CCTV images.

Isn’t that the same as motion detection?

Answer: Motion detection is a simple form of video analytics, but ordinary motion detection systems have their limitations. For example, if the lighting conditions suddenly change – perhaps the sun comes out from behind a cloud – false alarms are often produced. They’re also likely to be fooled by localised conditions such as movement of the camera pole or insects/birds passing through the field of view. For these reasons, plain motion detection systems are normally only useful for indoor applications.

So what does video analytics do?

Answer: The most useful video analytics software can detect objects of a specific size – typically persons – while ignoring irrelevant objects. It can then track these objects reliably, taking into account perspective, and generate an alarm when specific conditions are met, which is configured in the software by the user. In deciding whether to generate alarms, the software can take into account many factors, such as the size, speed, direction of travel of the target, and the distance it has moved. It also ignores environmental effects such as lighting changes, and weather patterns, and therefore produces a very low level of false alarms. Finally, if any attempt is made to tamper with or disable the cameras, good video analytics software will provide a warning.

Where are sophisticated video analytics likely to be used?

Answer: Applications are extremely diverse but tend to be external. These typically involve protecting the perimeter of a high-security installation such as a prison, airport or military compound. In these applications, CCTV combined with video analytics is often more cost-effective and less problematic than using conventional systems long term. The applications do need to be evaluated very carefully, however, as the number of cameras needed and their placement is likely to be different from an ordinary CCTV installation.

Applications for video analytics are changing all the time. The most common applications in the security arena are currently perimeter protection, object tracking, object abandonment, people counting and ANPR. 

There’s been a lot of discussion about desirability of using audio recording in CCTV systems. What’s the current position?

Answer: Unfortunately, nothing has yet been definitely resolved; the issue is one of liberty versus law and order, and there is no easy answer. However, a draft Code of Practice from the Information Commissioners Office (ICO) warns that audio recording in CCTV systems is “highly intrusive and unlikely to be justified”. Note that this document is only a draft at the moment, and it refers specifically to recording conversations between members of the public, rather than to recording audio in sensitive situations where the public wouldn’t normally be present. It does, nevertheless emphasise the need for caution, and to discuss fully the implications and desirability of audio recording with the client.

In what sorts of applications can audio recording be used?

Answer: Usually it’s only applied in high sensitivity areas, such as near a cash office, but it can also have other applications. It could be used near disabled facilities, for example, to enable someone who has problems but is out of camera shot to call for help. Whatever the application, audio recording can only be used to cover small, well-defined areas, usually in conjunction with a fixed camera and a directional microphone. It’s not really practical to provide audio coverage for large open spaces, such as the general areas of a shopping mall.

What about signage?

Answer: The current ICO Code of Practice doesn’t make specific mention of signs in relation to audio recording but depending on the application, the installation of signage to indicate that audio is being recorded as well as images should at least be considered.

Are there any specific technical issues relating to audio recording?

Answer: Microphones must be selected and sited with care if the recordings are not to be swamped by background noise – suppliers should be able to give guidance. It’s also worth noting that audio recordings are of limited evidential value unless they are lip-synchronised to the CCTV images. Lip synchronisation can, however, only be achieved with images recorded at rates of 12.5 frames per second and higher. This means that a CCTV installation with audio may need considerably more storage capacity than one without, where much lower frame rates may well be acceptable.

What’s the best option for monitors - CRT or TFT?

Answer: An interesting question. The most popular option is now TFT (Thin Film Transistor Liquid Crystal Display), as these monitors have so many benefits. They take up less space, they’re lighter, give off less heat, lower power consumption, offer good resolution right across the screen and, unlike CRT monitors, they don’t degrade significantly in performance as they get older. You should also consider price and availability. CRT monitors are tumbling in price but probably won’t be widely available for much longer due to reduced demand etc.

What do I need to consider when buying a TFT monitor?

Answer: The first thing is whether it will connect to the system you want to use it with! If you’re replacing a CRT monitor, you’ll need your TFT to have an analogue input, and this is something you need to be particularly careful about. Some otherwise quite good monitors perform badly when used with analogue inputs. Ideally you should try before you buy. If you’re working on a new digital installation, be aware that there are several kinds of digital video connectors (BNC, S-VHS, DVI etc), so make sure your monitor has the one you need.

What about resolution?

Answer: For TFT monitors, resolution is not quoted in lines, but in pixels. Usually, however, you’ll just see an abbreviation like VGA which is 640 x 480 pixels, SVGA which is 800 x 600 or XGA which is 1024 x 768. For comparison, a standard UK TV picture is 576 x 350 pixels, so you can see that even a VGA monitor is adequate for most purposes. You may want higher resolution for a large-screen monitor that will display multiple pictures. Note that the screen resolution should (at the very least) match the signal fed to the monitor.

Is refresh rate important?

Answer: For a TFT monitor, refresh rate is more accurately called response time. If it’s 20mS or longer, it can give smearing effects with fast moving targets. However, modern monitors typically have response times of 12mS or better, so this is no longer a big problem. This is very important to check as it can exacerbate some of the latency seen when looking at moving images.

Are there any other significant points to consider?

Answer: Look for a monitor with a wide viewing angle, and choose one with a glass screen. Without glass, the soft screen of the monitor will soon get marked and scratched as users point at targets, and soft screens are often hard to clean.

What’s the difference between RAID and ordinary hard disk storage systems for CCTV images?

