Disqus - Latest Comments for GeoffMoore

Re: Cisco Admits Defeat, Partners with Pelco

GeoffMoore — Tue, 22 Sep 2009 08:49:05 -0000

If I had to pick a word to describe any of this "coherent" would not be it. There are a few different threads to the story.

Cisco undoubtedly suffer from the typical "biggest kid on the block" Hubris that's usually the downfall of the biggest kid on the block. The marketing machine (or brain-washing machine if you like) from their traditional networking business works really well, and without rigorous decontamination it seems that once you're a Cisco guy you'll always be a Cisco guy, even when the facts about you and your competition no longer put you in a very favorable light. I think this same Hubris-factor leaks into other areas of their business too. I recall watching presentations by Cisco when they first came into the security market, and they really seemed to believe that they were just going to stroll in and clean up. That was a bit silly of them, I think, and it just hasn't happened.

There's obviously a larger strategy behind what Cisco were doing when they moved into CCTV, and this appears to be borne out by their move into building management and control systems too. In my view it looks like there's been a board meeting somewhere and someone's scribbled a picture up on a white board saying "yeh...if we do all the networking in the building, then let's see...what other devices could take advantage of that connectivity?". It makes sense for them to do that, and I'd have thought that given how far they lag behind all the other companies who are already in those sectors they would have sought partners back then...not now, so far down the line.

Pelco??Strategy??? Ya think???? They're now part of Schneider, who already have the TAC and Integral products...so I don't really know where Schneider thought they were going with that merger. Schneider's core business is building managament and control systems, so I don't really understand how Pelco can effectively "go out on a limb" and partner up with Cisco who compete in that area too...

Ok, so maybe Schneider see some advantage to cosying up to a networking company so they can claim to deliver complete building systems from end to end including the connectivity stuff in the middle, but if that's the case why Cisco? When they're going to run up against Cisco partners all over the world who will be fighting for that end of the business and getting in Schneider's way.

But where's the sense in Cisco being tied up with Pelco AND Tyco? To be honest I think American Dynamics have a better product offering in the IP space to what Pelco are doing with IP (for recorders at least). Sure, that's just my opinion, but even without that, where's the value in dividing Cisco's allegiance between two competing recording solutions??

I just don't get it.

For me Cisco are up on a shelf somewhere at the moment, waiting for them to produce something that is relevant to what I'm doing. I'm going to find it irritating if I now have Cisco butting into the conversations I have with Pelco...particularly when we've already decided they are not our networking provider of choice.

Admit defeat? Cisco?? Never! It's not in their vocabulary. But I do think there have probably been a few worried conversations upstairs at Cisco HQ with someone asking that awkward question : "why is nobody buying our stuff??" and after a few blank looks somebody has grabbed a Pelco brochure and said "quick everyone! Hide under this!!"

Re: GE Security for Sale -- Anyone Interested???

GeoffMoore — Sat, 29 Aug 2009 06:14:04 -0000

I'd go along with what a lot of people are saying here *on the surface of it*, and looking at GE Security as another manufacturer in a crowded sector, but I imagine there are a lot of "Wall Street" types who will look at it a different way.

We in the "real security industry" look at it from the perspective of how well the products they produce serve the security community, and in that regard I think GE are probably the worst of the Big Corporates. Unfortunately, I don't think that even GE themselves have ever looked at their business from that perspective.

GE got into the security sector proper when they purchased Interlogix some years back. As my information stands, that wasn't their first choice, but they were prevented from getting their hands on Honeywell Ademco by European monopoly legislation.

As it is, their primary motivation for getting into security was the same as their primary motivation for getting into any industry - that is (was) to force suppliers to go to GE Finance for credit. (I use the word "force" quite intentionally, because the procurement policies of GE businesses are extremely dictatorial, *telling* suppliers what their payment terms are going to be and *telling* them where they have to get their finance from).

GE have a large manufacturing facility in Ireland, and I got to know the way they worked fairly well through a couple of sources. The GE factory in Citywest (formerly Interlogix, formerly SLC Technologies, formerly Aritech) mostly assembled intruder alarm panels, some fire detetion and a bunch of low value detectors. Ireland (as can be seen in the results of various surveys) is a very epensive country these days, with some of the highest labour rates in the world, so it never made sense to me for them to build some of the lowest value components in the security industry there - especially when you add on a bunch of people whose jobs it is to do nothing except spend their days trying to enforce expensive SixSigma methodologies onto an unwilling workforce, or try to keep their leviathon poorly designed and crudely implemented Oracle based ERP system from falling over every five minutes.

