… but what about the poor installer? … asks Dave Mulcahy of Vision Factory
A senior and distinguished (slightly greying) member of our industry, hobbled up to me at IFSEC this year and said, "Dave, I've been around most of this show and haven't seen one multiplexer, VCR or matrix … it's like I'm at the wrong exhibition!"

There obviously were examples of 20th century technology at the NEC in May, (probably the majority being in the centre's control room!) but the overwhelming impression was one of PCs and flat screens on every stand.

If my friend, with all his experience, was intimidated by this, what was it like for the average small installer, who has spent the last 10 to 15 years getting to grips with analogue multiplexers and VCRs?

Most installers now know what they will get if they order a simplex or a duplex multiplexer. All multiplexers do basically the same thing. Some have more or different features than others and the picture quality and speed can vary, but at the end of the day, the installer knows from experience what to expect and what questions to ask the supplier.

With the advent of digital recording the situation is far more complex and there is significant confusion on the part of the average installer with many differing approaches.

A lot of the systems on the market especially those based on PCs and Windows, can be complex to configure and operate. This is a significant obstacle for the hard-pressed installer who is probably working on low margins and cannot afford to have delays due to complex configuration or dealing with operational problems from the end user.

Digital video recording owes more to the computer world than to VHS technology. Therefore, many manufacturers of the "new technology" who are attempting to enter our industry do not have a background in analogue recording or control equipment, i.e. they don't speak the installer's language. This can often lead to misunderstandings and potentially disastrous consequences for all parties involved.

With this article, I hope to allay some of the fears expressed to me recently by installers and end users who are hoping to upgrade to "digital", by establishing some facts about analogue and digital recording. These will allow me to make comparisons between the two approaches and provide some useful "rules of thumb" when upgrading to digital.

The main variables in any CCTV
recording system are:

  • Number of Cameras in the system.
  • Picture Quality
  • Update time (How often a snapshot is recorded from any particular camera)
  • Recording duration (At what time interval can the tape or equivalent, be changed)
  • Archive time (How long the pictures can be kept before they must be erased)

Comparison of Analogue and Digital recording Analogue multiplexers are generally available with 4, 8, 9, or 16, inputs (with the Vision Factory Cameo giving 18 and Montage Plus providing 24 or 32).

Digital recorders have progressed a long way in the last two years, with a number of established CCTV equipment manufacturers launching digital equivalents to their mainstream multiplexers. (In the case of the Vision Factory, Cameo D1 was launched at IFSEC2001 and is the digital recording equivalent to Cameo, our medium system multiplexer. The Cameo D1 for example, is available with 9 or 18 inputs and is designed to be virtually a drop in replacement for a Cameo multiplexer and VCR. Operation and set-up is similar to the analogue machine and so will be instantly recognised by the many installers who have used our products in the past.)

Analogue picture quality from the VCR varies from LP VHS, through VHS, LP SVHS to SVHS, (LP is of course long play). It is also affected by the quality of the multiplexer, however there is no control over the quality of recorded images from individual cameras, other than that of the camera and lens in the first place.

Digital picture quality – I shouldn't need to explain why digital recorders must compress pictures to record them efficiently. Most of us have by now heard of the various compression methods such as JPEG, MJPEG, Wavelet and MPEG2 if not H261, H263 and MPEG4. These can be split into two groups, which I call Still Compression and Motion Compression.

Still Compression
JPEG – The standard for still pictures in PCs and digital cameras. Frames of video that are multiplexed need to exist as snapshots for analysis and to be used in evidence.

MJPEG – JPEG images replayed quickly in sequence – nothing more than this.

Wavelet – Also now known as JPEG2000. Looks slightly better subjectively than JPEG at high compression (lowest quality) but not as good at the higher quality end.

Motion Compression
MPEG2 – As used by your DVD player and digital TV. OK for movies but achieves high compression by mostly only recording the differences between frames so cannot really be used with a multiplex of images where each frame must exist by itself.

H261 and H263 – Used by video conferencing and transmission equipment, works in a similar way to MPEG2 and so cannot usually be used in multiplexing.

MPEG4 – The newest standard for DVDs and digital TV. This is far too clever! It doesn't just record the differences between frames, it looks to see if the changes are like shapes it has seen before. If so, it uses the prestored shape instead. A bit like converting movies to computer graphics – frightening!

So most Digital recording multiplexers will use JPEG or Wavelet (the Vision Factory Cameo D1 uses JPEG). But what does this really do to the pictures? Most people think there is a direct correlation between the "file size", i.e. the amount of compression applied, and the picture resolution. This is only partly true as, more importantly, compression throws away the lower contrast bits of the picture first.

This concept is best demonstrated. The picture of the car (fig1) was recorded on a Cameo D1 at various quality levels giving file sizes between 10KB and 50KB. The images were then written to CD and replayed on a PC. This example is the 40KB image.

Fig 2 shows high and low contrast regions from each of the different quality recordings.

The high contrast sections of each image, i.e. the number plates, retain their resolution even when compressed down to 10KB, whereas the lower contrast areas illustrated by the edge of the road and the man hole cover become "blocky" much sooner. (Cameo D1 allows the record quality to be set for each camera and changed dynamically on an alarm.) Increasing the quality to maximum gives the best chance of seeing detail in lower contrast areas and can help with facial recognition where lighting is poor.