Answer: From a CCTV perspective, digital storage of video images on a hard drive (typically using a DVR) is no different from other media stored on a conventional PC. Consequently the system processor makes decisions about where data is stored, depending on available disk space. With RAID configured systems, video data is managed in a more logical and secure fashion. The ultimate aim is to achieve a much greater level of data protection and provide disaster recovery in the event of disc failure.

What types of RAID system are commonly used in CCTV installations?

Answer: RAID 0 systems are often used to increase overall archive capacity, and thus provide cost efficiency. However, RAID 0 systems don’t provide any extra data security; for that RAID 1 or RAID 5 configurations are commonly used. With RAID 1, all CCTV images are written in duplicate (requiring at least two disks). This doubles the number of disks needed, but it also means that if one disk fails, or if video evidence in impounded, no images are lost because the mirrored archive is still available. RAID 5 is a little more refined. Data for the images is striped over several disks. The system also produces and stores additional parity data. If a disk fails, it can use this parity data to rebuild the images that would otherwise be lost. There is a disadvantage with RAID5 in that if the data is impounded by the police, there may not be a ‘mirrored’ version of the removed data available.

What are the arguments for and against removable storage?

Answer: Removable storage is very useful when the police need to impound images for use as evidence, since they don’t need to impound the whole recorder. However, this advantage is lost if a mixture of fixed and removable storage is used as, in these cases, the police will always want to impound the recorder and the removable storage. It’s also worth noting that removable storage is less secure than internal storage. It’s a lot easier to make a disk cassette ‘disappear’ than a DVR!

Is external RAID configured storage available?

Answer: Certainly, and this is a common approach where large amount of data is archived. It’s important to make sure that you get the right interface for the application. There are two in common use – NAS and SCSI – each offering different benefits. Internal RAID management systems are, of course, also offered with high end DVRs.

How do I decide how much storage capacity to allow when I’m specifying a DVR (digital video recorder)?

Answer: There’s no simple answer – every installation must be assessed individually. Key factors affecting storage are picture quality, frame rate, compression method and the length of time for which the images are required.

What are the picture quality options?

Answer: The lowest resolution now normally adopted is CIF (352X288). CIF is generally the rule of thumb when calculating storage capacity, but higher resolutions, such as 2CIF (704X288), 4CIF (704X576) and D1 (720x576), are now often specified. As a guide, CIF images recorded using MPEG4 compression are around 10Kb, 2CIF images around 20Kb and 4CIF around 40Kb. Megapixel cameras usually produce images between 80 and 200Kb each. A balance must be struck between resolution, archive time and budget.

What about frame rates?

Answer: Always allocate frame rates appropriate to the application. Live motion is 25 full frames-per-second (fps) but each image can be 40Kb or more (4CIF). This means about 1Mb of storage per second of data from each camera – about 3.6Gb per hour.  Using 12.5 fps halves storage requirements and still permits lip-sync audio. Where lip sync isn’t needed, 4 fps is often acceptable, with corresponding savings in storage.

How do compression methods affect the amount of storage?

Answer: Significantly! The challenge is to reproduce high quality, high-resolution video using the smallest amount of drive space, but remember that there are no free lunches! If a DVR claims much smaller file sizes than comparable machines with the same compression method, beware – reduced file-sizes usually mean reduced quality. 
   
How long should recordings be kept?

Answer: This depends on the application, but don't automatically adopt the “31-day standard”, a hangover from VHS tape. Digital recording is much more flexible. Discuss the options and costs with your client. In general, look at periods where video data cannot be recovered. If this happens to be 20 days, then 20 to 22 days of archive are appropriate.

Deciding on storage capacity seems complicated. Is there no easier solution?

Answer: When in doubt, seek the advice of several professional suppliers to ensure a balanced view. Also, remember that storage is now much less expensive, so over-specifying a little won’t significantly affect project costs.

When I’m planning CCTV installations, do I still need to consider PTZ cameras, or is it now OK just to use domes?

Answer: Modern domes are very capable, but there are applications where PTZ cameras are more appropriate. Important factors to take into account include, for example, lens parameters. Domes utilise modular camera/lens combinations with up to 30x optical zoom in some cases. However, overall quality is rarely as accurate as a bespoke camera/motorised lens combination housed within a PTZ head.    

Can digital zoom be used to extend the zoom range of a dome?

Answer: Be careful! Digital zoom can only enlarge what the camera has already captured. It may make distant objects easier to see but, unlike optical zoom, it doesn’t actually help to capture extra detail. This is important when, for example, facial recognition is required. With digital zoom, you may be able to make a face fill your screen, but it may also be too blurred to recognise.

Is image distortion always an issue?

Answer: Domes will always produce somewhat poorer pictures under certain circumstances because of the distortion introduced by the curvature of the hemisphere. However, in good light and at reasonable zoom ratios, the difference in quality between the pictures produced by a dome and those produced by a PTZ can be negligible, particularly if you use larger 7” domes which give better optical performance.

You mention good light. Does that mean light levels are also a factor?

Answer: Yes, always. Domes utilise modular cameras which are not quite as good at coping with low light conditions. Usually the result is more noise in the images at low light levels. This not only makes it harder to see what’s going on, it also makes the images harder to compress, which means they take up more hard disk space on a digital recorder.

Are there any other factors to consider?

Answer: Cost and durability. Domes cost less and are easier to install, but they’re less durable and harder to clean. Wash/wipe can, for example, be added to clean the glass of a PTZ camera, which still cannot be achieved using a standard dome.

To sum up, a dome will suffice in most CCTV applications. However, when considering high quality image reproduction, excellent low-light performance, supplementary IR lighting and harsh operating environments, I would recommend the use of a PTZ.