Aritech was quite a respected company in the good old days, and their products still appear in a lot of Irish consultant's specifications, but really they're way over priced for what they are and they don't match the type of systems that are being installed in the local market (with the exception of the Eircom Phonewatch contract, which has to be the only thing that's kept that factory afloat over recent years). GE canned the R&D department a couple of years ago but continued to spend huge amounts of money on inneffective marketing that really didn't have any measurable impact in a market where plenty of also-rans where beating them at every hurdle.

Add to that constant product recalls, changes and defects and you get a big fat company that doesn't do anywhere near so well as many of the leaner, meaner companies that just make simple products at sensible prices.

I went to numerous product launches where I pulled holes in their new product with ease, where their technical sales people couldn't answer basic questions about how their products worked or how to get around even the simplest of issues.

I have had no cause to bother with GE Security products for years. I think the possibility of splitting the organisation back down into its component parts could be a very good one for those businesses GE purchased over the last few years, trying to make themselves into something they were not, but my skeptical head thinks that either (a) GE will already have done irrepairable damage to most of these businesses, or (b) the businesses that fell to GE's sword were probably not going to survive in the modern industry anyway.

It continues to be an admired organisation by those who measure business in terms of market capitolisation or even group turnover, but to me they've always just been a company that's big because it's big, not because it's good at anything

Re: How Do You Solve the Problem of Bad Surveillance Video?

GeoffMoore — Fri, 03 Jul 2009 01:43:59 -0000

The UK still has the dubious honor of having more CCTV cameras per capita than any other country in the world, and we've been doing it for so long that it's inevitable that a lot of work has gone into working out how to do it properly.

However, there are still plenty of badly installed systems in the UK with poor image quality. There are very reputable integrators who do it right but there are plenty of people who don't.

I'm constantly alarmed by the number of people in the industry who simply don't know how the technology works and hence have no real idea how to apply it in the real world. I'm even more alarmed now by the increasing size of the "IP only" generation of CCTV vendors who think that the systems they're installing are providing good images.

Re: How Do You Solve the Problem of Bad Surveillance Video?

GeoffMoore — Fri, 03 Jul 2009 01:27:19 -0000

I've seen that SWGIT document before. I think it's very wishy-washy and a bit old fashioned in its approach. Interestingly, I just had a look at the references at the back of the document and nearly all of them are UK Home Office Police Scientific Development Branch documents, including the 1994 Operational Requirements Specification Manual that I mentioned in my post, that was the fore-runner of the updated version that I linked to (http://www.homeoffice.gov.u.... It's very obsolete.

For those who don't know, I'm based in Dubai now, and here we have a thing called Dubai Police Law 24 (2009) (the revised version of Police Law 13 from abour 2007), which actually DICTATES how you use CCTV in a lot of premises. Hotels, banks, shopping malls, leisure facilities, filling stations, high-value shops (jewellery stores mostly) and pretty much anywhere that cash changes hands is audited annually and must pass Law 24 compliance if they are to have their trading license renewed. If they don't have a trading license they're dead in the water.

Although the law isn't massively technical in the way it's put together it is very specific in how you put systems together.:

As an example of how severe the rules are, here are just a few things that 4star and about hotels must have :

- a camera in every guest lift lobby on every floor to provide identification grade images
- a camera in every service lift lobby on every floor to provide identification grade images
- a camera in every guest lift to provide identification grade images
- a camera in every service lift to provide identification grade images
- a camera in every stairwell on every floor to provide identification grade images
- a WDR camera at every entrance or exit
- a camera at every utility area (gas storage, water storage, electrical intake, main telephone room)
- a camera at every point of sale
- a camera at every vehicle entrance and exit to provide license plate and overview recognition
- all digitally recorded for a minimum of 31 days
- building must have a security control room
- each dvr must have at least 2 monitors
- all recordings at minimum of 2CIF 5fps continuous at high quality
- security system must be on a physically separate network
- no wireless allowed

There's a bunch of other stuff too, but these are some of the most costly rules that are causing the most problems. And there's no arguing or negotiation. If you don't pass the audit you don't pass the audit.

The system has ended up being this way because of all the "apples and oranges" systems designs out there. It doesn't address operational requirements but it levels the playing field for the vendors a little by at least defining the quantity of cameras that need to go in and how they should be recorded.

It's certainly interesting. There are a whole lot of rules about manned guards, cash in transit, patrols etc too, all as part of the same law, and there's also a Federal Law for Abu Dhabi that doesn't go quite so far as this, but is due for revision and we believe that it's going to be similar to Law 24 when it hits.