The digital recorder's resolution before compression is also important. (The Cameo D1 conforms to the broadcast TV standard of 720 pixels per line allowing better than SVHS recording - over 500TVL with mono cameras). However, some PC based systems use the CIF standard of 360 pixels per line as used by webcams. Results with CIF will never be better than VHS. Although I have just argued that resolution does not always depend on file size, we all still need some rules of thumb to compare analogue to digital so here we go:

  • A VHS quality picture requires 20KBytes
  • An SVHS quality picture requires 30KBytes

Analogue Update time – If the VCR is in a timelapse mode then this defines the record speed and thus the update time per camera. The record speed does not depend on the multiplexer unless the VCR is in real time, then it is down to the multiplexer to record as fast as possible. Update time tables are available from most manufacturers and also CCTV text books (although I have seen a few that are wildly inaccurate). Standard 24 hour mode is always 50/9 = 5.55 pictures/second, however it is usually good enough to assume: 24 hour mode is about 5 pictures per second.

Most installers have put in hundreds of 16 way multiplexers with the VCR in 24hour mode so all they have to remember is that: 16 cameras in 24hour mode are recorded in about 3 seconds.

Digital Update time – It helps to think of a digital recorder in terms of a conventional multiplexer and a VCR combined. The overall record rate must be selected just like the TL mode on a VCR. The Cameo D1 record set up asks the user to select 1 in N frames. 1 in 1 gives 25 pictures per second (like a VCR in 3 hour mode) so all 18 cameras are recorded in about 0.7 seconds. 1 in 5 frames is 25/5 = 5 pictures per second and is the equivalent to a multiplexer with the VCR in 24 hour mode – again 16 cameras will be recorded in about 3 seconds.

(Both Analogue and Digital Cameos can modify the recording on an external alarm or on activity detection. Cameras in alarm can be prioritised so they are recorded more often at the expense of the other cameras. This does not reduce the recording duration. The overall record speed can also be increased but this reduces the overall record time. A Cameo D1 can also increase the record quality on an alarm as described earlier, which of course cannot be achieved with an analogue system.) Analogue Recording Duration – How often the site can be visited to change tapes is an important issue that often results in a system, which is a compromise. If the site is manned 24 hours a day, tapes can be changed more often, speeding up the record process, but many sites cannot be visited over weekends and holidays. The use of cascaded VCRs (end of tape triggers the next VCR) is often suggested and sometimes employed but many users don't want to purchase and service extra VCRs! Even where there is 24 hour manned security, many systems run in 24 hour mode for no other reason than "the manager doesn't want any more tapes!"

A useful fact to remember is: An E180 tapes holds about half a million pictures It follows that:

  • An E180 (VHS) tape is equivalent to 10GB: 500,000 x 20KB
  • An SE180 (SVHS) tape is equivalent to 15GB: 500,000 x 30KB
Digital Recording Duration – Until recently, digital recorders were only available with limited internal hard disc capacity. Just two years ago a 10GB Hard drive was about the norm for digital recorders, i.e. they could hold about one VHS tape's worth of pictures. Most of these early machines were only really useful as event recorders, or if left to record all the time, had to be backed up at least once per day, usually on expensive and fragile DAT tapes or similar. Granted, they gave instant access to very recent information but for anything more than a day old, tapes had to be searched again.

(Cameo D1 was one of the first machines to use the new 80GB discs in 2001 and will soon be supplied with the latest 160 GB devices. Although other manufacturers use the same drives, the Vision Factory Cameo D1 is probably unique in its position in the market by allowing up to six hard drives in one unit. Currently 80GB, 240GB and 480GB versions are available and upgradeable but this is soon to double. The current 480GB machine holds the equivalent of over 30 SVHS tapes which is a months worth of recording in "24hr" mode! Thus a unit can be left unattended while always allowing instant access to anything recorded in the last 30days! Not even my MD has a holiday that long.)

Analogue Tape Archive – It is a common misconception that the Data Protection Act requires tapes to be kept for 31 days, then erased. When a CCTV system is registered, the applicant must state why he needs to record at all. The archive must be kept for the shortest period that meets the normal operational requirements of the system. 31 days is the maximum allowed without a special case and is usually agreed to mainly on tape management issues. For example, a videotape's life is normally quoted as 12 days. Thus a VCR in 24 hour mode with 31 tapes will need a new batch of tapes once per year. A seven day archive would need the tapes renewed every three months which, of course, would never get done although the same number of tapes would need to purchased eventually. There are many more issues such as having to ensure the security and authenticity of the tape archive.

Digital Archive – As long as there is sufficient internal hard disc capacity, most systems can now completely dispense with the conventional removable "archive of everything". The digital recorder holds all the recording "on line" for the period allowed under data protection. This "archive" can be made secure by only allowing replay via a PIN number and, in the case of the D1, it is extremely unlikely anyone could decode pictures if the machine was stolen and taken apart as the hard discs would mean nothing to a PC.

(The D1, for example, has a built in CD writer to allow important events such as potential evidence to be protected.) A CD will hold about 20 minutes of reatime video and proportionately more of timelapse recordings. With a CDR now costing less than 30p it's about the same price per image as SVHS tape and is falling. A busy CCTV system may only generate a few CDs a week, which are much easier to manage than cupboards full of tapes. We consider the CD to be the first verifiable, removable recording, just like a first generation VHS tape. The hard disc can be thought of as a temporary store just like the memory in a conventional multiplexer. The former holds images for up to 30 days before making the recording, the latter, holds images for 1/5th second before recording to tape but the difference is just one of scale.

In summary
I hope, in this article, I have not to blinded you all with science but have just managed to get across the point that the digital recorder/multiplexer has now come of age and can be used without too much worry.

Its advantages over the old technology are almost too numerous to mention – I haven't even touched on networking for example. I think the main point to emphasise is that the belt and braces approach of throwing hard disc capacity at a job, makes installation far less time consuming and risky.

Why spend hours setting up a system to only record on alarms to save disc space when you can now record all the time and are guaranteed not to miss much?