One of the big challenges is that these laws are published in Arabic only, and the translations were a bit confusing at first, but now that the auditors are in full swing we know exactly what they're looking for, and there are a lot of unhappy hotels, banks and shopping malls as a result.

Re: How Do You Solve the Problem of Bad Surveillance Video?

GeoffMoore — Thu, 02 Jul 2009 07:15:22 -0000

Operational Requirements Specifications guys. It's the only proper way to specify CCTV.
If you design systems on the basis of what you're trying to achieve rather than what tools you'd like to use to do it, you're much more likely to get vendors to provide you with the correct solution, and the customer is much better able to determine whether they actually got what they wanted.

This methodology has been around in the uk since 1994. The original 1994 document was actually only superceded in 2007, but the new one is a very good general guide to implementing CCTV systems for non-technical people.

You'll find the document on how to do it here

http://scienceandresearch.h...

Re: Should You Use Cameras With Built In Storage?

GeoffMoore — Mon, 22 Jun 2009 12:04:44 -0000

I think that maybe there are applications in which this makes sense, but the problems that always come up when I start thinking about using something like this are (a) the risk associated with having the video out there where it can be lost, (b) the fact that in most forensic investigation of events you want to look at multiple cameras simultaneously, in which case you still have to get the video back to your central location somehow and (c) the attractiveness of letting the edge device decide what's useful video and what's not doesn't match a lot of real situations where you're looking for "the guy in the red sweater who walked by some time between dusk and dawn along with a gazzillion other people".

Of course, people are selling these solutions so there's obviously some market for them, and I see the benefit for small remote applications where you can't get the budget for a DVR/NVR, but personally I think these things are a side-show product, not mainstream stuff.

Re: HDcctv Alliance Competitive Analysis

GeoffMoore — Thu, 18 Jun 2009 10:03:41 -0000

Ugh...Mr IP, the reason that everywhere you go you see this is because this is the real world.

The only people who beleive there is a "battle" between analogue and IP are the ones who're willing to buy in to the idea that so-called *IP cameras" are superior to so-called "analogue" cameras just because they're IP.

I don't know - what industry are you actually in Mr IP? Maybe you're in the IP camera industry, and that would explain why you think that blindly aligning yourself with one technology or another is a good and worthwhile thing. I work in the security industry, and for me CCTV is one of an array of techniques and technologies that I use to try to help my clients mitigate risk. I have no vested interest in any technology, only in doing a good job for the clients, reducing their risks - which usually correlates in some way to reducing their costs.

For our customers it's about getting the right solution at the righ cost, not about using the latest technology.

Where I think that you and many others get blinded is that you can't make the separation between CCTV and networking. Clearly a network is a sensible way of transferring large amounts of data from one point to another and you have a lot of different highly resilient ways to do that. If we were thinking about connecting two computers together to send data from one to another, how else could we do it?? We couldn't use RS232 or USB if they were a long way apart. We could copy the data onto some portable medium and walk it over, but that wouldn't be convenient. A network makes sense, so that's what we use.

With CCTV - on the other hand - we have another choice, there has always been a simple medium for connecting cameras from one point to another, and the only limitations it has are that it's not bidirectional (which we don't care about all that much), it can't be used to transmit other media at the same time (which again, we don't really care much about - we can sent telemetry down the coax if we want so it serves all our needs). It doesn't have too many limitations on distance, and we actually have an option of using either coax or twisted pair, both of which are very cheap and can even be used to provide power to our end devices. The final limitation is that the signalling medium cannot support cameras higher than around 4CIF resolution, which is a shame, but doesn't exacltly take away our reason to live. We can always put up multiple cameras if it's that important.

Of course we're all sensible enough to realise that in some difficult connectivity situations it's easier to encode the video as an IP stream and use some IP media or other to connect, but that's only in certain circumstances where it makes technical and financial sense. Why spend all that extra money in locations where it delivers zero additional value?

"IP cameras" are a separate issue. There are an increasing number of people around who have never actually seen what a good image looks like - and I realise I may be sounding like a Betamax head arguing against VHS - but I'm afraid that it's been my job to look at images for a very long time, and I do know the difference between a good one and a bad one when I look at it. I can get good enough images from dozens and dozens of $100 cameras that work just fine and help me do my risk mitigation job for my customers. Where will I find a $100 IP camera? The cheapest IP cameras out there do not give me good images.

So it's not about cameras at the bottom of the range? It's about high spec, high functionality cameras? How much will I have to pay for a 5megapixel camera that will deliver that resolution at 25fps (PAL) and give me good performance below 1lux with wide dynamic range? Show me an IP fully functional camera you'd use for streetside applications that would give you zero latency connectivity so you can follow someone about at full zoom. It's nonsense. There is no battle. If you need these types of functionality you chose the *best* solution not the *IP* solution.

Look Mr IP, maybe you work in some niche sector of the industry where ivory towerism applies but you clearly are not working in the mainstream security sector where there are far more reasons to stick with analogue than there are to move to IP.

The last major compelling reason that has not been entirely blown out of the water is the ability to use megapixel cameras. It's the last one that I don't have a good argument for, but if I could get a megapixel camera and operate it along side all my $100 minidomes over existing cabling, with just an upgrade of the DVR I know that I could continue to sell it, and I'd beat all the guys like you who insisted on bidding IP every time without any really good compelling reason why the customer should pay the extra price every day of the week.

Come and be my competitor. That would be great. I could increase our margin.

Geoff
www.ipvideo.ie/blog

Re: Genetec, Milestone and OnSSI NVR Appliances Launched

GeoffMoore — Tue, 16 Jun 2009 23:48:46 -0000

JVC (and I think Sanyo..?) already produce small NVRs that run Milestone, one of which is a hybrid with a few BNCs on the back thereby enabling you to make a fairly smooth upgrade from analogue to IP without having to replace all of your cameras.

Conceptually I like this idea, because in really small applications you may just want a couple of megapixel cameras to work alongside your cheap and cheerful minidomes. Also it gives you a path to the future when you perhaps have a lot of tiny distributed sites but may want to integrate them in to a much larger scheme in the future.

Where the problem lies is that the units themselves are really quite expensive. When the price tag is high end users don't usually think "oh, ok. I'll spend all that extra money", they think "oh, that's expensive. Maybe I don't need this after all".

Re: How Well do IP Cameras Work in Low Light?

GeoffMoore — Wed, 10 Jun 2009 15:43:42 -0000

indeed, Sanyo make some great products.
I really like their "PanFocus" products. No depth of field issues, everything the camera sees is just in focus (which means you don't need to focus the camera in the first place).

Personally my favourite WDR camera remains the Bosch DinionXF in either the box or FlexiDome versions and in both the analogue and IP versions. They just perform, straight out of the box in low light and in WDR environments. Not cheap but you know they're going to work. The IP version runs a little hot for my liking, but otherwise a very solid camera with very low noise.

It's just a shame that they haven't taken the same technology over into their speed domes.

Geoff

Re: How Well do IP Cameras Work in Low Light?

GeoffMoore — Tue, 09 Jun 2009 14:53:26 -0000

To be honest John, I've found Pixim based cameras (both analogue and IP) to be noisy in low light.

Maybe it's all become fantastic over the last year and I'm missing out on something fantastic, but I have other less expensive ways of skinning the same cat that I know work. So I'd need to be convinced by some very compelling real world tests.

Side by side tests of different camera types are the only way to get a proper feel for the comparative performance of different camera technologies, and although advances have been made in a lot of areas with Megapixel cameras they're still not a patch on some of the old reliable performers.

Geoff

Re: How Well do IP Cameras Work in Low Light?

GeoffMoore — Tue, 09 Jun 2009 14:46:43 -0000

This is a big issue with the way CCTV systems are specified in general. We have consultants who write performance specs that define a camera in terms of parameters lifted from a datasheet like "the camera shall have a resolution of blah blah and will have a sensitivity of blah blah lux."

It's like they've forgotten that what the end user wants is "the camera will produce identification grade images from location x for targets of size range y moving at z meters per second under the lighting conditions prevelant at that location throughout a 24 hour period in all expected weather conditions".

Defined in terms of such operational requirements instead of meaningless technical terms, it becomes the integrator's responsibility to do a proper survey to understand the prevailing conditions under which the camera needs to work, select the right camera and lens for the job and put it in the right place and set it up properly.

And end users tell me that because they're not experts in CCTV they have to employ consultants to write their specs. What's technical about knowing what you want the camera to do?

Sure, there's technical stuff to be defined but it's mostly just detail and if it's important to the operation of the system then it can be defined in non-technical operational terms. Getting good pictures of the right thing is what a CCTV system is all about.

Geoff

Re: How Well do IP Cameras Work in Low Light?

GeoffMoore — Tue, 09 Jun 2009 09:55:19 -0000

And another thing...(and this is not really about IP cameras, it's about cameras in general), I've always thought it would be a good idea to be able to attach a bunch of other parameters to the day/night switching event, because day and night are such completely differnet environments that I sometimes find I that if I set a camera up to give me an optimum image not be the best image possible during the day.

So in *some* situations I might want to set the gain parameters or white balance differently at day and at night, and it might be nice to attach those to the day/night switching event.

This sort of thing lends itself more to an IP camera, simply because you've generally got a much richer user interface through which to configure the camera than a few buttons and a simple on-screen menu.

That's just a by-the-way comment...

Cheers
Geoff
.

Re: How Well do IP Cameras Work in Low Light?

GeoffMoore — Tue, 09 Jun 2009 08:05:35 -0000

Hi John
Although low-light performance is something everyone gets quite excited about I'd actually contend that it's as much about dynamic range than just the ability of a camera to produce an image when there's not much light about. If it *is* dark, then frequently the object you actually want to see are going to be *very* bright by comparison, and you often get issues with standard cameras of washed out, overexposed images or just very bright lights in a sea of darkness and no discernable detail at all.

Also extremely important (and mostly not understood by a lot of people selling cameras) is that the sensitivity of the camera defines its ability to detect light at the sensor, not out in the scene. You may wave your light meter around and see acceptable levels of light in the areas where your camera is aimed, but that doesn't mean that those same levels of light are going to get back to the camera. The colour and reflectivity of the materials in the scene define how much incident light is reflected back to the camera, and the aperture of the lens determines how much of the reflected light makes it onto the sensor itself.

It's a big challenge with a lot of standard cameras to get them setup to give you a good picture, only to discover that when it rains you can't see anything because of all the reflected street lights.

I'm frequently putting cameras into applications in daylight where there's 10,000lux + hitting a target in broad daylight, but every shadow just looks black because the camera can't cope with the overall contrast in the scene. There aren't many cameras that work really well in these types of environment, but the ones that do it best are still standard resolution analogue cameras (or the equivalent analogue camera with an encoder built in).

Cheers
Geoff

Re: How Obama Impacts the Security Industry

GeoffMoore — Fri, 07 Nov 2008 08:31:43 -0000

Hi John
Yes, that pretty much bears out my understanding of the US market, and it's exactly why I refered to it as "ambulance chasing".

Geoff
www.ipvideo.ie/blog

Re: How Obama Impacts the Security Industry

GeoffMoore — Thu, 06 Nov 2008 17:44:39 -0000

Whilst the election of the new president *may* have some influence on the homeland arena (and I actually think that it'll be foriegn policy that gets changed rather than the situation back home) it won't have any bearing on the more general day-to-day uses of security technologies - i.e. the fight against crime.

Look at the UK, for instance. The country with the highest number of CCTV cameras per capita in the world. Where are all the cameras? Town centres, shopping malls, banks, airports, roads, rail stations. These are the sectors in which the industry is concentrated and the continued growth in the market over the last 20 years+ has had little or nothing to do with which government happens to have been in power at any particular time.

Sure, there was a surge in anti-terrorism related technologies after 9/11 and 7/7 but those events were blips on a constantly rising slope.

I'd go so far as to accuse those companies who're banking on large Homeland Security projects (the ones that might get cut now) of *ambulance chasing*. Why don't they just start working towards selling systems to cities to reduce crime and anti-social behaviour and actually do something to improve the quality of people's lives?

OK, you have different problems over in the US to those we have here in Europe. You have long porous borders where we just have ocean. But all the same, you have plenty of internal problems that could be worked out if people would just apply themselves rather than fretting about foriegn matters.

Geoff Moore
www.ipvideo.ie/blog

Re: Video Surveillance Book - 2nd Edition Released

GeoffMoore — Tue, 28 Oct 2008 17:28:43 -0000

Well done and congratulations, John!

When people produce stuff like this and put it out in the public domain it raises the bar for everyone, and takes away the excuse of ignorance from those who fall prey to the untruths and missinformation peddled by so many in this industry.

Keep up the great work!

Geoff Moore
www.ipvideo.ie/blog

Re: How Strong is the ROI / TCO of IP Video Surveillance?

GeoffMoore — Fri, 24 Oct 2008 11:02:45 -0000

I see a lot of instances where people use IP connectivity with analogue cameras, though. The Verint S1100 type wireless point to point gear actually encourages this (ok, maybe it doesn't *encourage* it, but it really *facilitates* it very strongly).

I rambled on about the whole megapixel camera situation a while ago (http://ipvideo.ie/blog/?p=19) and I pointed out that the re-usable cabling argument isn't always very strong either (http://ipvideo.ie/blog/?p=39).

I don't think the RoI argument for IP video surveillance systems based on like-for-like replacement of analogue systems stands up. However, I *do* think there are applications where you can use the technology to do things you cannot do with analogue technology. That's not really an RoI argument, it's a risk management argument or at least a cost/benefit argument.

When we get cameras that work at megapixel resolutions, in wide dynamic range lighting conditions, at high frame rates, AND we can convince clients to put them in all the right places around their sites, THEN we'll get CCTV surveillance systems that actually do what we'd want CCTV surveillance systems to do. There's no increase in RoI for that other than the fact that the analogue way with poor resolution cameras operating at low frame rates etc is a BAD INVESTMENT because systems like that are of little value other than as deterrants (and we know that a determined intruder is not going to be deterred by CCTV, he'll just put on a mask and break in anyway).

Re: Credit Crisis Accelerates Convergence

GeoffMoore — Tue, 21 Oct 2008 11:54:50 -0000

Hmm...

"Credit Crisis Will Accelerate Convergence" - no, I'm afraid I see that as more of a hope than a fact. Facts need evidence to back them up and I'm not seeing it.

Once again we're back peeling away layers of the emperor's new clothes in search of a practical reality (if that's not too much of a mixed mataphor). Some customers are making *ideological* or *policy* decisions to go the IP route based on a fairly vague premise that it provides increased ROI and reduced COO. I probably agree with that decision made on that basis (although I think most people would be in for a surprise if they really did a completely unbiased analysis of the costs), but I still reserve my right to disagree on the basis that a significant number of the stated advantages will rarely ever be realised in the majority of real-world implementations. That's okay, these guys are making policy, and I like policy.

However, when I look at my customer base I see comparitively few organisations who base their purchasing decisions on policy or ideology outside of the non-commercial government, education and transport markets. Most decide what to buy on the basis of a very inadequate and innaccurate understanding of the technology (often guided by equally misinformed *expert advisors*) and cost of purchase (or cost of purchase plus 3 years maintenance, say).

Where there are no specific requirements that cannot be fullfilled any other way than with IP camera technology, IP based connectivity or IP recording solutions, a like-for-like analogue implementation will generally win on cost.

We continue to see the biggest market for video surveillance in commercial applications. Without the growth in sales volumes of IP equipment that rapid inroads into the commercial sector will bring, I do not envisage a rapid decrease of component prices or a rapid expansion of market familiarity.

The absense of such factors is likely to slow the rate at which companies adopt IP technology *as a policy*, and hence the procurement decision in this sector is likely to remain highly influenced by cost of purchase.

Of course, generalisations don't help. There will be (as there always are) market segments that will benefit from the so called "credit crunch" but I think the benefit of forums such as this is for those of us who are out in the real world to share real information about how the market is actually developing, rather than speculating on how we'd like it to develop.

Geoff Moore
www.ipvideo.ie/blog

Re: Does Lower Expectations Solve Video Analytics Problems?

GeoffMoore — Thu, 09 Oct 2008 16:24:40 -0000

Well, I guess that's the big question John. What "working well" means to different people.

I've trialled a few of the "analytics in a box" type solutions with varying degrees of success (I'm not going to detail all of the results right now), and we're currently protecting the perimeters of a couple of small compounds using Mate Trigger boxes (Bosch DinionXF LTC495 cameras with Raytec RAYMAX50 infrared illuminators connecting into a Heitel CamServer through to a remote monitoring centre). The cameras loop through into the Mate units and the alarm outputs connect back into the alarm inputs on the DVR. The cameras are covering area inside the fenceline to give early warning of intruders entering the compound before they reach the buildings (which are all individually alarmed too).

This is a really primitive way to use these units but we were constrained a little by how the monitoring centre needed to receive alarms and by how far the customer would "believe us" that this was going to work!

This is a very remote location with no artficial lighting at night (apart from the IRs on the cameras), no traffic or passers-by in the area, quite a lot of foliage, plenty of small wild animals but no cattle/farm animals.

Each camera is only covering quite a short length of perimeter (because the compounds are small) and the fenceline is straight (no curves or complex geometrical shapes). The cameras are up on 6m columns looking down the fenceline.

We stipulated to the client that this had to be a really clean area with no clutter, and the remote nature of the site made it easy for us to cut down on things that *look like* people.

Unfortunately I didn't do the commissioning so I can't give you any gorey details about how long it took to set this up, or if there were any specific things that tripped the guys up during the process - but I do know that the systems were set up over the course of about a week with testing during day and night. Although I don't have exact figures I also know that the false alarm rate is very low (the monitoring centre would have pulled the plug if they were getting excessive alarms). They do a video patrol a couple of times a night then respond to alarm triggers if necessary - from what I've seen these are rare.

We have scheduled quarterly maintenance visits to site to routinely check on all of the security equipment, this includes cleaning and checking focus on cameras. I'm not sure if anyone's done a thorough walk test on the analytics lately...I'd have to check that.

I'm looking at two other remote sites for the same company at the moment, both of which are going to be Milestone based. I was strongly considering using Cernium Perceptrak for these, but the integration of Mate with Milestone is now done so I think I may stick with them (I've also had no fun at all trying to get a sensible response from Cernium distributors on this side of the Atlantic - Cernium seem to have a hard time understanding what it's like for us guys at the coalface trying to get these products into new markets).

So, in this application, analytics are doing what we want them to do - but we do have the advantage of a very clean environment and a relatively small area. I'm currently looking at another site where we're aiming to protect a building in a much busier area. I'm quite keen to see how we get on with that as I think it's going to be a lot more "challenging". It's also going to be an IP solution with an analytic server sitting alongside the storage server - and that will be the first time we've done that.

Re: Does Lower Expectations Solve Video Analytics Problems?

GeoffMoore — Thu, 09 Oct 2008 11:06:54 -0000

We (in this industry at least - but probably in others) have a real habit of wandering around with a solution looking for a handy problem. I think "the industry" has done this with video analytics.

The video analytic everyone used to show (with much flourish and fanfare) was the old "left object/removed object" one, but really, when you think of that in any practical situation it would be extremely hard to find a situation when they'd be much *real* use. So the industry itself set the bar pretty high by telling the lay-person that functions like that were going to be possible and practical.

Good video motion detection and virtual trip-wire analytics are available and they work very well...but they're not very exciting, because lay people think it's trivial to do stuff like that. People counting works too (as demonstrated again by me quite recently with the Cognimatics people counting solution over at http://ipvideo.ie/blog/?p=5) but laypeople don't think of people counting as being a video application until you show them how well it works.

Some of the very best analytics I've seen have been designed to solve very specific problems for very specific applications. I'd point to a tiny UK company called Pepperdog (http://www.pepperdog.net/in... who produced a system for railway level crossing that detected anyone on the tracks not wearing a hi-vizibility jacket (thereby indicating that they were not rail workers - or that they were railworkers who weren't following the health and safety procedures.

They also put in a people counting system at Stanstead Airport (I think) that automatically counted people using the public toilets and sent an SMS to cleaning staff every 50 uses - rather than having the cleaners visit on a time schedule regardless of whether there'd been any useage.

These are good, real-world things with solutions aimed specifically at a tangible (and relatively easily defined) problem.

I remember reading some market research some time ago in which they discovered that non-computer users expected computers to be capable of the same level of functionality as the computers they saw portrayed in Star Trek. I think we as an industry need to help the real world get real about what analytics are capable of, and stop hyping up things like "trip'n'fall" and "spill-alerts" that just aren't practical.

Re: Benefits of Standards for Specifying IP Cameras

GeoffMoore — Wed, 08 Oct 2008 08:08:56 -0000

Hmm...now, that looks like a very interesting document. I'm going to have to sit with this one for a while...

I'm a very big fan of operational requirements specifications as opposed to *technical* requirements specifications. The technical ones are not really helping - they're often "written" by non-experts and don't reflect what the client is really looking for. They also concentrate the vendor on the technical detail - encouraging everyone to play *spec-manship*,. where you're just trying to place the lowest cost against an *interpretation* of what the specifier wrote (often taking into account little grammatical errors to get your price lower than the next guy's).

You can drive a bus through most technical specs, not least because you won't see two manufacturers specify the performance of their cameras in precisely the same way, and ALL manufacturers go to great lengths to make their products look as good on paper as they possibly can (you can't blame them for that really, but it makes like-for-like comparisons between equipment specs very tough).

In my view, if you have an operational requirement stated, you take away all the technical mumbo-jumbo that the consultant loves (because it makes him look like he's earning his fee) and converts the spec into a bunch of non-technical descriptions of how the system ought to work. I think you also place the onus of responsibility to achieve the customer's requirement upon the vendor, and if they cannot prove that the system operates the way it was specified then there is a basis for the customer to withhold payment (the prime motivator towards a vendor doing a good job).

Recognised specifications are also good (in all things) because they also define a vocabulary that everyone can recognise and use. When you're at the wrong end of an un-level playing field the whole game just becomes one big uphill struggle :)

Re: Is Axis IP camera dominance coming to an end?

GeoffMoore — Wed, 08 Oct 2008 06:31:04 -0000

I think sometimes a company's early stage desperation to get sales can translate into a strategy that they end up being stuck with (to some extent). I really don't think the "everyone's a partner" idea was a good one. As Syndrome said in The Incredibles movie "when everyone's special, nobody is" - but it's a very difficult U-Turn to change that now.

But at the same time, the fact that they're EVERYWHERE gives them a big viral advantage. Everyone who talks about IP can't help but mention Axis somewhere along the line. People who know next to nothing about IP probably recognise the brand.

I still get blank looks from people when I mention Mobotix and ACTi and IQeye, but you say "Axis" and the little light goes on in people's eyes. When these other guys catch up with their brand visibility I'd say Axis are screwed.

On the *cost* issue - hey, look, I've been designing embedded products for years. It's hard to find a micro these days that doesn't come with an IP stack and network interface included. Throw away all that horrible analogue drive circuitry and stick in digital land!! Okay, you probably need a little more memory in the camera and you're going to have to run a webserver in there too, but you know - sacrifice a tiny bit of camera performance and you could make and sell cheap IP cameras for the bottom end of the market. Get the volumes up and all the prices will fall, and Axis will have to compete or move on.

I think I said that I'd like a 4CIF minidome - heck, I'd be happy with a CIF IP minidome if I could get it at the same price as a reasonable analogue equivalent! Most of the cheap DVR solutions I'm trying to compete against are recording at 2CIF or worse anyway.

Re: Is Axis IP camera dominance coming to an end?

GeoffMoore — Wed, 08 Oct 2008 06:10:04 -0000

_blush_ thanks John... :)

Re: Will Security Integrators Survive IT Convergence? : IP Video Market Info

GeoffMoore — Wed, 08 Oct 2008 05:39:04 -0000

Maybe it's different in different parts of the world, but speaking for the industry here in Ireland I definitly believe the traditional installer/integrator sector is going to lose a massive amount of ground to IT/networking companies over the next few years. Here are a few of the reasons why :

We have no proper systems integrators. We've got plenty of installers, a few of whom have rudimentary networking knowledge (enough to set a static IP address in a camera or DVR but usually not enough to set up network from scratch). Sure, not many network installers know a whole lot about camera technology or the physics of light, but - surprise surprise - not many installers actually do either.

The larger installer/integrators here are significantly unionised, and their engineers actively reject the possibility of retraining and upskilling - mad though that may seem. It makes them inflexible and costly (because job demarcation forces them to use several specialists instead of one flexible guy), and basically unattractive.

IT people want to deal with IT people. I don't trust the majority of security technicians to work on a customer's network and neither does the customer, but they'd trust a trained and qualified network technician to do something relatively simple like put up a camera or install an alarm (in the majority of cases they're likely to be able to do just as good a job as the security technician, the only question is whether the network technician would *stoop* to that level and do something so *menial* as security work).

Larger corporations who're taking organisational security seriously are beginning to look at larger frameworks such as ISO27001 because they fit in neatly with other corporate standards such as ISO9001 and ISO14001 that they're required to meet because of the type of business they're in. This is *framed* as an information security standard and the majority of what's in there relates to IT and organisational issues, but it includes a significant section on physical and environmental security too - and after all, if you're looking at security in the modern context you simply cannot look at physical security in isolation. I don't know of a single integrator that's currently offering end-to-end security of this nature either on their own or as part of a consortium of service providers, but I know several IT/networking companies who're offering ISO27001 auditing and certification - so they are (by implication) involving themselves in physical security.

I don't think that integrators are going to be *wiped out* by convergence, but I do think that the landscape is going to change significantly and some integrators are going to be left behind.

It's strange. Intregration to me used to mean writing software and understanding weird and wonderful protocols and doing serious design work. These days it just seems to be about plugging a few alarm inputs together and making something go "beep".

Re: Video Surveillance Platforms

GeoffMoore — Tue, 07 Oct 2008 16:52:51 -0000

I've been banging the Open Platform drum for a long time now. From the perspective of the security management "platform" product (because you can't consider a *video only* solution to security, you have to look at the whole security operation end-to-end, with physical security and surveillance existing as significant components within the whole) I'm quite impressed with Orsus (www.orsus.com).

It's got a big price ticket but that's ok. The early entrants are often overtaken by lower cost equivalents at a later date.

In my own particular geographical region nobody would buy something like this anyway. Nobody's willing to pay for an integrator to do any serious integration, and hence there aren't any serious integrators. Chicken and egg. We also have an extremely conservative market in which nobody will do anything unless they know somebody personally who's already done the same thing. This makes new product introduction incredibly frustrating.